Research on Quantitative Analysis Methods for the Spatial Characteristics of Traditional Villages Based on Three-Dimensional Point Cloud Data: A Case Study of Liukeng Village, Jiangxi, China

Abstract

Traditional villages are important carriers of cultural heritage, and the quantitative study of their spatial characteristics is an important approach to their preservation. However, the rapid extraction, statistics, and estimation of the rich spatial characteristic indicators in these villages have become bottlenecks in traditional village research. This paper employs UAV (unmanned aerial vehicle) and handheld laser scanners to acquire three-dimensional point cloud data and construct a spatial feature three-dimensional calculation workflow of “field data collection—data processing—data analysis and application”, which enables the rapid acquisition, processing, and analysis of three-dimensional village data. Typical case studies are conducted in Liukeng Village, China, focusing on the quantification of village spatial characteristics at three levels: topography, streets and alleys, and individual buildings, as well as comparative studies of multiple villages across different regions. The quantification of three-dimensional data reveals the regularity of village spatial characteristics and uncovers the spatial wisdom embedded in the site selection and spatial structure of traditional villages. This paper establishes a complete technical route for the quantitative analysis of villages, deepens public understanding of the diverse value of traditional villages, and provides technical support for research and practice in related fields.

1. Introduction

Traditional villages are important carriers of cultural heritage and are the largest active inheritance carriers of agricultural civilization in the world. They are also comprehensive entities that integrate both material and intangible cultural heritage [1]. In January 2017, the Chinese Traditional Village Protection Project was incorporated into the national strategy [2], aiming to achieve full coverage and comprehensive inclusion of village spatial elements, thereby strengthening cultural confidence. Since 2012, the Ministry of Housing and Urban-Rural Development, along with related departments, conducted identification of traditional village resources, and a total of 8155 villages with important conservation value were included in the Chinese Traditional Village List [3]. As an important carrier of Chinese traditional culture, the protection and development of traditional villages have received increasing attention.

However, with the rapid advancement of urbanization and the imbalance of urban-rural development, the appearance of traditional villages in China has been severely damaged [4]. For example, market-driven tourism development has led to the transformation of village blocks, and to cope with increased traffic and people, some houses have been demolished, and roads widened, damaging the spatial texture of traditional villages [5]. Current village planning neglects the overall value and interrelationships of village spatial layouts, resulting in cold and monotonous “city-like” new village forms, which disrupt the organic and rich spatial characteristics of traditional villages and sever the traditional context of village spatial form.

The spatial characteristics of traditional villages reflect regional climate features, lifestyles, and traditional customs [6]. They also express Chinese agricultural culture and traditional rural construction ideas and concepts. The inherent order of traditional village space is the essential “rurality” that distinguishes rural areas from cities [7]. Studying the regularities of spatial characteristics in traditional villages can help trace back to the original appearance of these villages, reveal the construction wisdom of ancient people, and is an important approach to the continuation of traditional village preservation. The statistical analysis of spatial feature indicators at the levels of individual dwellings, street spaces, and topographic environments is the basic work shared in various studies on traditional villages [8,9]. In previous studies [10,11], however, due to the lack of effective feature extraction and quantification methods, the rural wisdom is difficult to accurately reveal, and the interaction between village space and rural environment is difficult to visualize and illustrate. In addition, traditional villages in China are scattered in different geographical environments, and they have the characteristics of “self-organization” and “self-growth”. Therefore, they have spatial complexity at multiple levels, including topography, street patterns, and courtyard units. Regardless of the perspective of spatial defense, ecological livability, or regional characteristics, the study of traditional villages requires corresponding spatial feature indicators for quantitative description or analysis, and through quantification, to grasp the regularity of village characteristics from a macro perspective.

Currently, there are over 6000 traditional villages at the national level in China, with even more at the provincial level. It is time-consuming and labor-intensive to implement “field measurement-manual drawing-data extraction-cross-regional comparison” for a large number of villages. Therefore, it is imperative for research on traditional villages to explore suitable way to quickly obtain spatial indicators for a large number of settlements, and achieve a collective understanding of village spatial characteristics and cross-regional comparison. In recent years, the rapid acquisition and extensive application of 3D point cloud data have introduced new technological methods for quantitatively analyzing village spatial characteristics [12,13,14]. Through the use of unmanned aerial photography and handheld laser scanners on-site [15,16,17,18], the rapid acquisition of three-dimensional data can be achieved, especially through direct calculations based on 3D point cloud data, supplemented by other spatial analysis methods. This can semi-automatically extract quantitative indicators of village spatial characteristics from 3D point clouds in a short period [19]. In addition to point cloud direct calculations, data vectorization processing supported by Geographic Information System (GIS) [20], spatial syntax [21,22,23], and other analytical theories or methods are employed to comprehensively determine the optimal combination of calculation methods for spatial indicators at various levels.

This paper utilizes 3D point cloud data to develop a complete technical route for the quantitative analysis of spatial characteristics in traditional villages. The route includes “field data collection—data processing—data analysis and application”, enabling the rapid acquisition, processing, and analysis of three-dimensional village data. The paper addresses the following problems: (1) In terms of field data collection methods, how to efficiently and accurately obtain high-precision 3D point cloud models considering the dense buildings and narrow streets characteristic of traditional villages; (2) In terms of data processing methods, how to process point cloud data and classify various elements such as vegetation, roofs, walls, and road networks; and (3) In terms of data analysis and application methods, how to choose relevant quantitative indicators to describe the spatial characteristics of village morphology at the levels of topography, streets and alleys, and individual buildings, thereby enabling the efficient extraction of diverse spatial feature indicators.

Taking Liukeng Village in Fuzhou City, Jiangxi Province, China as an example, the paper conducts a typical case study on the quantification of village spatial characteristics at the levels of topography, street spaces, and individual buildings. Furthermore, based on the example study mentioned above, cross-regional comparisons of spatial characteristics are conducted for different types of residential forms, thereby verifying the rationality of the original regional divisions and highlighting the uniqueness of each region through visual results and quantitative data. This paper adopts a digital technology application approach to expand the study methods of traditional village spatial characteristic analysis, providing technical support for related fields.

2. Literature Review

2.1. Research on Spatial Characteristics of Traditional Villages

Research on the spatial characteristics of traditional villages includes the study of the compositional elements of village spatial morphology [24,25], morphometric methods, the evolution of influencing factors [26,27], the evolution of driving forces [22], and research on evolution mechanisms [28,29]. Fractal theory [30,31], boundary theory [32], and Euclidean geometry [33] have also been introduced into the field of village research, with a comprehensive explanation of their conceptual significance, utility value, and computational methods. Various other theories have also been introduced into the field of village research, with their conceptual significance, utility value, and computational methods being comprehensively explained [34]. In the multidimensional spatial analysis of village scale, models based on Space Syntax [35,36] and models for spatial parameterization and reconstruction based on the CityEngine platform [37] have been developed. For larger-scale village systems, Lin, G., et al. have attempted to use gravity models to measure spatial interaction and employ hierarchical analysis–geographic information system composite models to assist with spatial reconstruction [38].

In terms of quantitative analysis of spatial characteristics in traditional villages, there are still the following technical challenges: (1) Low data precision and a lack of high-precision centimeter-level data. Previous studies have primarily relied on satellite digital elevation models to support macro-scale spatial analysis, but have been limited by approximately 30-m precision, making it difficult to analyze complex or subtle terrain changes and accurately record village streets and buildings. This limitation hinders micro-scale village research. Additionally, these research directions have strict demands for spatial data, leading to a lack of micro-terrain and high-precision data resources. For example, a previous study [39] compared the terrain undulations and slope directions in the vicinity of traditional villages in Guizhou Province but could only provide approximate values for slope and slope direction without reflecting the detailed changes in micro-terrain, making it difficult to compare and analyze the various spatial characteristics of multiple villages. (2) Emphasis on two-dimensional (2D) data with a lack of three-dimensional (3D) spatial data. Previous research by Pu Xincheng et al. [32] primarily relied on computer-aided design (CAD) programming and the use of 2D village plans to complete spatial calculations, which is time-consuming and limited to two-dimensional scales. With the availability of 3D computational software and suitable 3D calculation methods for villages, various spatial characteristics can now be calculated quickly and semi-automatically [40,41]. This is a key requirement for many scholars in this field. Therefore, the current method of programming calculations based on 2D line drawings will eventually be upgraded to new methods based on 3D point cloud calculations, which will greatly increase calculation speed and result accuracy. (3) The application of data is not comprehensive, and it is difficult to accurately describe the spatial pattern of villages by analyzing a single indicator. Current research lacks comprehensive methods to express various aspects of the studied objects, merely describing spatial characteristics within a single dimension or locally. The traditional quantification approach focuses on a specific feature, which is limited in its ability to comprehensively cover all aspects of the morphological features and lacks persuasiveness, making it difficult to achieve a comprehensive evaluation of the characteristics of the village. Village spatial forms possess highly complex characteristics, making it difficult to describe them using a single indicator.

2.2. New Developments in the Study of Spatial Characteristics in Traditional Villages Brought by 3D Point Cloud Data

3D point cloud data is a collection of data composed of a large number of discrete three-dimensional points, containing information about the position, shape, color, and other attributes of the objects being described [38]. 3D point cloud data provides a realistic, three-dimensional, and temporal representation of traditional village spaces, breaking free from the constraints of traditional two-dimensional information representation. It enables the intuitive display, analysis, and application of various elements of villages and their natural resources, shifting from planar to three-dimensional, from abstract to concrete, and from static to dynamic. In the context of digital China construction and rural revitalization, the construction and development of rural areas are becoming digitalized. 3D point cloud data provides digital technology support for village protection, promotes the digitization, internetization, and sharing of information, and contributes to the study, preservation, and planning of traditional villages.

The acquisition of 3D point cloud data primarily includes low-altitude UAV aerial photography and three-dimensional laser scanners [18]. Among these, low-altitude aerial photography using multi-rotor drones has the greatest advantage in data acquisition for villages. It has the fastest operation speed and can capture large areas in a short period. Aerial photography is less affected by factors such as mountainous terrain and complex building shapes, and continuous downward photography along flight paths can capture more surface images with higher overlap. Considering the characteristics of narrow streets and small building spacing in villages, three-dimensional laser scanners can supplement more accurate ground data, such as the space under the eaves of houses, courtyard spaces, spaces under trees, and indoor spaces of residential buildings.

Currently, 3D point cloud data has been widely applied in surveying and mapping, geographic information systems, architectural design, cultural heritage preservation, industrial manufacturing, and other fields. It can be used for digital modeling of buildings, three-dimensional scanning of cultural relics, measurement of terrain and landforms, and various other applications. For example, tilt photography technology is applied to the upgrading of old buildings, solving the problems of lack of as-built drawings [42], low efficiency, and poor visualization effects of traditional measurement methods; the use of low-altitude UAV aerial photography and image processing technology for preliminary field observation [43]; and the use of low-altitude UAV aerial photography technology for the protection of traditional settlements [44].

In conclusion, 3D point cloud data provides accurate information and new analytical techniques for studying the spatial characteristics of traditional villages. It has broad application prospects in traditional village studies, providing effective support and services for village protection, planning, cultural heritage preservation, and tourism. Through the quantification of three-dimensional data, researchers can conduct in-depth analyses, understand the development trends and changes of villages, and provide recommendations for their protection and sustainable development. In the context of big data and smart cities, the protection of traditional settlements urgently requires high-tech digital support to lead the industry’s development and promote the digitization of information.

3. Materials and Methods

3.1. Study Area

This study selects Liukeng Village in Le’an County, Fuzhou City, Jiangxi Province, China, as the research area (Figure 1). The village spans an area of 3.61 square kilometers and is located at a longitude of 115.771697 and a latitude of 27.265745. The average elevation of the village is 427 m, and the region is characterized by a subtropical monsoon climate. Liukeng Village was established during the Shengyuan period of the Southern Tang Dynasty in the Five Dynasties period and initially belonged to Yongfeng County in Jizhou. It later became part of Le’an County in Fuzhou during the Southern Song Dynasty. The village has a history of nearly a thousand years and has been listed in the first batch of the Chinese Traditional Village List and the first batch of Historical and Cultural Famous Villages.

First, the location of Liukeng Village fully embodies the traditional Chinese Feng Shui concept. The village is situated on a “peninsula” surrounded by water on three sides and backed by a mountain on the west side. Wujian river passes through the village, representing a typical example of traditional village water management system planning and design, and demonstrating the wisdom and strategies applicable to micro-terrain environments. Second, the street space of Liukeng Village was planned and built by Dong Sui in the mid-to-late Ming Dynasty, and the street form has been preserved intact. Third, the individual buildings in Liukeng Village represent traditional Chinese “courtyard-style” residential structures. The village preserves more than 260 traditional architectural structures from the Ming and Qing Dynasty, including 19 Ming Dynasty buildings, 59 archways and pavilions, and a variety of courtyard forms in residential buildings. In summary, Liukeng Village is representative in terms of its location, street space patterns, and individual building forms among traditional villages in China. This paper takes Liukeng Village as the research object, utilizing 3D point cloud data to analyze the village’s topographic environment, street spaces, and individual buildings at three levels, achieving the quantification and visualization of traditional ancient construction wisdom in China.

3.2. Methods

This study constructs a technical route for the quantification and analysis of spatial characteristics in traditional villages, which includes three parts (Figure 2): field data collection, data processing, and data analysis and application. In the field data collection stage, low-altitude UAV aerial photography and handheld laser scanners are used to obtain 3D point cloud data. Afterwards, the point cloud data is processed, including denoising, fusion, and object classification. Finally, data analysis and application are conducted based on quantifiable indicators of village spatial characteristics. The following sections will focus on the technical characteristics of each step and explain the data acquisition methods and analysis techniques.

3.2.1. Step I: Field Data Collection: Obtaining 3D Point Cloud Data through Low-Altitude UAV Aerial Photography and Handheld Laser Scanners

Given the characteristics of traditional villages, such as high building density, small distance between individual buildings, and complex terrain, field data collection involves the combined use of low-altitude UAV aerial photography and handheld laser scanners to obtain 3D point cloud data.

For aerial photography, a DJI Mavic3 RTK UAV is used, equipped with a 20-megapixel camera, an 84° field of view, a 24 mm equivalent focal length, and an aperture range of f/2.8–f/11. The study area is divided into four flight zones based on the UAV’s flight endurance, control radius, and the terrain undulations of the area. The route length of each zone is not greater than 60% of the remote control radius of the unmanned aerial vehicle (UAV) platform to ensure flight safety. Each zone is flown using a five-way flight pattern to capture all facades of the buildings and minimize occlusions between buildings, ensuring the acquisition of three-dimensional data in narrow spaces (Figure 3). To ensure the accuracy of 3D point cloud model results and to avoid the UAV touching obstacles such as high-voltage lines during the flight (normally the height of high-voltage electricity pylons is 40–75 m), the flight altitude is set at 80 m, with a flight speed of 5 m/s. The resolution of the aerial images is 3.8 cm/pixel, and the overlap is set at 80% in the flight direction and 70% laterally. The total flight distance is 50 km, resulting in 1679 captured images, covering an area of approximately 458,700 square meters within the village. Handheld laser scanners are used to supplement three-dimensional data for indoor spaces, spaces under eaves, and streets and alleys (Figure 4). The study utilizes a Feima SLAM100 handheld laser scanner, which has a 360° rotating head and can generate point cloud coverage of 270° × 360°. It has a point frequency of 320 kpts/s and a maximum measurement range of 120 m. The scanning speed is approximately 0.5 m/s, covering a total length of 2.4 km for streets and alleys and 21 individual buildings for indoor scanning.

3.2.2. Step II: Data Processing: Point Cloud Preprocessing and Object Classification

The image data obtained from field data collection requires further processing to generate three-dimensional models suitable for analysis. This involves point cloud generation, point cloud preprocessing, and point cloud object classification. Agisoft Photoscan (1.4.5) software is used for modeling the data obtained from aerial photography. Through processes such as aerial tape splicing, denoising, smoothing, matching, geometric correction, homonymous point matching, and area network leveling, spatial three-dimensional processing is achieved. After completing the spatial three-dimensional processing, the preprocessed point cloud data and trajectory line files are combined for matching and fusion to perform three-dimensional reconstruction and generate a 3D point cloud model.

The classification of objects in the 3D point cloud model is a fundamental step for extracting spatial feature indicators and conducting data analysis. In this study, the CloudCompare (2.12.2) software is used for three-dimensional processing. By considering the color, surface orientation, height, and position features of the 3D point cloud, the originally chaotic point cloud or three-dimensional model is semi-automatically classified into separate data blocks representing different types of objects, such as vegetation, roofs, walls, and road networks. The main principle of object classification is to utilize the basic information provided by the point cloud, such as spatial coordinates, color, and normal vectors, to extract elevation features, texture features, and geometric forms required for village research. The specific methods for utilizing and classifying point cloud information are as follows: (1) Spatial coordinates: Using the three-dimensional coordinates (X, Y, Z) of the points and the relative positional differences between points, the size, height, and distance of the target objects can be calculated. (2) Color information: Differences in RGB color values can be used to classify buildings and vegetation with different colors. (3) Normal vector information: By setting an angle threshold for the point cloud, data blocks with significant slope differences, such as village ground, building walls, and roofs, can be classified, and the orientation and slope of the roofs can be analyzed. The resulting 3D point cloud model of the research area, after data processing and object classification, is shown in Figure 5.

3.2.3. Step III: Data Analysis and Application: Extraction of Quantitative Indicators of Village Spatial Characteristics and Interpretation of Construction Wisdom

This study primarily focuses on the quantification of typical spatial characteristics in villages at three levels: topographic environment, street spaces, and individual buildings. Different quantifiable indicators and analysis methods are extracted for each level, as shown in

Table 1. Quantitative indicators of spatial characteristics of villages.

Hierarchy	Spatial Characteristics	Quantitative Indicators	Analysis Software/Explanation
Topographic Environment	Interpretation of Feng Shui Forest	Topographic elevation, Feng Shui forest protection range	Arcgis 10.8
	Village Water Management	Water source, water volume, irrigation paths	Grasshopper 7.0 Rainwater runoff
Street Spaces	Street Scale	Street width	Direct calculation from point cloud
	Street Hierarchy	Distribution of streets Placement of House Doors in Streets	Combination of point cloud model statistics and field surveys
Individual Buildings	Courtyard Space Area	Building footprint area	Direct calculation from point cloud
	Saturation	Ratio of courtyard area to the area of the circumscribed rectangle	Higher value indicates more saturated courtyard space
	Boundary Coefficient	Ratio of courtyard perimeter to the perimeter of the circumscribed rectangle	The value closer to 1 indicates a more regular courtyard shape
	Courtyard Form	Combination and expansion relationships of courtyards	Point cloud model statistics

. Through the empirical study of typical examples, the work flow is demonstrated, enriching the study outcomes of village spatial characteristics and summarizing the technical approaches for studying spatial characteristics at each level, providing technical support for similar research.

4. Results

4.1. Topographic Environment: Visualization and Interpretation of Feng Shui Concepts and Site Selection Wisdom

4.1.1. Interpretation of the Connotation and Wisdom in Feng Shui Forest

Using 3D point cloud data, the quantifiable interpretation of Feng Shui concepts in traditional village site selection in China can be achieved, allowing for the quantitative evaluation of the Feng Shui environment in villages. The site of Liukeng Village in the research area was selected by representatives of the Feng Shui faction, Yang Junsong, a national teacher of the Southern Tang Dynasty, and his disciple, Zeng Wenchan. The Feng Shui pattern of the village consists of a geographical feature with a higher elevation in the south and a lower elevation in the north, with the village’s site facing a broad river channel to the northwest. In the winter, the impact of cold air is more pronounced due to the increased humidity caused by the river. The prevalent wind direction in Liukeng Village is from the northwest and northeast. There is a Malan Islet on the northern side of the village, which serves as a barrier against the cold winds from the northwest and northeast. Large areas of Feng Shui forests have been planted at the village entrance and near Malan Islet. Feng Shui forests are common in traditional villages throughout China, and their wisdom and characteristics are difficult to describe textually. However, three-dimensional data can visualize the village’s topography and showcase the wisdom of Feng Shui forest’s hindering effect on cold winter winds.

This study utilizes CloudCompare (2.12.2) software to convert the 3D point cloud source data into a pseudo-color elevation map (Figure 6) and adjust the height threshold to emphasize the subtle changes in the village’s elevation while minimizing the differences between high and low mountains. Compared to regular photos, this analysis map provides a clear and concise representation of the village’s topographic environment. In Figure 6A, the purple solid line in the diamond-shaped box represents the Feng Shui forest on Malan Islet, and the key characteristics of the defense line (Figure 6B) are highlighted. Malan Islet is the only gap in the defense line, with natural mountains on both sides blocking the wind. It is necessary to plant Feng Shui forests here to block the wind. Additionally, Malan Islet is a small plateau with an elevation slightly higher than the farmland. After planting trees, there is a height difference of 30m between the tree canopy and the farmland. According to the basic rule that the depth of the tree shadow area can reach 20 times the tree height, this Feng Shui forest can protect half of Liukeng Village during the cold and humid northwest wind in winter, significantly reducing the wind speed near the ground (indicated by the white dashed line in Figure 6C). Utilizing the architectural wind environment Phonics software, the wind speed within the village’s micro-terrain environment is simulated, considering the presence and absence of wind and water forests at a 6 m/s wind speed. Figure 7A shows the wind environment simulation without the Feng Shui forest, while Figure 7B depicts the wind environment simulation with the Feng Shui forest. Analysis of the wind speed distribution within the village reveals a marked reduction in wind speed at the Wu River surface (Area 1) due to the presence of the Feng Shui forest. Furthermore, the wind speed within the built-up area (Area 2) located in the eastern section of the village is also decreased. The analysis of the village’s topographic environment using 3D point cloud data allows for a visual understanding of the ancient people’s behavior, precise decision-making, and site selection wisdom, and also enables the visualization of traditional Chinese Feng Shui theories.

4.1.2. Micro-Topography-Based Quantification of Water Management

3D point cloud data can be used for the quantitative analysis of water management in villages. It is common to find natural water systems or channels and water ditches in traditional villages, which generally serve functions such as washing, landscape enhancement, and water storage. However, the exact capacity and specific functions of these water systems have been difficult to interpret using previous research methods. In the following section, the source, quantity, and irrigation paths of the main water source in Liukeng Village, Longhu Lake, are simulated and analyzed using 3D point cloud data, revealing for the first time the agricultural irrigation functions of Longhu Lake as an artificial water body within the village.

Regarding the source of Longhu Lake water in Liukeng Village, some scholars speculate that it comes from the Wu River [42]. This hypothesis can be supported in modern times with advanced water lifting technology. However, in ancient times, considering the significant elevation difference between the Wu River and the village base, this assumption is not realistic. Li Qiuxiang wrote that the water of Longhu Lake comes from mountain streams but did not specifically indicate which mountain stream. This study utilizes Grasshopper [45] to analyze the water convergence of the three-dimensional topography of Liukeng, unraveling the puzzle of “where does Longhu Lake’s water come from?” Based on the simulation results of rainfall runoff and the pseudo-color elevation map, the maximum water convergence range and convergence points of the nearby mountains can be clearly delineated. With the guidance of the simulation results, the entry point of the mountain stream into the lake was eventually found beneath the thriving aquatic plants on the lake’s bank. The hidden underground channels and diverting outlets that lead the water into the village were also discovered (Figure 8). The water inlet is well concealed, unknown to the local villagers. However, through the quantified results of the three-dimensional topography, the overall spatial characteristics of the village can be accurately understood.

The water quantity of Longhu Lake can be calculated using three-dimensional data. According to the measurement results shown in Figure 9A, the water convergence area of the mountain stream reaches 2.5 million square meters, while the area of Longhu Lake is approximately 40,300 square meters. With water supplied from the mountain stream to Longhu Lake, even in drought years with minimal rainfall, the A plot of farmland irrigated solely by Longhu Lake can obtain life-saving crops, enhancing disaster resilience and self-sufficiency. Based on the calculation that 1 square meter of paddy field requires an annual irrigation volume of 0.9 cubic meters, the irrigated area of 109,300 square meters in plot A requires 98,370 cubic meters of water, indicating that Longhu Lake with an area of 40,300 square meters would need to have a depth of 2.48 m. Based on the actual point cloud measurements during the summer desilting of Longhu Lake, the maximum depth of the lake is approximately 2.5 m, with a maximum water storage capacity of approximately 100,750 cubic meters, which precisely meets the irrigation needs of farmland A. From the topography of Liukeng Village, the entire village gradually descends from south to north, and the irrigation routes of the farmland also follow the natural terrain (Figure 9B). Several mountain streams converge at the mountaintop and flow into the mountaintop reservoir, which then enters the artificial water channel through a sluice gate. The water channel descends along the mountain slope, irrigating terraced fields along the way. At the entrance of the village, the water channel forms a pond and then splits into two paths through a diversion gate: one flows northwest, converging with the Wu River through farmland B; the other flows northeast, entering the village through an underground route and using Longhu Lake as a transit station to supply farmland A from the north end. The overall layout is shown in Figure 9.

The analysis of 3D point cloud data proves that Longhu Lake in Liukeng Village is a village-level water storage and drainage reservoir, serving not only daily life but also as an agricultural water management facility that does not require any mechanical pumping equipment. This paper fills the gap in the quantification of spatial indicators in previous conceptual research, deepening the understanding of the site selection, planning, and construction wisdom of traditional village topography.

4.2. Street Spaces: Spatial Scale and Hierarchical Division

4.2.1. Quantification of Street Space Scale

By utilizing semi-automated object classification of 3D point cloud data, the street facade point clouds along the streets of Liukeng Village can be extracted from the source data, enabling automatic calculation of the street widths in the village and achieving the quantification of street space.

Figure 10 shows the relationship between distance and the number of point clouds for each street in Liukeng Village. By comparing multiple charts side by side, the following patterns of street scale can be observed: (1) Street widths in Liukeng Village range from 2 to 6 m, as indicated by the significant number of point clouds within this range. (2) Compared to typical villages in Zhejiang and other areas, the average street width in Liukeng Village is slightly larger, at 3.62 m. Figure 10 presents box plots of street width values for four villages, including Shiliping Village, Longmen Village, Peilin Village, and Xinye Village, in Zhejiang, compared to Liukeng Village. From Figure 11, it can be seen that the median and mean street widths in Liukeng Village are larger than those in the other villages. The digital results, compared to previous textual summaries, provide a more intuitive and accurate reflection of the width variation and spatial distribution of the streets in Liukeng Village, quantifying the spatial characteristics of street spaces in traditional villages.

4.2.2. Hierarchy of Street Spaces

The street spaces in Liukeng Village represent a unique ancient urban-rural planning pattern in China known as the legacy of “neighborhood system”. These spaces are formed by a rich hierarchy and structure of streets, lanes, and doors that are dispersed and nested at different levels.

Based on the quantification of street scale discussed earlier, the street spaces in the traditional “neighborhood system” can be hierarchically divided according to the position, width, and accessibility of the streets. By evenly sampling the point clouds within each lane, the average width of the streets can be obtained, and the street spaces can be divided into four levels based on the average street width (Figure 12A). The average width of Level 1 roads is approximately 3.86 m, Level 2 roads have an average width of 3.16 m, Level 3 roads have an average width of 2.11 m, and Level 4 roads have an average width of approximately 1.14 m. The widths of the streets at each level progressively decrease, forming a stepped distribution. Different levels of street spaces not only exhibit regularity in terms of street width but also have distinct functions. Level 1 roads are the main thoroughfares, characterized by a “seven horizontal (east-west orientation) and one vertical (north-south orientation)” pattern, with two-story tower-style door pavilions at the entrances. Level 2 roads are short lanes connecting the main thoroughfares in the north-south direction and serve as important transportation networks connecting residential buildings to the main streets. Level 3 roads are east-west short lanes that complement the Level 2 lanes. Level 4 roads are narrow supporting lanes within building clusters, primarily for passage, with lower through traffic. In locations with relatively dense building layouts, they appear as enclosed alleys.

Within the hierarchical division of street spaces, each street’s “door” also exhibits a certain level of hierarchy. Enclosed doors can enclose buildings into independent areas, forming building clusters with distinct characteristics of the “neighborhood system”. In this paper, point cloud data was used to classify and quantify all the doors in Liukeng Village, showing their scale and hierarchical characteristics (Figure 12B). A total of nearly 70 surviving or variant door structures, including lookout-style inverted doors, double-sloped doors, eight-character doors, bottle-shaped doors, and regular-scale doors, were recorded. The doors can be divided into four levels based on their location and size: Level 1 doors are village doors, with an eave height of up to 8.5 m, and there are six village doors in total. Level 2 doors are inner lane doors, such as Yegui door, Juren door, and the bridge gallery located on the east bank of Longhu Lake. These doors do not exceed a total height of 5 m, providing a more intimate scale and smaller dimensions compared to village doors. Level 3 doors are Qu-street doors with a typical height of around 4 m. They are used to further divide the internal and external levels and serve as boundaries within the curved lanes. The most common forms are scale doors and double-sloped eaves, or a combination of both. Level 4 doors are household doors. These doors have rich decorations and are not limited by the hierarchical regulations of the “neighborhood system”. Figure 13 shows typical examples of doors at different levels in Liukeng Village, divided into four levels: village door, inner lane door, Qu-street door, and household door. The grid is utilized to compare the scale of different-level doors at the same scale.

4.3. Building Elements: Characteristics and Morphology of Residential Courtyards

4.3.1. Quantitative Analysis of Residential Courtyard Scale

A total of 30 well-preserved samples of Ming and Qing Dynasty residential buildings in Liukeng Village were selected to quantify the courtyard space area, building footprint area, area ratio, and saturation degree. The courtyard space area provides an understanding of the scale of the courtyards. The area ratio is the ratio of the courtyard area to the overall building’s external envelope area. The calculation of saturation degree refers to the study of graphic morphology, defining the degree to which the courtyard fills the external rectangular shape by the ratio of the projected area of the courtyard to the area of the circumscribed rectangle. By calculating and analyzing these indicators, the consistent patterns of residential courtyard spaces’ morphology can be comprehended at an overall level.

According to the statistical results from the 3D point cloud data: First, in terms of area (Figure 14A), the average courtyard area of the selected samples in Liukeng Village is 12.7 square meters, and 76.5% of samples are concentrated within 15 square meters. The average building footprint area is 223.6 square meters. It can be inferred that the courtyard spaces in Liukeng Village are generally small in scale, creating a more intimate spatial experience. Second, in terms of area ratio (Figure 14B), due to the varying scales of traditional residential buildings and the different constraints in courtyard design, there is significant variation in this indicator. The average area ratio of the courtyard spaces is 0.056, which is relatively smaller compared to northern courtyard spaces. Third, in terms of saturation degree (Figure 14C), the average saturation degree of the samples is 0.97, with 94.2% of the courtyard spaces concentrated between 0.95 and 1. This demonstrates that the majority of courtyard spaces in Liukeng Village exhibit a square and full appearance, influenced by the traditional concepts of stability and regularity in architectural design. By quantifying the characteristics of residential courtyard spaces, it can be concluded that the courtyards in Liukeng Village possess small-scale and square-shapedcharacteristics. The calculations of area, area ratio, and saturation degree provide a comprehensive and accurate understanding of the characteristics of residential courtyards in the village.

4.3.2. Summary of the Form of Residential Courtyards

Based on the quantification of courtyard spatial scale, 3D point cloud data of the research samples were obtained. By using semi-automatic point cloud segmentation technology, the roof boundaries were separated and orthographic roof images were generated. This allows for a quick understanding of the form of the residential courtyards in Liukeng Village and a clear identification of the roof combinations. This information can then be used to summarize the prototype and evolutionary types of residential courtyard spaces through visual diagrams.

The prototype of the residential courtyard form in Liukeng Village is the “one center, two sides” three-room courtyard with a central courtyard. The architectural plan unfolds vertically with the central courtyard as the main axis, consisting of a central room and a secondary room. The secondary room is closely connected to the side room. In addition, the transition between the central courtyard and the main hall is facilitated by the use of covered corridors. Other forms of courtyard layouts can be varied on the basis of this prototype, resulting in various combinations of courtyards [46]. The expansions mainly involve combining the basic type with additional courtyards to obtain better physical environments. This includes horizontal expansion with “cross-courtyard” layouts, vertical expansion with “multi-entrance” layouts, and comprehensive combination expansions. While the expansion of courtyards leads to more complex spatial hierarchies, the elements within each courtyard still adhere to the original structural relationships. Figure 15 below illustrates the basic forms of courtyard layouts, providing a clear visualization of the types and combinations of residential courtyard forms in Liukeng Village in the form of a map.

5. Discussion

5.1. Strategy for Utilizing 3D Point Cloud Data

High-precision 3D point cloud data forms the foundation for the quantitative analysis of village spatial features; consequently, the accuracy of the entire area and individual building models in Liukeng Village warrants examination. A spatial 3D coordinate system, based on the entire area model, was developed, and 15 feature points were selected for coordinate position accuracy analysis. The results indicate that the median errors for the X and Y coordinates are 0.01 m and 0.09 m, respectively, both under the threshold of 0.14 m, thus complying with the ±0.15 m accuracy standard stipulated by China’s external digital mapping technical regulations (GB/T 14912-2005) [47]. Additionally, the median error for the Z coordinate is 0.07 m, aligning with the 1/3 accuracy requirement for the 0.5 m isometric distance of flatland in the 1:500 elevation topographic map. Six individual buildings at diverse locations within the village were selected, and their dimensions—length, width, and height—were measured. These measurements were then compared with data collected at the same sites in the field using ground-based portable rangefinders to ascertain whether the digital surface model satisfies the buildings’ measurement accuracy requirements. The outcomes indicate that the errors—0.03 m in length, 0.08 m in width, and 0.09 m in height—all remain below 0.15 m, thereby meeting the high-precision requirements stipulated by various field digital mapping technical regulations. Drawing upon the analyses regarding the accuracy of feature coordinates and the mapping of building units via 3D point cloud data, it is evident that the precision of such data comprehensively fulfills the fundamental accuracy requirements essential for quantifying the spatial features of villages.

After segmenting the point cloud data, quantitative indicators can be directly calculated. To verify the effectiveness of the data obtained through point cloud segmentation, the roofs were extracted and the building heights were calculated using the roof point cloud. By summarizing the vertical distances (relative heights) between all roof points and the adjacent ground points of each building, Figure 16A was obtained. The highest point cloud density was observed in the height range of 3.5–7.1 m, and the number of point clouds above 7.1 m is only 23.5% of the total, and the slope of the curve in the image is increasing rapidly, showing a downward trend. Based on this, it can be inferred that the ridge height of the main buildings in traditional Liukeng Village residences is concentrated around 7.1 m (on average), as there are fewer buildings with ridge heights exceeding this value. To validate this inference and the reliability of this statistical method, 158 out of nearly 200 randomly selected traditional residences in Liukeng Village were manually measured for their ridge heights. The results are shown in Figure 16B, with an average ridge height of 7.13 m and an eave height of 5.11 m. The comparison between the point cloud calculation and manual measurement confirms the reliability of this method.

Using the research method described in this paper, in addition to the quantitative analysis already conducted, 3D point cloud data can be used to statistically analyze various spatial characteristics of villages at different scales, including individual buildings, street patterns, and terrain environments. This data can also be used for spatial morphology comparisons across different regions and multiple villages. By combining the spatial characteristics of villages, the main quantitative analysis indicators at each level can be listed, as shown in

Table 2. Indicators for quantitative analysis of village levels.

Hierarchy		Basic Features	Specific Content	Computational/Analytical Methods
Terrain environment	Ground	Slope, direction, elevation	Mountain and ground slope, slope direction, elevation	In combination with gis
	Forest	Shape	Shape index of tree crown projection in forest land	In combination with gis
		Area	Area of tree crown projection in forest land	Direct calculation from point cloud
		Height	Average height and height variation in forest land	Direct calculation from point cloud
	Water	Length, width	Length and width of the minimum bounding rectangle of pond contours	Direct calculation from point cloud
		Perimeter, area	Perimeter and area of pond contours	Direct calculation from point cloud
		Volume	Water storage capacity of ponds	Physical simulation based on 3D data
Street hierarchy	Space	Width	Width of streets, i.e., distance between outer walls of buildings on both sides	Direct calculation from point cloud
		Height	Height of streets, i.e., height of buildings along the street and variations along the road	Direct calculation from point cloud
		Height-to-width ratio	Height-to-width ratio of streets and variations along the road	Direct calculation from point cloud
		Visibility analysis	Visibility analysis of landmark features	Physical simulation based on 3D data
	Interface	Slope, direction	Slope and direction of street surfaces	In combination with gis
		Sinuosity	Ratio of actual length of street centerline to endpoint length	Direct calculation from point cloud
		Distribution density	Public space ratio: ratio of public space area to total base area	Direct calculation from point cloud
Building hierarchy	Space	Building height	Height from ground to flat/sloping roofs, ridge height and eave height of sloping roofs (individual/average)	Direct calculation from point cloud
			Overall height distribution and height differences of building clusters	Direct calculation from point cloud
	Roof	Number of buildings	Number of buildings divided by roof type	Direct calculation from point cloud
		Slope, direction	Slope and direction of roofs (individual/average)	Direct calculation from point cloud
		Shape	Length-to-width ratio of roofs (individual/average)	Direct calculation from point cloud
		Length, width	Length and width of roofs (individual/average)	Direct calculation from point cloud
	Ground	Shape	Length-to-width ratio of building footprints (individual/average)	Direct calculation from point cloud
		Area	Building footprint area = roof projection area—eave area (individual/average)	Direct calculation from point cloud

. The primary calculation methods for each indicator include direct calculation based on point clouds, calculation using spatial syntax and GIS tools after vectorization editing, and physical simulation calculation based on three-dimensional data. The content discussed in this paper pertains to the direct calculation method using point clouds. The advantage of direct point cloud calculation is that there is no need to convert the point cloud data obtained from photogrammetry into other formats, and the details of objects, such as the width and height of a door, can be accurately measured, and the accuracy of the data can be as high as 0.01 m. Point cloud calculations overlap with GIS calculations in extracting environmental features such as terrain slope and aspect; therefore, when high accuracy is not required, point cloud elevation data can be converted into Digital Elevation Model data for analysis using GIS. Spatial syntax is mainly used to calculate indicators related to street and alley spaces and village patterns, further analyzing the connectivity and correlation between streets and alleys. The third category involves extracting structural information by converting three-dimensional data formats into simplified vector models, which are then input into thermal environment simulation software to obtain simulation results for wind, heat, humidity, and sunlight. Environmental simulation based on micro-topographic three-dimensional data is also an important quantitative method for assessing the wisdom and livability strategies of traditional villages.

5.2. Innovations

This paper utilizes 3D point cloud data and establishes a complete technical route for the quantitative analysis of the spatial characteristics of traditional villages. Taking Liukeng Village in Fuzhou, Jiangxi Province, China, as an example, the paper quantitatively analyzes the spatial characteristics of the village at three levels: terrain environment, street space, and individual buildings. The innovations of this paper are as follows:

1.

Establishing a complete work chain for calculating village spatial characteristics

Based on 3D point cloud data, a fast data acquisition, processing, and deep utilization mode can be established, covering the entire process from field data collection to direct spatial calculation of 3D point clouds. This breaks through the limitations of manual mapping, three-dimensional roaming, and the bottleneck of indoor processing of three-dimensional data. In the current context where low-altitude unmanned aerial vehicle (UAV) photogrammetry is widely used in village research, effectively utilizing the obtained three-dimensional data is a key factor in improving the value of equipment utilization. In addition to direct point cloud calculation, data transformation is utilized to support GIS, spatial syntax, and various analysis theories or methods. By integrating these approaches, an optimal combination of methods for calculating spatial characteristics of traditional villages is sorted out. This forms a bridge connecting source data to valuable information and constitutes a comprehensive quantitative method for analyzing village spatial characteristics.

2.

Addressing the challenge of rapid extraction of spatial characteristics

By using the “air-ground coordination” approach to quickly acquire 3D point cloud data on-site in villages, especially through the calculation of 3D point clouds combined with other spatial analysis methods, spatial indicators can be semi-automatically extracted from the point clouds in a short period. This avoids the need for manual mapping and individual statistics, allowing for the rapid completion of estimation work. This enables researchers to obtain quantitative spatial indicator data for unfamiliar villages, large-scale villages, or complex street villages within a few days, greatly improving fieldwork efficiency and the depth of research results. The obtained three-dimensional source data of the village can also be used for heritage data archiving, computer model construction, virtual roaming, and other traditional purposes.

3.

Achieving three-dimensional quantification and visualization of traditional Chinese construction intelligence

Through the analysis and application of 3D point cloud data, methods for extracting various indicators of terrain environment, street space, and individual buildings at different levels in a three-dimensional context can be summarized. This gives deeper meaning to three-dimensional village data, allowing for accurate analysis of the current state of the village and the summarization of intelligent and livable strategies for village location and construction. By intelligently identifying digital surface models, automatic classification and overlay comparison can be performed, generating multiple three-dimensional vector information layers for terrain, streets, and individual buildings. This allows for comparative analysis and visual presentation. This technological innovation not only eliminates interference and analyzes individual systems but also expands the recognition and understanding of the interrelationships between various elements. It helps analyze the adaptability intelligence in the formation process of traditional village spatial characteristics, which is particularly important for the study of villages with complex terrains. 3D point cloud data not only serves as a model for visually displaying the spatial structure of villages but also becomes an important foundation for exploring and interpreting the diverse values of villages. It deepens the public’s understanding of the multiple values of traditional villages, strengthens their sense of identity and pride in local culture, and establishes cultural confidence, thereby assisting in the inheritance of village cultural heritage.

5.3. Limitations

The limitations of this paper lie in the extraction of spatial characteristics and the revelation of patterns in a single village sample, as well as the cross-regional comparison of certain typical indicators. Due to regional differences among traditional village samples, there are significant differences in spatial characteristics between different villages. Therefore, it is necessary to further systematically summarize the differential indicators and improve the categorization of indicators. Future research will utilize the proposed technical route to acquire large amounts of 3D point cloud data from villages, incorporate statistical data into standardized databases for digital museums, achieve rapid aggregation and quantitative statistics, and conduct comparisons of cross-regional architectural features. This will enhance the application potential of 3D point cloud data in village big data nationwide and deepen the understanding of traditional Chinese villages. The paper will also use 3D point cloud data as a basis to support in-depth analysis of various aspects of village research, such as defense and safety, disaster resilience, architectural preservation, and spatial perception based on quantitative indicators. The visualization results will naturally contribute to the study of traditional village intelligence.

5.4. Future Application Scenarios

Based on the case study of the quantification of spatial characteristics of Liukeng Village, the application potential of 3D point cloud data is vast. Therefore, a proposal is made to construct a 3D point cloud digital cloud platform based on village data. Through the construction of this cloud platform, comprehensive utilization of village 3D point cloud data can be achieved, facilitating the convergence and utilization of relevant digital information, breaking through the information exchange bottleneck between upstream scientific research platforms and downstream cultural and tourism service industries. This will create data cloud services to support various aspects of traditional village construction, realizing the informatization, spatialization, and visualization of traditional village spatial characteristics. The cloud platform consists of three parts: the village spatial basic information data layer, the village historical and cultural knowledge application layer, and the village public display and tour service layer.

5.4.1. Village Spatial Basic Information Data Layer

The village spatial basic information data layer includes detailed records of the village’s basic data. Three-dimensional data mainly consists of detailed three-dimensional models and 360-degree panoramic roaming results obtained through low-altitude photogrammetry techniques. It also includes tangible traditional resources such as the traditional village site selection pattern and surrounding environment, traditional buildings, historical environmental elements, as well as historical documents, various planning, policies, and management systems related to the protection and development of traditional villages. These materials are systematically organized and archived in the form of drawings, documents, videos, recordings, etc.

5.4.2. Village Historical and Cultural Knowledge Application Layer

The village historical and cultural knowledge application layer aims to popularize and educate the public about rural history and culture. It comprehensively interprets two types of social relationships—lineage and geographical—and their changes and developments through different types of heritage elements. It uses architecture, space, landscapes, and legends to understand how the social structure and local identity of traditional villages were constructed during specific periods. This systematic interpretation enriches the scope of rural heritage, expanding from material remains to landscapes and material cultural heritage. It also reveals the regional context and structural processes through in-depth exploration of local historical and cultural aspects, enhancing the local and temporal aspects of interpretation. This layer contributes to the dynamic inheritance of folk culture and the revitalization of intangible cultural heritage representative projects, as well as the activation and utilization of other traditional production methods, lifestyles, social relationships, rural customs, and folk skills.

5.4.3. Village Public Display and Tour Service Layer

The village public display and tour service layer involves the construction of online digital museums for villages. It focuses on one or multiple traditional villages, selecting topics, planning, and producing content that showcases the cultural characteristics of the villages. Through the tour service layer, various interactive scenarios are provided for the public to participate in and immerse themselves in. Traditional village virtual simulation models and various digital presentations are established through panoramic roaming and other methods, allowing viewers to have a richer and multidimensional understanding of traditional villages. Emotional experiences can promote good interaction between digital museums and users, enabling users to engage in emotional communication and interaction, thereby promoting the cultural inheritance of traditional villages.

In summary, the construction of a 3D point cloud digital cloud platform for traditional villages can achieve the following aspects: (1) Comprehensive aggregation of basic data and diverse information of the village, reproducing the authenticity, integrity, and systematic nature of historical buildings and local heritage across different periods. (2) Digital representation of traditional village culture and historical features, telling the story of China, and serving the major strategic goal of national cultural rejuvenation. (3) Filling the gap in public communication channels (platforms), improving cultural and educational experiences, and implementing the functions of cultural inheritance carriers.

6. Conclusions

This paper utilizes 3D point cloud data to develop a complete technical route for the quantification analysis of spatial characteristics in traditional villages. The route includes field data collection, data processing, and data analysis and application. It enables the rapid acquisition, processing, and analysis of three-dimensional village data. Using Liukeng Village as an example, the paper quantifies the spatial characteristics of the village at three levels: terrain, street space, and individual buildings. Through the quantification and visualization of three-dimensional data, the spatial intelligence inherent in traditional Chinese villages is revealed. The paper efficiently achieves rich village spatial indicator estimation results, without the need for manual mapping, meeting the requirements for on-site investigations and supporting the depth and breadth of cross-regional and multi-village comparative research.

Through the quantification analysis of 3D point cloud data, the paper identifies the regularities in village spatial characteristics across three typical examples at different levels. The preliminary conclusions are as follows: (1) Terrain and site selection: Traditional villages consider Feng Shui theories in their site selection, using Feng Shui forests to protect against cold northwest winds in winter. The precise location of water ponds within the village demonstrates the high precision and scientific nature of ancient water engineering through irrigation paths and rainwater runoff analysis. (2) Street space: The scale of street space is highly coupled with functional requirements and forms a multi-level spatial hierarchy based on the traditional “neighborhood system” pattern. The division of spatial hierarchy is related to the scale of streets and the size of neighborhood gates. (3) Individual buildings: Residential courtyard spaces have small-scale and square-shaped characteristics. They evolve from the prototype of the “one center, two sides” courtyard with a central courtyard, and the paper provides a visual diagram summarizing the evolution of courtyard spaces in traditional houses.

This paper establishes a complete technical route for the quantification analysis of villages and, through this analysis, explores the value of traditional village spatial intelligence, deepening the public’s understanding of the diverse values of traditional villages. The research method developed in this paper is not only applicable to the quantification analysis of traditional villages but also provides reference and guidance for other scholars facing similar analytical needs. It promotes the popular application of this technical method in the field of digital architecture and provides technical support for research and practice in related fields.