Analysis and Optimization of Landscape Preference Characteristics of Rural Public Space Based on Eye-Tracking Technology: The Case of Huangshandian Village, China

Abstract

As an important part of rural space, the public space landscape has a positive influence on tourists and local residents, and also has an important guiding significance for the sustainable development of rural areas. This study expands the previous research on rural public space, taking the “user-oriented” landscape preference as an important entry point, taking Beijing rural public space as the research object, using eye-tracking technology to objectively reflect the degree of users’ preference for the landscape, making quantitative analysis through eye-tracking objective data and subjective satisfaction evaluation to explore the subconscious and aesthetic laws of the subjects, and summarizing the internal driving factors leading to its evolution from the summary of phenomena. This paper explores the visual behavior information and significance behind the “subjective space” created by the countryside for the users, intending to provide strategies for landscape optimization and the management of rural public space in an effort to aid the sustainable development of the countryside.

1. Introduction

In recent years, the population migration and construction of new villages in China’s urbanization process have intensified the socioeconomic development and changes in the countryside [1,2,3,4,5]. According to China’s 2021 National Bureau of Statistics, China’s rural population has decreased by 164.36 million in the past 10 years, and the proportion of the urban population has increased by 14.21%. China’s urban–rural dichotomy is polarized, with megacities dramatically expanding on the one hand, and some small and medium-sized cities, especially in rural areas, severely shrinking on the other. For example, the urbanization rate in Beijing-Tianjin-Hebei has risen from 34.1% in 2000 to 62.5% in 2015, with an average annual growth rate of 1.89% [6]. Wuhan has experienced rapid urban expansion over the past three decades, which has been driven by topography, transportation, and socioeconomic factors, with the proportion of rural arable land decreasing by 6.18%, and the proportion of built-up land increasing by 6.51% [7].

1.1. Rural Public Space

In the context of China’s rapid urbanization, the habitat of the countryside has undergone dramatic changes. As the physical place in which people carry out various ideas and social activities, such as interpersonal communication and cultural and information exchange, the public space of the countryside has borne the brunt of the changes. In terms of spatial function, a large number of rural public spaces have been swallowed up by private spaces in the process of urbanization, losing their public nature and hosting a single type of activity [8]. In the establishment of village buildings in a rural village in Zaozhuang, Shandong province, the public spaces at the peripheries of the buildings are increasingly far from private space, reducing the use of public space [9]. In terms of the countryside appearance, rural public space construction lacks character and is seriously homogenized [10]. For example, the horse head wall, the representative element of the Huizhou school of architecture, one of the most important schools of traditional Chinese architecture, is blindly used in the construction of major villages across the country, which has resulted in a uniform countryside appearance. In terms of emotional identity, the public space consists of a large number of buildings that either do not match the rural appearance or are commercialized, and there is the phenomenon of the excessive consumption of vernacular culture, which causes villagers to fall into a crisis of identification with their own vernacular cultures [11]. For example, Li discovered that there are more than 271 million villages in China today compared to 363 million communities ten years ago. Although new rural building benefits certain common communities, it nonetheless obliterates thousands of traditional villages that are rich in cultural, scientific, and aesthetic characteristics [12]. Obviously, in the context of rapid urbanization, rural public spaces are experiencing more serious trends in terms of the lack of humanistic care [13], loss of local characteristics [14], the disintegration of intimate communities [15], unfamiliarization of villages [16], and spatial shrinkage [17]. A contradiction exists between the existing supply of rural public spaces and the demands of residents [18,19], and the relationship between the people and land is also acutely tense.

Compared with urban public spaces, researchers in various countries have focused less on rural public spaces. Most of the existing studies on rural public spaces focus on the creation of physical space. For example, Chen et al. studied the construction of public space in villages based on the theory of space syntax, summarized the problems in its construction, and proposed three optimization strategies, especially increasing the public activity space and optimizing the public environment of villagers [20]. Ruilian et al. took 37 districts and counties in Chongqing as the research objects, used the entropy-value and spatial-autocorrelation methods to empirically measure the rural function levels, and analyzed the spatial and temporal evolution of the three major functions (rural production, living, and ecology) and the interactions among them to provide theoretical support and a practical reference for rural revitalization and sustainable rural development [21]. Erickson et al. examined the reasons for the rapid changes in the rural woodlands in the rural landscape of Michigan (the United States) compared with elements of public space, and they found that aesthetics and environmental protection were important motives for maintaining rural woodlands [22]. Some scholars have also studied the development of rural public spaces from a time-dimensional perspective [23]. Dong et al. used a timeline to sort out the dialectical relationship between the function and form in the development of rural public spaces in China, and they proposed a reshaping of the elements of the public space, such as the scale, proportion, structure, and quantity [24]. Other scholars have attempted to apply different research methods to assess the elements in rural public spaces, such as Suh et al., who studied the establishment of an assessment system for the suitability of rural facilities and public spaces, identifying eight assessment categories and 32 indicators to assess the suitability of public spaces and facilities in rural areas in terms of the use, environment, and visual value for use in the development of guidelines for the planning and design of public spaces and facilities for rural communities in Korea [25]. Moradi et al. employed field studies to observe the surroundings of Azghad village, conducted interviews with locals using a qualitative research methodology, and investigated the physical aspects of this rural hamlet using documentary studies. The components of the social-physical system that make up the residential type of the village were retrieved via an examination of the physical pattern of the village in the region under investigation [26]. According to the current research progress: (1) Most of the existing studies on rural public spaces are focused on the physical space, such as the scale, type of space, natural environment, evaluation of elements, and other aspects of the physical space, and mostly from the perspectives of designers and builders, while there are relatively few quantitative studies on public spaces from the perspective of users. (2) In view of the current social background of rapid urbanization, villagers’ subjectivity and needs have been easily neglected for a long time. To solve the social problems brought about by urbanization, we need to pay attention to and identify the evolving psychological demands of users. Therefore, in the construction of rural public spaces, the focus should not only be on the construction of the external rural physical space and landscape environment but should also identify the landscape preferences and psychological demands of different user groups from the perspective of the user’s needs.

1.2. Landscape Preference Theory and Eye-Tracking Technology

As the cognitive judgments of and preferences for environmental information that are formed by users through sensory reception and empirical interpretation in the process of the human–environment interaction [27], “landscape preference” is not only an important theory for studying the human–land relationship [28], but it is also a breakthrough link for a focus on the relevance between rural public space construction and the users. Researchers use many theoretical approaches in the study of environmental psychology for the evaluation of landscape preferences, such as content identification [29], for understanding the different types of landscape-preference patterns [30], the semantic difference approach [31], and landscape-scenic-beauty analysis [32], which they have widely used to evaluate preferences and aesthetic values, such as the vegetation and landscape topography [33,34]. In such studies of the spatial cognitive perception of the landscape, most of the studies use questionnaires and scales [35], and a few studies combine physiological measurement techniques [36]. Although the aesthetic evaluation of the environmental landscape is more tangible and contributes to the evaluation of user satisfaction, the formation of a definite and objective measurement of user perception is still a challenge and often calls for the use of more holistic and innovative approaches and methods [37].

Therefore, in this paper, we introduce eye-tracking technology based on the study of landscape preferences in rural public spaces. Eye tracking is a new research tool that has improved in price, portability, and accuracy, heralding a new era of big data collection and analysis for landscape preferences [38] as it offers a set of measures used to detect and record eye movements and is useful to analyze visual attention, perception, and connection between the brain and outside world [39], and has unique advantages in the processing of visual information [40]. In psychological experiments, researchers can use it to collect the relevant physiological indicators reflected by the visual information of the public space landscape [41], visualize the subjects’ visual concerns, and analyze their mental activities and perceptual processes with objective data [42], which objectively reflect the preference data of the user population for the landscape in the space [43], and which researchers can use to analyze the visual needs of the subject through objective data [44].

To provide new ideas in landscape perception research, many scholars are now introducing eye-tracking technology into the study of different landscape spaces. For example, Schirpke et al. collected 19 indicators that described the features of the hotspots that were discovered using the Visual Attention Software by 3M and that were based on 78 panoramic landscape pictures that represented the key landscape types of the Central European Alps (3M-VAS). The potential of using eye-tracking simulation software in mountain landscapes was investigated and explored [45]. Amati et al. performed eye tracking on 35 participants as they walked through two different parks in the urban center of Melbourne. The growing mobility, sophistication, and affordability of eye-tracking technology were used to investigate its utility in analyzing landscape preferences [38]. Meanwhile, Wu et al. examined the effects of urban green space (UGS) design intensity on landscape preference, restorativeness, and eye movement, using treatment images with 200 students as participants. The most feasible regression equations between design intensity and preference, restorativeness, and eye movement were obtained [46]. In addition, Ding et al. used eye-tracking technology to study the extent of people’s preference for native greenery plants from the perspective of visual preference using pictures of plant organs such as leaves, flowers, and fruits as stimulating materials to improve rural greening techniques, with a study conducted by students from the Central South University of Forestry Technology and villagers in Changkou Village, Fujian Province [47]. Some studies explore the application of eye-tracking technology in the landscape from different populations. For example, Li et al. divided 90 participants into two groups of outsiders and insiders to view photos of Han Buddhist temples. Consensus and differences in their visual preferences and eye movement metrics were assessed [48]. Dupont et al. divided the experimental group into 21 landscape experts and 21 amateurs and asked them to observe 74 landscape photographs to explore whether expertise in landscape-related issues affects the way landscapes are observed [49].

From this, the following conclusions can be drawn: (1) More and more studies of landscape preference have introduced eye-tracking technology and obtained reliable experimental results, and the application of eye tracking in landscape preference has great potential. (2) Most of the studies on landscape space using eye-tracking technology are on urban space, and the studies on rural public space only focus on elements such as plants, or the mountains and forests on the periphery of the countryside, and relatively little on the use of eye-tracking technology in rural internal public space. Currently, as rural public space in China is undergoing urbanization changes, the public attributes of the countryside are constantly changing. As an important tourist village, local residents and visitors are two important types of user groups in Beijing’s Huangshandian Village, and if the visual landscape preferences of local residents and visitors in using rural public space are unknown, it will be challenging to make effective recommendations for landscape enhancement as well as resource management in Huangshandian Village. Therefore, this paper aims to combine subjective assessment and eye tracking to explore the perceptions of local residents and visitors on different rural public spaces, and to enrich the application of eye-tracking technology in landscape preference studies. It also provides a quantitative index and a powerful guide for the optimization of rural public space landscape, which is more conducive to the optimization and construction of rural public space landscape.

1.3. Research Objective

In this study, we took the public space of Huangshandian Village in Beijing as the research object. We extend the previous research on rural public space and take the “user-oriented” landscape preference as an important entry point, mainly by using the eye-tracking instrument as a powerful objective quantification tool to study the preference elements of different types of public space landscape from the perspective of landscape users, filling the gap that landscape preference research creates by mostly relying on subjective evaluation, focusing on the visual information of different participants when enjoying the public space landscape, and objectively reflecting the degree of users’ preference for landscape. This study attempts to answer the following questions: (1) What are the participants’ visual preferences for different types of rural public space landscapes? and how do the landscape characteristics of these spaces affect participants’ landscape evaluations? (2) What are the differences in viewing the same public space scenes among different user groups? (3) Is there a correlation between participants’ objective data and subjective evaluations when viewing various types of public space scenes? We sought the combination of subjective feelings and the optimization of the village spatial landscape, and we built a strategy for the healthy cyclic development of both to enrich the research system of rural landscapes, provide new ideas and a basis for the optimization of rural public space landscape construction, and to guide the sustainable development of rural spatial landscapes.

2. Study Area

2.1. Study Site

Huangshandian Village is located in the shallow southwestern mountains of Beijing, the capital of China, 50 km from downtown Beijing, located in Zhoukoudian in the middle of the Fangshan District. The village has a geographical area of 20.2 square kilometers, and it is located in a low mountainous area that is deep in the Yanshan Mountains at the junction of the North China Plain and Taihang Mountains, with a temperate semi-humid monsoon continental climate, four distinct seasons, and simultaneous rain and heat. Since the beginning of the 20th century, Huangshandian Village has experienced three different stages of development: traditional agriculture, industrial mining, and tourism services, which have relied on its rich tourism and mineral resources and its advantageous location, being adjacent to the capital. Rapid urbanization has had a significant impact on the rural habitat and public space.

The village has proximity to major scenic spots, a beautiful natural landscape, and numerous historical relics, with a characteristic backdrop of mountains and water (Figure 1), and it has become one of the first “Key Villages for Rural Tourism in China”. At present, Huangshandian Village has formed a development framework of a rural-ecological-life industrial district and whole-area tourism complex, integrating leisure, recreation, and vacation. As a typical suburban village of a megacity and demonstration site for rural revitalization on the cultural belt of western Beijing, the use of the public space [50], public attributes [51], and public activities [52] in Huangshandian Village have all substantially changed. As one of the first villages in China to start the pilot construction of new rural communities, Huangshandian Village has a typical theoretical background and research relevance in the field of landscape optimization and the enhancement of rural public spaces. In this study, we introduced eye-tracking technology to fill the gap in this rural landscape research direction from the perspective of users to improve the quality of rural public spaces, and to provide a reference for the construction of rural public spaces and sustainable rural development in other similar areas.

2.2. Public Space Selection

The rural public space in this study mainly refers to a public space in the narrow sense, which is an open space [53]. In order to better reflect the landscape characteristics of the public space in Huangshandian Village and the representative elements of the landscape theme, we adopted the comprehensive opinions of villagers, experts and the government for the experimental site selection and conducted several field studies. Firstly, we visited Huangshandian Village and talked with local residents to find out 62 public spaces where they often gather, and selected 15 spaces with a large number of people and high frequency of use, and divided them into four categories according to the shape, scale, and location of the public spaces, combined with the characteristics of interaction behaviors and functional needs and roles: the linear street space, surface-shaped square space, waterfront space, and point-shaped public green space [54]. We then invited several experts and teachers who have at least 10 years of experience in landscape design, environmental psychology, rural planning, and other related fields to visit the 15 sites after the initial screening. Finally, through discussions with the government, the most representative site in each of the four categories was selected and considered by the government to have the potential for subsequent practice. These four types of spaces are highly recognizable, have distinctive landscape features, and are important landscape space types in Huangshandian Village that are closely related to the villagers’ daily lives and have various functions, such as leisure, communication, access, etc. The villagers and tourists can move freely and exchange information in these specific public spaces:

(1)

Square space landscape (Figure 2a): The cultural plaza is the recreational activity center of Huangshandian Village, where villagers and tourists concentrate to participate in recreational activities and information exchange, and where most of the village’s important activities take place, such as festival and wedding celebrations, creating an important space for the rich and colorful cultural activities and intimate neighborhood relationship in Huangshandian Village [7]. The plaza space is located in the center of the village, at an important position at the intersection of streets and alleys, the paving of which consists of the characteristic local stones, and it has simple and ancient vignettes.

(2)

Street space landscape (Figure 2b): The street space landscape is a typical landscape of Huangshandian Village. During the field research, we often observed the local residents walking in the street space and talking and communicating with each other, as well as general village activities [55]. The street space of Huangshandian Village is mostly north–south in depth, connecting small public spaces and organizing traffic, and the buildings on both sides have the obvious characteristics of the mountain dwellings in western Beijing, with many stone walls and tile and slate roofs.

(3)

Waterfront space landscape (Figure 2c): The waterfront space landscape is an important characteristic of the landscape of Huangshandian Village. The main river in the jurisdiction is the Xiekuo River and the landscape is smooth in the peak season, but weak in the dry season, which is an important space for recreation, distribution, and sightseeing in Huangshandian Village. Both sides of the water area are hard revetments. People carry out spontaneous activities of interaction and communication in the open space beside the river. As an obvious and important linear node space in the village, it is one of the most natural living spaces in the village. The selected sites for the experiment include stone bridges across the river and open green spaces beside the river. This stone bridge is located in the central area of the village waterfront system, which bears the important traffic function of Huangshandian Village, and is the only way for villagers and tourists to pass. At the same time, the riverside green space connected by the stone bridge has been determined by the government as the next reconstruction area, so this area was selected as the experimental site. As shown in Dupont, L’s article [49] on landscape identification by professionals and non-experts, the experiment excluded such distracting experimental perspectives because non-experts often spend more time and attention on specific objects, such as parked cars, signs, and other non-landscape elements. In the actual space-use process, the riverside green space is located at the extinction point of the forward direction of the stone bridge; this angle can show the main landscape surface, and finally determine the angle shown as the experimental observation angle.

(4)

Public green space landscape (Figure 2d): The public green space is an important space for villagers and tourists to perceive nature. The space is related to the image and appearance of the whole village, and it consists of the old theater, residential buildings, waterfront walkways, and green landscape. During the research, we found that the public green space is mostly a stopping space for tourists on their journeys, and it takes part in the organization of traffic. The space is relatively open, which increases the relevance and integrity of the whole rural landscape. The selected public green space used to be the place where the old stage in the village was located, and it was the space for holding festival ceremonies in the village. After the relocation, out of emotional attachment to the familiar venue [56], local residents still spontaneously choose to gather in this area for assembly and exchange activities [57]. At the same time, experts believe that compared with the fragmented distribution of other public green spaces in the village, the public green space here has a large area and is located in the area where tourists and local residents are concentrated, which can be used as a representative of this kind of activity space. Considering the regional characteristics and historical inheritance of the government, it is also considered that the public green space landscape here has considerable practical potential in the future.

Figure 2. Four types of public space landscapes in Huangshandian Village. (a) square space landscape; (b) street space landscape; (c) waterfront space landscape; (d) public green space landscape.

Figure 2. Four types of public space landscapes in Huangshandian Village. (a) square space landscape; (b) street space landscape; (c) waterfront space landscape; (d) public green space landscape.

3. Methodology

Under the boom of rural tourism, it is more conducive to verify and judge the subjective feelings of two types of users towards the objective environment by studying both the tourist’s experiences and the villager’s feelings about its use. Analysis and exploration using eye-tracking technology can enrich the research on the perception of rural public spaces [58,59]. In the early studies of spatial perception, scholars used photographs and other means to extract cognitive information on the surroundings of the study subjects [60,61,62]. Visual images as reproductions of the particular ways of viewing specific groups of people [63] can be used to explore and analyze their perceptions of places and spaces [64]. Researchers have reported differences in subjects’ cognitive levels when observing different types of public space landscapes [42]. We can use eye-tracking technology to collect the eye-movement data recorded using infrared pupil cameras, which can be used as a source of research data and combined with subjective evaluation to explore the visual behavioral information and meanings behind the “subjective space” produced by the rural public space of the users, analyze the visual landscape preferences of the subjects for the rural public space, and provide further guidance and a basis for the landscape optimization of rural public spaces.

3.1. Design of Eye-Movement Experiment

According to the main types of public spaces in Huangshandian Village summarized in the previous analysis, we used the four typical public spaces (square space, street space, waterfront space, and public green space) as the spatial observation scenes for the eye-movement experiment. The research team individually led each of the 20 subjects to the four designated spatial sites and then started the formal experiment:

(1)

The researchers explained the procedure and precautions of the experiment to the subjects to ensure that they had a full understanding of the experimental methods and requirements.

(2)

To begin the eye-movement calibration work, after the calibration, the researchers instructed the subjects not to swing their heads, and they gave each subject a uniform instructional phrase: “The following is the official start of the experiment, I will take you to watch a public space scene in the village, will stay for a period of time, please watch carefully and try not to move position suddenly.”

(3)

The researchers separately took each of the 20 experimenters to the same position to watch the same public space scene. When the subject entered a relaxed state and started watching, the researchers pressed the timer to start recording the eye-movement data for 30 s. The researchers then asked the subjects to fill out a subjective satisfaction questionnaire based on their subjective perceptions of the viewing.

3.2. Participants and Apparatus

Through communication with the school and the local government, we believe that this study can be used for eye-tracking experiments and investigations without the approval of the Ethics Committee. Then, we recruited some tourists visiting the area as participants in the experiment. The study process was conducted anonymously to effectively protect the privacy of the subjects, and we respected and guaranteed the subject’s right to know and their right to make their own decisions about participating in the study. The final subjects were 20 local residents and tourists from Huangshandian Village, half of whom were male and half female. The subjects were required to have a naked-eye vision or corrected vision of more than 1.0, normal color vision and no visual disease. Before the experiment, all participants were informed of the research purpose, main procedures, etc.

The study equipment is a portable head-mounted eye-tracking device, Eye Tracking Core+ (ETC), which is capable of measuring eye movements at 120 Hz per second without restricting the subject’s movement and actions, with a foreground camera field of view of 81° horizontally and 60° vertically, and which records the experiment while the participant is observing a real scene. It was equipped with post-eye movement analysis software (Core Studio) installed on a PC with Nvidia GTX1060 graphics card configuration, 1920 × 1080 resolution, and a Windows 10 Pro 64-bit operating system. There are also 20 paper questionnaires, pencils, timers, etc.

3.3. Data Analysis

Since the eye-tracking experiment was conducted in a real environment, the subject’s visual field may have angular changes, so later, when processing the eye-movement data collected and output using the eye-tracking analysis software (Core Studio), each fixation point information in the experimental video is checked one by one with the location information of the pictures taken until all fixation points in the video are mapped to the pictures (no loss of time information). After all the experimental data were checked, the statistical software Excel and SPSS were used to analyze the data in three parts: (1) a statistical analysis of the oculomotor data of the different subjects. Output eye-movement index data cover a lot of information, and in this study, the main data indexes in the eye-movement experiment were related to the eye saccade and fixation. We selected a total of four indexes as the eye-movement-index data in this experiment: the Average Fixation Duration, Fixation Count, Average Saccade Amplitude, and Saccade Count (

Table 1. Selection of eye-movement indicators and their relevance.

Eye-Movement Indicators	Relevance
Average fixation duration (s)	Represents the average duration (in seconds) of all the selected gaze points in the experiment; the longer the average duration of the gaze points in each scene, the greater the subject’s interest and involvement in an element of the space [65].
Fixation count	The number of times a subject looks at an element per unit of time is an indicator that reflects the degree of the importance of the evaluation area; the more times a subject looks at the area, the higher the processing efficiency of the subject, and the visual information can quickly be used to form psychological information, representing the higher degree of attention or interest present in the subject [66].
Average saccade amplitude (px)	Refers to the average size of all the selected eye beat amplitudes in the experiment (in terms of the viewing angle); the larger the average eye beat amplitude, the more distinctive the characteristics of the picture, and the greater the range of information acquired [67].
Saccade count	Refers to the number of eye jumps per unit time, which is an indicator of the subject’s search behavior; the more eye jumps per unit time, the greater the search volume and less obvious the characteristics of the image [68].

). (2) an analysis of the visual preferences for the oculomotor landscape, including heat-map and path-map analyses; (3) an analysis of subjective satisfaction.

4. Results

4.1. Statistical Analysis of Eye-Movement Data

4.1.1. Analysis of Differences in Eye-Movement Indicators in Different Types of Public Spaces

We collected the study data and organized them. According to the data, the mean values of the subject’s eye-movement indexes for different public spaces differed. We tabulated the selected indexes of the average fixation duration, fixation count, saccade amplitude, and saccade count according to the different mean values (

Table 2. Mean values of eye-movement indicators in different types of public spaces.

Public-Space Category	Average Fixation Duration (s)		Fixation Count		Average Saccade Amplitude (px)		Saccade Count		Total Observation Time (s)
	M	SD	M	SD	M	SD	M	SD
Square space	0.28	0.06	75	15.89	123.91	24.78	74	15.89	30
Street space	0.41	0.10	62	17.06	132.29	42.50	61	17.06	30
Waterfront space	0.40	0.03	84	16.82	152.95	41.89	83	16.77	30
Public green space	0.38	0.11	91	22.85	179.66	18.55	90	22.85	30

) to facilitate further analysis.

Average fixation duration analysis: The average gaze durations of the four types of public spaces fluctuated between 0.28 s and 0.38 s (Figure 3a). The descending order is as follows: street space (0.41 s); waterfront space (0.40 s); public green space (0.38 s); square space (0.28 s). The longer the gaze duration, the more information is acquired by the subject when perceiving the spatial environment and scene with more effort. According to the experimental data, the subjects needed to find more gaze points to observe the street space, and they paid less attention to the square space.

Average saccade amplitude analysis: According to the results, there were differences in the mean values of the average eye jump amplitudes for the different types of public space landscapes (Figure 3c). The mean values in descending order are as follows: public green space (179.66); waterfront space (152.95); street space (132.29); square space (123.91). The larger the average saccade amplitude, the more distinctive the features of the pictures. The spatial features of the public green space were more distinctive, and the subjects could directly reach the target area, while the square space was lower than the other three types of spaces, which indicated that the information search range of the square space is larger and not distinctive enough.

Saccade count and fixation count analyses: From the results, the mean values of the numbers of the saccade count and fixation count for the four different types of public-space landscapes showed a positive correlation trend (Figure 3b,d, respectively), with floating range values between 74 and 91. Among them, the public green space had the highest number of eye bounces and gazes (90 and 91, respectively), while the square space had the lower numbers of saccade counts and fixation counts, which means that the subjects searched for more information on and were more interested in the landscape composition of the public green space. The information search for the square space was smaller, the importance of this space was lower, the scene characteristics were clear at a glance, and the landscape composition was not rich enough.

In summary, when the subjects observed different types of rural public space scenes, they were more interested in the public green space landscape than the other three spatial landscapes (in descending order: public green space, waterfront space, street space, and square space). In terms of feature distinctness, the square space was the most prominent, and the subjects searched for the most information on the public green space landscape. The most cognitive effort was made for the street space landscape. The subject’s visual behavioral information was influenced by the feature distinctness of the public space landscapes.

4.1.2. Analysis of Differences in Eye-Movement Indicators of Different Types of Public Space Landscapes Using Different User Groups

Based on the classification of the four types of public spaces in Huangshandian Village, we calculated the average fixation duration, fixation count, saccade amplitude, and saccade count, and we summarized them based on the screened eye-movement-index values (

Table 3. Comparison of eye movement indicators of different users in different public spaces.

Public-Space Category	User Group	Average Fixation Duration (s)		Fixation Count		Average Saccade Amplitude (px)		Saccade Count		Total Observation Time (s)
		M	SD	M	SD	M	SD	M	SD
Square space	Local residents	0.27	0.05	46	11.47	122.10	25.49	74	14.80	30
	Visitors	0.29	0.06	76	17.79	125.72	25.46	75	17.80	30
Street space	Local residents	0.41	0.13	58	21.40	127.99	50.74	57	21.40	30
	Visitors	0.41	0.08	65	11.66	136.59	34.95	64	11.66	30
Waterfront space	Local residents	0.38	0.03	92	12.76	126.39	22.00	91	13.60	30
	Visitors	0.42	0.03	77	17.60	179.50	40.66	76	16.84	30
Public Green space	Local residents	0.29	0.03	102	25.84	171.21	20.25	101	25.84	30
	Visitors	0.47	0.07	78	8.91	189.16	11.11	77	8.91	30

). The suitability of this test was verified through a homogeneity test for all four spatial categories of eye movement index values before analyzing the comparison among variables. ANOVA was used to verify whether there was a significant difference between the eye-movement indicators of the two user groups in different public spaces. The differences and effect sizes were further analyzed for the data with significance. The Cohen’s $f$ [69] and ${\eta }^2=z^2/N$ [70] were adopted to calculate the effect size (Small: >0.1 and <0.25; Medium: >0.25 and <0.4; Large: >0.4). We observed substantial differences in the subject’s eye-movement indicators for different types of public spaces and users, and the comparative study between the two had some relevance.

Average fixation duration: The average fixation duration of four kinds of spaces passed the homogeneity test (p = 0.397, p = 0.171, p = 0.245, p = 0.147), which verified the applicability of this test. As seen in Figure 4a, the average fixation durations of the visitors were all higher than the local residents, which indicates that the visitors worked harder than the local residents to perceive the public spaces. The longest average fixation duration was 0.41 s for the street space, and the shortest was 0.27 s for the plaza space, figures which were demonstrated by the local residents. The longest average fixation duration was 0.47 s for the public green space, and the shortest was 0.29 s for the plaza space, figures which were demonstrated by the tourists. The shortest average fixation duration was 0.29 s. According to the analysis of variance, the average fixation duration of the local residents and visitors in the waterfront space (p < 0.05, η² = 0.315) and the public green space (p < 0.01, η² = 0.746) were significantly different, and the value expressed by the latter was significantly greater than that of the former, and the degree of difference was different to a large extent, however it was not suitable for the square space and the street space.

Fixation count: The fixation count of four kinds of spaces passed the homogeneity test (p = 0.242, p = 0.176, p = 0.514, p = 0.178), which verified the applicability of this test. According to Figure 4b, the average fixation count for the different types of public space landscapes varied widely among the different users, with an overall fluctuation range from 40 to 100 times. According to the analysis of variance, it can be seen that there are significant differences between local residents and visitors in the times of staring at square space (p < 0.05, η² = 0.496), waterfront space (p < 0.05, η² = 0.209), and public green space (p < 0.05, η² = 0.294), and there are significant differences. From the average value, we can see that both the local residents and visitors had the highest average fixation counts for the public green space, while the lowest average fixation counts were for the square space, which indicates that both the visitors and local residents were more interested in the landscape of the public green space, and that the landscape distribution of the public green space attracted people’s attention to a certain extent. Some information areas of the public green space were more important to the local residents than the tourists.

Average saccade amplitude: Before analyzing the variables of the two groups of users, the average saccade amplitude of the four kinds of spaces all passed the homogeneity test (p = 0.72, p = 0.572, p = 0.037, p = 0.071). According to Figure 4c, overall, the local residents’ average saccade amplitudes for the square space, street space, waterfront space, and public green space were both lower than those of the visitors; in addition, the average saccade amplitude in waterfront space (p < 0.05, η² = 0.423) and public green space (p < 0.05, η² = 0.247) between the two groups was significantly different, which indicates that the visitors had stronger performances in acquiring the landscape information compared with the local residents. In terms of the public space types alone, both the local residents and visitors had higher saccade amplitudes for the waterfront space and public green space landscapes, however the lowest for square space, and the difference between the two kinds of users was small.

Saccade count: The saccade count of four kinds of spaces passed the homogeneity test (p = 0.504, p = 0.176, p = 0.736, p = 0.178), which verified the applicability of this test. We observed little difference in the saccade counts of the visitors compared with the local residents for each type of public space (Figure 4d). However, from the results of ANOVA, it could be seen that the number of saccade counts of local residents and visitors for the waterfront space (p < 0.05, η² = 0.211) and the public green space (p < 0.05, η² = 0.294) presented significant differences. According to the figure, the average saccade counts of the local residents for the waterfront space and public green space were slightly higher than those of the visitors (by 91 and 101 times, respectively), which indicates that the local residents searched for information on the waterfront space and public green space landscapes more frequently than the visitors.

4.2. Visual Preference Analysis of Eye Movement in Landscape

4.2.1. Eye-Movement Hot-Spot-View Analysis

The hot-spot view can efficiently and intuitively display the visual attention areas of multiple subjects, and the visual attention areas of different types of public space landscapes can be generated after the experimental results of the subjects are aggregated and overlaid for analysis. Red indicates the key areas of visual attention, yellow and green indicate the areas with less visual attention, and green indicates the areas with the least visible attention (a gradual weakening from red to yellow to green).

Square space: The visual range of the subjects for the square space was more likely to focus on the specific height differences of the landscape structures and artifacts and the traditional features in the center of the scene because of the open view (Figure 5a); therefore, the residential buildings and landscape artifacts (lanterns) attracted more of the subject’s attention, which then spread to the surrounding green landscape and architectural facilities, and the subjects were more inclined to watch the vanishing point in the scene.

Street space: The streets in Huangshandian Village are mostly of a north–south longitude, with granite or stone pavement, and decorated with groundcover greenery. The buildings on both sides of the street have the obvious characteristics of the mountain dwellings in western Beijing. The main focus of the subjects when observing the street space was towards the disappearance of the sight line of the street (Figure 5b), and it then spread to the building structures (doors, windows, roofs) and green landscape on both sides.

Waterfront space: The subject’s visual range for the waterfront space was mostly focused on the landscape facilities, vegetation greenery, and buildings at and around the end of the water system (Figure 5c), while the gaze points near the water scene and for the water scene itself were relatively less focused, which indicates that, in the waterfront space scene with open sight lines, the distant and background landscape more strongly attracted the subject’s attention to a certain extent.

Public green space: The subject’s observation points for the public green space in Huangshandian Village tended to be on the old theater, with open sight lines, and on the distant residential buildings with a green background, which indicates that the subject’s eyes tended to fall on the vegetation landscape and favorable landscape buildings, with stronger depths of field and richer landscape elements (Figure 5d).

4.2.2. Eye-Movement Trajectory Map Analysis

The analysis of the number sequence of the oculomotor trajectory can objectively reflect the trajectory and sequential route of the subject’s eye activity. The direct line between two gaze points represents the sweeping behavior of the subject from one gaze point to another, and the size of the number-sequence circle in the diagram represents the length of the subject’s gaze time at a certain point, which more specifically reflects the subject’s elements of interest [2].

By compiling and analyzing the eye-tracking-path trajectory map (Figure 6a–d) of all the subjects observing the four types of public space landscapes in Huangshandian Village, we concluded the following: (1) in the scenes with open lines of sight or strong depths of field, such as the plaza space and street space, the subject’s observation points first focus on the disappearance of the line of sight (i.e., the extinction point in the picture), and then on the landscape structure (residential buildings and facilities) and plant groups; (2) the subjects were more inclined to focus on the branching points of the trees when observing the plant groups, and to move around with the sparseness of the vegetation, and then to focus on distant peaks or the proximal groundcover; (3) the subjects observed the waterfront space by focusing their eyes on the riparian landscape and repeatedly focusing their observations.

4.3. Visual Preference Analysis of Eye Movement in Landscape

In order to verify whether there was some connection and pattern between the subject’s subjective evaluation and the eye-movement experiment data, after the subjects finished the eye-movement experiment, the researchers asked them to evaluate the landscape satisfaction for each type of public space scene with a score from 1–5. We used Excel to calculate the average value of satisfaction for each type of public space landscape. According to Figure 7a, the subject’s landscape satisfaction ratings for the four types of public spaces, from highest to lowest, were as follows: street space, waterfront space, public green space, and square space. Overall, both the visitors and local residents rated the street space the highest, with an average score of 3.98. The lowest score was given to the square space, with an average score of 2.69 (Figure 7b). The street space and waterfront space, with higher scores, have well-organized landscape resources, clear structures, distinctive landscape color pairings, and higher ornamental value, and thus the subjects preferred them.

4.4. Subjective and Objective Correlation Analyses

We applied the statistical analysis software SPSS software (version 11.01; SPSS Inc., Chicago, IL, USA) to quantitatively analyze the correlation mechanism between the subjective and objective correlation analyses. According to the relationship between the eye-movement data and subjective satisfaction variables in the four types of public spaces (

Table 4. Correlation between eye movement data and subjective evaluation.

Type of Space			Average Fixation Duration	Fixation Count	Average Saccade Amplitude	Saccade Count	Satisfaction Evaluation
Square space	Average fixation duration	Pearson correlation	1
	Fixation count		−0.202	1
	Average saccade amplitude		0.110	−0.219	1
	Saccade count		−0.202	1.000 **	−0.219	1
	Satisfaction evaluation		0.610 **	−0.231	0.670 **	−0.231	1
Street space	Average fixation duration	Pearson correlation	1
	Fixation count		−0.507 *	1
	Average saccade amplitude		−0.329	−0.117	1
	Saccade count		−0.507 *	1.000 **	−0.117	1
	Satisfaction evaluation		0.499 *	−0.469 *	−0.236	−0.469 *	1
Waterfront space	Average fixation duration	Pearson correlation	1
	Fixation count		−0.752 **	1
	Average saccade amplitude		−0.260	0.011	1
	Saccade count		−0.752 **	1.000 **	0.011	1
	Satisfaction evaluation		0.604 **	−0.460 *	−0.104	−0.460 *	1
Public green space	Average fixation duration	Pearson correlation	1
	Fixation count		−0.525 *	1
	Average saccade amplitude		0.175	0.062	1
	Saccade count		−0.525 *	1.000 **	0.062	1
	Satisfaction evaluation		0.665 **	−0.469	0.514 *	−0.469	1

* Correlation is significant at 0.05 level (two-tailed). **. Correlation is significant at 0.01 level (two-tailed).

), subjective satisfaction in all four types of spaces was significantly correlated with the average fixation duration. Among them, the subjective satisfaction evaluation of the square space and public green space had a significant positive correlation with the average saccade amplitude, while the subjective satisfaction evaluation of the street space and waterfront space showed a significant negative correlation with the fixation and saccade counts.

5. Discussion

5.1. Correlation Analysis of Eye-Tracking Data and Subjective Satisfaction for Public Space

Combining subjective evaluations with eye-tracking data to quantify the effects of these two variables can provide a more comprehensive and objective assessment of the landscape quality. We can not only evaluate the importance of each public space according to the subjective-evaluation scale, but also analyze heat maps, trajectory maps, and various eye-movement indicators to explore the subject’s subconscious and aesthetic patterns. Thus, we can analyze the composition and proportions of the public space landscape elements and the matching law of each element to achieve guidance for enhancing the public space landscape design.

According to the correlation analysis of the eye-movement data and subjective satisfaction [71,72,73,74,75,76,77,78,79] (Table 3), there is a close relationship between them: the average fixation duration had a significant positive correlation with all four types of public spaces, and a negative correlation with the fixation count and saccade count, which suggests that the more attention people pay to the public space, or the less they search for elements in the space, the more likely they are to make higher evaluations of the spatial landscape. The lower the number of the fixation count and saccade count per unit of time, the higher the ornamental value of the landscape and degree of people’s interest. According to the analysis of variables, the subjective evaluation of both the street space and waterfront space was significantly negatively correlated with the fixation count and saccade count, which indicated that the street space and waterfront space have more obvious scene characteristics and are more likely to attract people’s interest compared with the other two types of spaces. The greater the average saccade amplitude, the more distinctive the features of the scene, and the more intense the performance of the subject when acquiring the landscape information. According to the subjective evaluations of the square space and public green space, there was a significant positive correlation with the average saccade amplitude, and the Pearson correlation coefficient for the square space was higher than that of the public green space, whereas in the previous analysis of the eye-movement data, the average saccade-amplitude values of the subjects in the square space were the smallest compared with other types of spatial spaces, which is due to the following: first, the subjective evaluation of the square space by the subjects was more varied; second, the average saccade amplitude in the eye-movement data analysis was small, indicating that the spatial composition of the open square space was more distinct and people could reach the target area at a glance. However, in terms of subjective evaluation, some subjects may not think that there are too many ornamental elements in the open square space scene.

The use of eye tracking in this study revealed that the perception of the public space landscape is influenced by various visual features. According to the eye-movement heat and trajectory maps above, the subjects paid more attention to the green landscape in the areas of public space interest, and favorable landscape structures, such as buildings and vignettes. Plant materials, buildings, pavements, and water features are the main elements of landscape architectural design [80], and they have different effects on the landscape evaluation. In previous studies, researchers found that natural elements and artificial elements have different emotional impacts on people’s perceptions, with natural elements (such as vegetation and water bodies) ranking positively, and artificial elements (such as cars, parking, and billboards) ranking negatively [81,82]. Moreover, in previous environmental psychology studies, researchers found that people spent more time and had fewer fixation counts when observing natural scenes compared with urban scenes; thus, the subjects required less eye movement when viewing natural environments [83,84,85]. Noland et al. found that the respondents spent the most time on natural elements, such as trees, with trees ranking highest in the element assessment, and artificial elements, such as paths, ranking lower [86]. Amati et al. found that the eye is attracted to artificial objects, such as light poles, trash cans, and signage, which have sharp edges and contrast in color with the surrounding vegetation or environmental background [38]. These results are all consistent with the findings of this study; thus, eye tracking is associated with both vegetation greenery and hardscape elements. The results contribute to the understanding of people’s emotional perceptions of the landscape during the evaluation process.

In conclusion, the eye-movement data can indirectly reflect the psychological changes in the subjects through the analysis of the relevant software, which is not analyzed using the subjective valence method. The combination of eye-movement analysis and subjective evaluation can reflect the psychological feelings of the subjects towards each element in the landscape pictures, which has strong guiding significance for landscape design.

5.2. Analysis of Differences in Landscape Preferences of Different User Groups

The public space users explored in this paper were local residents and visitors. According to the above experimental data results and questionnaire analysis, the two groups had different functional orientations and subjective judgments when making spatial observations, which we analyzed for the following reasons. In terms of the differences between the two experimental data, the tourist experimental group showed a stronger performance in terms of the perception of the landscape and information acquisition compared with the local residents. The countryside is a new scene for visitors, which requires them to make an effort to explore the important elements in the public spaces to seek their points of interest [87]. For the waterfront spaces and public green spaces, the lines of sight are more dispersed, and the visitors paid more attention to the large spaces, appreciating the overall atmosphere and focusing on feeling the overall landscape beauty of the countryside. Chen et al. found that visitors prefer landscapes that maintain the naturalness and rusticity of the rural landscape. The square space and street space mainly host daily leisure and entertainment functions, and visitors are less involved in these two types of sites and pay less attention to them [88]. The visual information is more concentrated and focused on landscape elements that easily stimulate people’s interest, such as lanterns, vignettes, flower boxes, notice boards, etc. The contrast between the white spaces and the details of the two types of spaces also creates more exploration opportunities for visitors [89].

The aborigine group more easily searched for information on the landscape in the square space and street space than the tourists, which is because these two types of spaces are their daily life spaces and they are familiar with their landscapes [90]. The landscape elements in the street space, such as the gables of the buildings and changing landscape walls, were more likely to attract the subject’s attention. On the contrary, the local residents searched for information on the waterfront spaces and public green spaces more than the tourists. The main reason is, for the local residents, the waterfront space has more landscape elements, and the public green space was previously a theater for village activities. The villagers generate more spontaneous activities on these two types of land, which the local residents carry more memories of and have certain emotional attachments toward [91].

From the commonality of the eye-tracking experimental data, both the local residents and tourists showed some attention to the public green space, waterfront space, and street space, and neither group was concerned about the square space. These commonalities reflect the fact that green space and waterfront space in the countryside are more attractive to people. The unique scenery and rustic nature of the countryside cannot be felt by the tourists in the city, and these two spaces are also the priority choice of the local residents for recreation after their daily work and leisure; thus, both groups showed a strong interest in the public green space. The space of the square is closer to the hard space of the city and lacks rural-specific content; thus, tourists are less interested in the space of the square, while a large number of hard white areas in the square causes the local residents to have little interest in it [92]. In terms of the subjective evaluations, we observed consistency among both groups, which indicates that both groups were more interested in the street space and less concerned with the square space in terms of subjective aesthetics. Combined with the eye-tracking data, the results of both analyses coincide to some extent. In summary, tourists are more concerned with the recreation service function of rural public spaces and the appearance of the countryside [93], while local residents are more concerned with the functional utility of the site and the suitability of the human living environment [94]. At the same time, there is a certain commonality of landscape preferences between both groups. The spaces with higher interest points should play to their strengths, and the square space with lower interest points should avoid disadvantages such as large white spaces and the integration of vernacular features.

5.3. Optimization Strategies for Different Types of Public-Space Landscapes

(1)

Square space: According to the above study, the current status of the square space is relatively empty, and compared with the other types of spaces, the villagers and tourists had shorter gaze times and paid less attention to the space, which means that the square is not distinctive enough. Therefore, in this study, we aimed to optimize the function and landscape of the spatial landscape of the square to reduce the sense of emptiness and enrich the information on the site. Considering the visual appeal of the vanishing point to the crowd, favorable landscape structures can be configured at the vanishing point of the scene. The control of its color and form would create a greater contrast with the existing pavement and surrounding scenes of the square [95]. In addition, considering the interest of the crowd in height differences in the experimental results, the square renovation should make full use of the facade elements in the space, such as the building walls, plant forest canopy line, and distant mountains as the mid-view and distant view of the scene, and create layers, beauty, and rhythm together with the square foreground.

(2)

Street space: The street space achieved higher scores in both the subjective and objective evaluations, and the crowd had more attention points for the observation of the street space and paid more attention to it. Therefore, as a unique street landscape in western Beijing, the key point of renovation should be to protect and enhance the current elements while using the spatial forms for extension and refinement. On the one hand, the space is segmented by making full use of the guiding nature of the street space, and the turning space is appropriately increased in the form of an enclosure based on maintaining the depth. The end of the space is decorated with important landscape nodes, such as characteristic sculptures [96] and solitary plants, in an attempt to create a sense of a “Different view of stepping” [97]. On the other hand, due to the inward view of the street space, extra attention should be paid to the details of the walls and the elements on them, and the original materials should be used for repair, protection, and renewal to fully integrate them with the surrounding environment [98].

(3)

Waterfront space: In view of the tourists and villagers the current waterfront space of Huangshandian Village is relatively average. The enhancement of the landscape of the Huangshandian Village waterfront space should be based on maintaining good water quality to increase the richness of the landscape while reflecting its characteristics [99]. As the crowd’s line of sight often stayed near the barge, the design of the water landscape should be focused on enriching the barge form to a natural meandering berm, as well as on arranging favorable water landscape architecture [100] on the premise of ensuring the water quality so that the open space, semi-open space, and closed space have a balanced coexistence in the waterfront space, which draws close-up attention to the water scene itself, with attention originally focused on the distance and background [101,102].

(4)

Public green space: According to the data, public green space is highly valued and distinct, but it is difficult to identify. Thus, the core of the public-green-space landscape enhancement lies in enhancing the depth of field of the space, increasing the landscape elements, and highlighting the spatial distinctiveness and identifiability. The entrance to the public green space can be marked to enhance its distinctiveness. From previous studies, we know that to achieve the effect of sharpness and security, the overall spatial color and texture of public green spaces should form a certain contrast with the surrounding environment, while the internal color of the space should be soft and comfortable to fully adapt to the psychology of tourists. In addition, the landscape difference of different green areas should be enhanced with the help of vegetation landscape and especially landscape architecture, which should be used as much as possible to introduce a vision into the space on the basis of perfect restoration to enhance the sense of the spatial hierarchy.

In addition, the plant landscape plays an important part in the beauty of the space, while color variation is an attractive visual element in landscape scenes [103]. For the spatial landscape enhancement of Huangshandian Village, planting should pursue color-phase change, the rich forest canopy line level, and sparse space while using colorful foliage plants, seasonal flowers, fragrant flowers [104], and recreational plants [105] to create a plant space with rich colors that appeal to all five senses, and should aim to pursue the practicality and beauty of the landscape on the basis of ecology [106]. For the optimization of the square-space landscape, plants play a role in the space division at the same time. Beautifully shaped trees dominate, with an appropriate increase in brightly colored native flowers in the understory. The street-space plants should be as graceful as possible while directing attention to both sides of the street in the form of three-dimensional greening. The waterfront space fully caters to the visual range, and the vegetation at the end of the water system and around the barge is enriched in layers to enhance the effect of the plant landscape from a distance. The height and color differences of the public-green-space plants should be large to fully attract attention.

6. Conclusions

In this study, we combined survey data and eye-tracking technology to identify the landscape preferences of different people for different spaces based on a quantitative evaluation of the eye-tracking data and subjective satisfaction of the different user groups in the rural public space of Huangshandian Village in Beijing, China. The results show that eye-tracking technology can provide an effective measurement method for the landscape preference of different people in different spaces and can also provide a direct quantitative evaluation and empirical support for the landscape preference theory, and we accordingly propose specific strategies for its landscape optimization. The main conclusions and implications of this study are as follows:

(1)

The subject’s eye-movement data differed when observing different types of rural public space landscapes, and there were also differences in the eye-movement data for the same spatial scene among different types of subjects. The subjects cognitively worked hardest on the spatial landscape of streets and alleys, and their visual information was mostly focused on the places where the focal points of vision were prominent. Both the visitors and local residents were more interested in the public green spaces, which means that the landscape features of the public green spaces are more obvious. The local residents searched for information on the public green space and waterfront space landscapes more than the tourists, and the information searches were more frequent. The subjects were interested in four types of spaces, which were, in descending order: the public green space, waterfront space, street space, and square space. In terms of the subjective evaluations, the subjects had the lowest subjective evaluations and satisfaction with the square space, and the highest satisfaction with the street space, which indicates there are similarities between the objective eye-movement-analysis results and the subjective-evaluation results.

(2)

The overall results of the eye-movement-hotspot view and trajectory map showed that the subject’s visual information was biased toward the center of the spaces and the points where their vision disappeared. The visitors were more concerned with the function of the rural public space recreation services and the appearance of the countryside, while the local residents were more concerned with the functional practicality of the site and the suitability of the human living environment. The architecture of the square space and the undulating structures are the main visual attraction elements. The details of the street space, and especially the doors, windows, and structures on both sides, are the main factors that attract people’s vision. For spaces with long depths and open sight lines, the subject’s visual information was mostly focused on the green landscape and architectural elements. For the waterfront space, the visual information was mostly biased towards the distant barge, skyline, plant groups, and other landscape aspects with multielement integration.

(3)

This study is important for understanding users’ landscape perceptions and guiding landscape design. It is hoped that the results of this study will help rural planners as well as designers to more clearly understand the relationship between landscape preference elements and users’ subjective perceptions, and to fully understand the activity needs and landscape preferences of visitors and villagers within the public space landscape. Moreover, it is hoped that the eye-tracking technology will be incorporated into the planning and management practice of rural landscape, and that the research methods such as landscape visual information and subjective evaluation will be used to reasonably organize each landscape element for different spatial types in order to precisely improve the spatial landscape characteristics.

In addition, this paper conducted an eye-tracking study on four main types of public spaces in Huangshandian Village, and it may not cover all the types of rural public spaces; thus, the comprehensiveness of the study needs further discussion and research. Although local residents and visitors are very meaningful as experimental subjects, however, due to the limited size of experimental samples, the analysis of eye movement data is not profound, and the preliminary study may have some limitations. The sample size and range of the population background can be expanded in future related studies to better reflect the realistic landscape judgments of visitors and villagers as much as possible, and to draw more generalized research conclusions. At the same time, in the future, we can consider asking the subjects to test the new design scheme again after the landscape environment transformation and upgrading, exploring the data differences before and after the strategy, which is the focus of our subsequent research.