Spatial-Temporal Evolution Characteristics Analysis of Color Steel Buildings in Lanzhou City

Abstract

With the continuous acceleration of China’s urbanization process, color steel plate, as a new type of building material, has been widely used in all kinds of temporary buildings and has become the spatial carrier of the specific development stage of urbanization. This study focuses on Lanzhou City as a case study to deeply analyze the spatiotemporal distribution and evolution of color steel plate buildings. Utilizing data extracted from Google imagery and GF-2 satellite images of the built-up areas in Lanzhou, spatial statistical and analytical methods such as centroid analysis, compactness index, and patch density are applied. Systematic analysis is conducted across different time periods and spatial scales to examine the evolution of indicators, including quantity, centroid distribution, spatial clustering, and distribution direction. The results show that from 2013 to 2021, the prevalence of color steel buildings in Lanzhou city initially increased and then decreased, and the number peaked in 2017, but there is a significant difference between distinct areas in the urban area. By quantitatively analyzing the spatial and temporal evolution characteristics of color steel plate buildings, this study reveals the important role it plays in promoting the urbanization process and provides a scientific basis for relevant planning decisions.

1. Introduction

In recent years, China’s urbanization rate has increased and the scale of construction land has shown explosive growth, while land finance revenue, residents’ income, urban population, transportation costs, and the level of agricultural land rent are all influencing urban expansion [1]. In the context of new urbanization, the current land use, development, and management are facing urgent challenges [2], and geospatial analysis and visualization can be used to understand the land use structure and its changes at different spatial scales, which is critical to unlocking the potential for sustainable land use development [3]. Zhou et al. simulated urban development through multiple scenarios of urban land to reveal the potential problems and conflicts of the future spatial expansion of urban land [4]. Zhang et al. designed a vector model to analyze the spatial structural characteristics of the buildings and portray the characteristics of the urban structure and the changing law of the city of Beijing [5]. Lu et al. used nighttime remote sensing light images to identify and analyze urban spatial structure and evolution characteristics [6]. Zhou et al. used slope data and multi-period land cover data to reveal the evolution of vertical spatial-temporal patterns of urban sprawl in China from 1990 to 2015 [7]. In the process of urbanization, color steel plate is widely used in public buildings, large-scale factories, and movable board houses [8], and as the main building material, it has been widely distributed in rapidly evolving areas such as urban villages, urban-rural combinations, and industrial parks, leading to large-scale and intensive distribution of small-scale color steel plate buildings in the above-mentioned plots. The data extracted from remote sensing images suggests that the number of color steel plate buildings in the Lanzhou city urban area reached 27,318 buildings in 2017, with a total area of about 15.39 km². This indicates that color steel plate buildings are an important part of urban space. Since the 20th century, in the process of economic development and urban expansion, major cities in Northwest China have used color steel plate buildings as new spatial carriers, which play an important role in actively promoting the expansion and optimization of the spatial structure of the city and the transformation of functional functions [9].

Research on the spatial distribution pattern and influencing factors of color steel plates has important practical significance for urban planning, land management, environmental monitoring, and other fields. Existing research on color steel plates mainly focuses on spatial analysis using remote sensing data. Yang et al. extracted data on colorful steel buildings from multi-source high-resolution satellite images, classified and integrated color steel plate buildings of different categories and sizes, and studied their spatial distribution and aggregation characteristics using spatial analysis methods [10]. To address the errors and spatial scale limitations in the data, the research direction has shifted towards the study of remote sensing image extraction algorithms. Heiden et al. used hyperspectral data to identify materials on the surface of urban buildings and analyzed them based on grayscale values [11]. Zhang et al. utilized the characteristics of Synthetic Aperture Radar (SAR) to develop a Triplet Network that integrates optical technology and SAR data to optimize the accuracy and performance of traditional semantic segmentation models in extracting color steel buildings [12]. Breakthroughs in algorithm technology have improved the accuracy and quality of research data on color steel plates, providing further possibilities for interdisciplinary fusion. Zhang et al. used the U-Net network to extract data on color steel buildings from satellite images from Google and GF-2, calculated weighted density, landscape indices, and other indicators, and explored the relationship between color steel buildings and urban heat island effects [13,14]. However, existing research focuses on the extraction technology of color steel buildings in remote sensing images, with relatively single and traditional spatial analysis methods, and rarely considers the role of their temporal characteristics in spatial distribution.

Rapid urbanization will produce many social and environmental impacts. In some of the new or under-construction industrial parks, there are large-scale color steel plate buildings, mainly used for factories, logistics warehousing, and so on. However, the existence of these temporary buildings not only affects the cityscape but also poses safety risks such as fire and collapse. Therefore, it is very important to detect its evolving characteristics and analyze its impacts [15]. Based on multi-temporal data from 2013 to 2021, this study employs various spatiotemporal analysis methods to focus on the spatiotemporal evolution characteristics and patterns of the color steel building clusters in Lanzhou City, including overall features, centroid deviations, and clustering degrees. Starting from the perspective of spatiotemporal integration, this research provides new insights into the spatial pattern research of color steel plates under the background of urbanization, thereby extending to provide scientific support and decision-making for the construction planning of sustainable development cities.

This study is divided into a total of six sections, and Section 2 describes the research status of the color steel plate building group. Section 3 of the research area, data, and specific research methods used. Section 4 carries out experimental analysis to get the evolution results of different spatial analysis methods in the Lanzhou color steel plate building group. Section 5 discusses the comprehensive results for color steel plates and compares and analyzes them with previous studies, as well as points out the shortcomings of this study and future research directions. Section 6 summarizes the specific findings of the article.

2. Related Work

Nowadays, the status quo of color steel building complexity has attracted the attention of many scholars, and its spatial logic analysis, spatiotemporal characteristics, spatiotemporal evolution, coating aging, and fire risk analysis are the focus of attention. However, the existing studies mainly focus on urban land use, urban expansion, industrial economy, environmental pollution, urban-rural differences, and other issues [16,17]. Research on the spatiotemporal evolution and distribution patterns of color steel plate buildings is still relatively scarce in terms of theoretical and methodological frameworks. This study combines GIS spatial analysis technology and spatial theory methods to study the spatial coupling relationship of color steel plate buildings and also analyzes the spatiotemporal evolution of color steel plate buildings in urban areas of Lanzhou City.

2.1. Study on the Spatiotemporal Evolution Rules and Characteristics of Typical Geographic Entities

Geographic entities are independent natural or artificial features in the real world with spatial location and common attributes [18,19], which may be objective houses and vegetation or political districts and contour lines abstracted by human beings. Geographic entities can act as a natural bridge between geographic information and various thematic information, enabling the integration and sharing of information. The data model based on geographic entities has many advantages. First, geographic entities as cognitive units are more in line with the cognitive habits of the general public and are easy to understand and use by non-specialists; second, geographic entities can provide a basis for linking geographic information and other thematic data; and third, they provide a reliable way to realize entity-level data updates [20]. Geographic entity data adopts the entity-oriented modeling method, with geometric elements as the basic unit of spatial data expression and classification hierarchical organization [21]. The conceptual model of geographic entity data is shown in Figure 1, which consists of an entity layer and an element layer. The element layer is the constituent unit of geographic entities, expressed as points, lines, surfaces, and bodies, which represent single, connected, and homogeneous geometric objects in the space, and these points, lines, surfaces, and bodies are uniquely identified by the element identification code. The entity layer consists of one or more graphic elements, usually expressed by geometric complexes but can also be expressed by geometric primitives identified by entity codes [22,23].

Geographic entity data is the foundation and bridge of the geographic information platform. The production of map data is built based on basic geographic information data, firstly forming geographic entity data; then adding social and economic information based on geographic entity data to form governmental electronic map data; adding comprehensive social information based on governmental electronic map data and forming public electronic map data after encrypted processing [8]. With the in-depth development of smart city construction based on geographic entity location information integration of natural resources, transportation, environmental protection, emergency disaster mitigation, population, economic and credit information, and other multi-sectoral data to build urban management, big data has become an inevitable trend in the development of the current smart city [24]. The extension of information from the natural attributes of houses to their social attributes realizes the specialized needs of different departments while ensuring that the basic results meet relevant technical standards of the surveying and mapping industry. This can provide refined and accurate geographic information and location services for smart city construction, with extensive applications in areas such as comprehensive renovation of residential communities, finding illegal houses, and de-scaling urban development [25].

2.2. Spatial and Temporal Evolution Law and Characteristics of Color Steel Plate Building Group

Rapid and disorderly urban growth has triggered urban expansion and irreversible land cover changes. Medium-sized cities, in particular, have become increasingly important in achieving sustainable urban development and have become the focus of research and policymaking. However, there is a clear gap in assessing urban sprawl in these cities [26]. To address this challenge, the United Nations established 17 Sustainable Development Goals (SDGs) in 2015, with a particular emphasis on the goal of achieving sustainable cities and communities. The achievement of this goal requires researchers to be able to visualize and analyze the status and trends of SDG indicators [27]. In this regard, Wang et al. developed a geospatial big data analysis engine based on SuperMap iObject for Java and Apache Spark [28]. And based on this, they proposed an innovative solution: a spatiotemporal big data visualization framework that integrates open-source map libraries, visualization libraries, and modern web development techniques and utilizes Spark Streaming for real-time data processing. Meanwhile, they mapped the results in real-time to DataFlowLayer, which supports high-performance geospatial big data analytics by using GIScript (https://github.com/skyswind/GIScript?tab=readme-ov-file, accessed on 25 May 2024) and iDesktop Cross (https://supermap-idesktop.github.io/SuperMap-iDesktop-Cross/, accessed on 25 May 2024) to support high-performance spatial and temporal big data spatial analysis for more accurate assessment and planning of urban development [29].

Among the many areas of sustainable urban development, the study of spatial differences in housing conditions is of particular importance. Housing is not only a fundamental issue in determining the sustainable development of cities and regions, but it is also crucial for understanding how housing conditions affect the economic and social dimensions [30]. Peter et al. assessed the spatial constraints associated with inner-city residential construction by creating an Index of Residential Development (IoRD), which can help policymakers identify spatial development targets for further planning [31]. Meanwhile, a study by Osman et al. revealed how the urban characteristics of unplanned settlements can affect sustainable development, especially in regions such as Egypt [30]. In addition, Bai et al. used a multi-level fuzzy integrated evaluation method to construct community sustainable development evaluation indicators from a micro perspective, which provided a quantitative tool for community development [32]. And de Siqueira et al. provided another perspective by evaluating the relationship between urban development actions and sustainability based on the LEED-ND indicator system [33].

Detection of temporary color steel buildings and their spatial and temporal patterns and characterization are important for many developing cities, and areas with dense distributions of color steel buildings are usually problematic with high population densities and high levels of sustainability and risk [34]. Therefore, in the context of exploring sustainable urban development, color steel buildings also serve as an important measure of sustainable urban development, which is distributed in large quantities in large and medium-sized cities and are widely used in buildings such as factories, warehouses, workshops, and private houses due to its advantages of low price and convenient installation [35], etc. Therefore, the study of the spatial and temporal distribution pattern of the urban color steel building complex is important for understanding the city’s operating conditions, solving urban problems, and promoting sustainable urban development. Hou et al. proposed a new benchmark dataset for retrieving color steel sheds from Google Earth images with a total number of 2407 remote sensing images [36]. Samat et al. used Sentinel-2A/B MSIL2A imagery to map blue and red color steel buildings across China [37]. Hong et al. utilized remote sensing imagery and instrumental experiments to analyze the influencing factors of the reflectance of color steel plates and to investigate their spectral characteristics, providing technical support for information extraction [38]. Sun et al. used CNN-based blue steel roof information extraction and Gaofen-2 imagery to analyze the distribution of colored steel buildings [39]. Dong et al., considering the safety hazards posed by lightweight and heavy floating objects in railway operation environments, innovatively combined a large model to propose a dual-branch semantic segmentation network for extracting color steel building structures [40]. This aims to mitigate safety incidents caused by heavy floating objects resulting from color steel building structures. Most of the existing research’s attention to the color steel plate building clusters in the city focuses on fire spacing [41,42], fire characteristics [43], fire rescue countermeasures, and the illegal remediation of color steel plate buildings in the urban area, etc. And there are fewer studies on the extraction of its spatial elements, spatial distribution characteristics, etc., especially in urban industrial parks, where the large-volume color steel plate buildings are clustered and are paid less attention to. Industrial parks appeared at the end of the 19th century to carry industries and promote industrial agglomeration, and in the long course of development, the function of industrial parks has gradually changed from single to diversified. Due to the slow development of industry in Northwest China, the construction of industrial parks is also relatively simple. Color steel plate construction is low-cost and can realize large-span architecture, so it exists in a large number of industrial parks in the city. Based on Google Earth image data, Gao et al. analyzed the coupling relationship between industrial parks and color steel plate building clusters in Yinchuan City and found that the core density area of color steel plate building clusters overlapped with national and provincial industrial parks [44]. Li et al. extracted data on color steel building clusters from Google Earth imagery and combined various spatiotemporal analysis methods to investigate the spatiotemporal distribution patterns and evolutionary trends of industrial parks in Xining City [45]. Industrial parks, transportation networks, and land use are the main factors affecting the spatial distribution of color steel plate building clusters. In the key cities of Northwest China, there is a certain overlap between the color plate building clusters and industrial parks, and the study of the spatial and temporal distribution characteristics of the color plate building clusters is of great significance to the study of industrial parks in Northwest China and the analysis of the urbanization process [46].

Li et al. investigated the spatial distribution characteristics of small color steel buildings and large color steel buildings by kernel density analysis, average nearest neighbor, and standard deviation ellipse correlation methods for Anning District, Lanzhou City, respectively [47]. Ma and others selected color steel buildings in the Anning District of Lanzhou City as the research object and studied color steel buildings in terms of temporal and spatial changes, stability, degree of fragmentation, and aggregation characteristics, concluding that the distribution characteristics of color steel buildings in the region are significant [48]. One of the main directions of current research is to deconstruct the geographical distribution of color steel buildings using remote sensing imagery data. Another important direction of research involves developing new methods and integrating various data sources to characterize aspects such as urban spatial morphology and development level related to color steel buildings. Zhang et al. used satellite imagery to dynamically monitor color steel buildings, revealing the spatial morphology and development level of the Munyaka Region in Kenya [49]. Wang et al. through the analysis of the spatial layout and agglomeration characteristics of color steel plate houses in urban areas, drew the connection between large and small color steel plate houses and the spatial structure of other cities and pointed out that large and small color steel plate houses are mostly located in industrial districts and that small color steel plate houses are mostly found in urban villages and urban-rural junctions [50]. Gao et al. selected the concentrated areas of color steel buildings in urban villages and industrial parks, carried out the determination of area and spacing, and registered the corresponding use [51]. The extensive spatial analysis research on color steel buildings provides scientific support for sustainable urban development.

As shown in Figure 2, we have outlined the current state of research on color steel plates. This research is driven by the trend of urbanization and the context of the Sustainable Development Goals (SDGs). The types of research on sustainable cities are diverse, encompassing urban infrastructure, urban buildings, urban greening, and more. Color steel buildings fall under the category of urban buildings in sustainable cities. Research on color steel buildings is based on foundational studies of base materials and chemical coatings, such as coatings research, and extends to geographic applications like remote sensing extraction and spatiotemporal analysis. However, these related studies mainly rely on government work reports, statistical bulletins, statistical yearbooks, and other statistical data for analysis, with large human influence factors and periodicity and timeliness limitations, making it impossible to conduct dynamic and continuous monitoring. There is a lack of intuitive and dynamic data that can effectively map the temporal and spatial changes of the color steel plate building complex, and the study of the spatial and temporal evolution law is insufficient.

3. Materials and Methods

3.1. Study Area

Lanzhou is located in the geometric center of China’s land map, is the central city within the Yellow River Basin, and is one of the typical mountainous cities in northwest China. Lanzhou has a mesothermal continental climate, with an average annual precipitation of 324 mm, an average annual evaporation of 1676 mm, and a dryness index of 5.17. The general topography is high in the northwest and low in the southeast, with an elevation of 1418~3677 m. The geomorphological landscapes are complex and varied, and the main green spaces are dominated by meadows, croplands, and shrublands. The Yellow River passes through the city from southwest to northeast, dividing the two mountains, creating a river-valley-type city form with an east-west length of about 35 km and a north-south width of 2~8 km, which leads to the tightness of urban construction land. It is a typical mountainous city in Northwest China, with drought and little rain, scarce construction land, a large population, and great development pressure. It is subject to the double pressure of urban construction expansion and protection of the ecological environment, and this phenomenon is concentrated in the development and construction of the city center area.

In 2013, there were 20,497 color steel plate buildings in Lanzhou City, with a building area of 11.71 km². Despite a decreasing trend in the city’s urban and suburban areas due to the ongoing transformation of shantytowns and comprehensive improvement of the urban environment, large steel sheet buildings in industrial parks have steadily increased. In 2021, Lanzhou had 26,425 steel sheet buildings covering an area of 14.72 km². The studied urban area of Lanzhou includes the following districts: Chengguan, Qilihe, Anning, Xigu, and Honggu, with a total area of approximately 1640.37 km². The research area is depicted in Figure 3.

This study obtained spatiotemporal data on Lanzhou’s steel sheet buildings from 2013 to 2021. The 2013 steel sheet building data were extracted from Google Imagery using manual vectorization. From 2014 onwards, steel sheet building data were derived from GF-2 satellite imagery using a neural network model for extraction.

3.2. Research Methods

3.2.1. Distribution and Transfer of Center of Gravity

The spatial centroid represents the geographical center of a set of spatial elements and can assist in determining the clustering location of steel structure buildings [52]. Through the transfer distance and direction of the centroid [53,54], the expansion trend of urban steel structure buildings can be inferred [55]. Simultaneously, this study takes into account the varying areas of steel structure buildings, using the area attribute as the weight value for calculating the centroid. The formulas for the weighted average gravity center [56,57], transfer distance (D), and transfer angle (β) of the centroid are as follows:

$$\begin{array}{c}\overline{X}=\frac{{\sum }_{i=1}^{n}Are{a}_i{x}_i}{{\sum }_{i=1}^{n}Are{a}_i} \end{array}$$

(1)

$$\begin{array}{c}\overline{Y}=\frac{{\sum }_{i=1}^{n}Are{a}_i{y}_i}{{\sum }_{i=1}^{n}Are{a}_i} \end{array}$$

(2)

$$\begin{array}{c}{D}_{t+1}=\sqrt{{\left(\overline{X_{t+1}}-\overline{X_t}\right)}^2}+\sqrt{{\left(\overline{Y_{t+1}}-\overline{Y_t}\right)}^2} \end{array}$$

(3)

$${\beta }_{t+1}=arctan\left(\frac{\overline{X_{t+1}}-\overline{X_t}}{\overline{Y_{t+1}}-\overline{Y_t}}\right)\left(\overline{Y_{t+1}}-\overline{Y_t}\ge 0\right),$$

$$\begin{array}{c}{\beta }_{t+1}=\pi +arctan\left(\frac{\overline{X_{t+1}}-\overline{X_t}}{\overline{Y_{t+1}}-\overline{Y_t}}\right)\left(\overline{Y_{t+1}}-\overline{Y_t}\le 0\right). \end{array}$$

(4)

where $Are{a}_i$ represents the area of steel structure building $i$, $\overline{X}$ represents the abscissa of the weighted distribution centroid of steel structure buildings during a certain period, $\overline{Y}$ denotes the ordinate of the weighted distribution centroid of steel structure buildings during a certain period, $\overline{Y_{i+1}}-\overline{Y_i}$ is the offset on the y-axis between adjacent two centroid points, $\overline{X_{i+1}}-\overline{X_i}$ represents the offset on the $x$-axis between two centroid points, $D_{t+1}$ represents the spatial centroid transfer distance from time $t$ to $t+1$, and ${\beta }_{t+1}$ represents the angle between the spatial centroid transfer direction from time $t$ to $t+1$ and the north direction.

3.2.2. Standard Deviation Ellipse

To further analyze the spatial distribution orientation of steel structure buildings in Lanzhou City, the standard deviation ellipse method [58] is employed. Through this method, the spatial characteristics of steel structure buildings, including central tendency, dispersion trend, and diffusion direction, are summarized [59]. This method calculates the standard deviation of $X$ and $Y$ coordinates from the average center as the starting point. The formula for calculating the error ellipse is as follows:

$$\begin{array}{c}tan\mathsf{\theta }=\left(A+B\right)/C \end{array}$$

(5)

$$\begin{array}{c}A=\left({\displaystyle \sum _{i=1}^n}{w}_i^2{\widetilde{x_i}}^2-{\displaystyle \sum _{i=1}^n}{w}_i^2{\widetilde{y_i}}^2\right) \end{array}$$

(6)

$$\begin{array}{c}C=2{\displaystyle \sum _{i=1}^n}{w}_i^2\widetilde{x_i}\widetilde{y_i} \end{array}$$

(7)

$$\begin{array}{c}\left\{\begin{array}{c}\widetilde{x_i}=x_i-\overline{X}\\ \widetilde{y_i}=y_i-\overline{Y}\end{array}\right. \end{array}$$

(8)

where $\mathsf{\theta }$ represents the orientation of the error ellipse, indicating the angle measured clockwise from the north direction to obtain the long axis of the error ellipse. $\widetilde{x_i}$ and $\widetilde{y_i}$ are the deviation values from the weighted average center to steel structure building i. $w_i$ is the weight, representing the fractional value of the steel structure building area in this study. The standard deviations in the $X$ and $Y$ directions are obtained according to the following formulas:

$$\begin{array}{c}\left\{\begin{array}{c}{\sigma }_x=\sqrt{\frac{{\sum }_{i=1}^n{\left(w_i\widetilde{x_i}cos\mathsf{\theta }-w_i\widetilde{y_i}sin\mathsf{\theta }\right)}^2}{{\sum }_{i=1}^n{w_i}^2}}\\ {\sigma }_y=\sqrt{\frac{{\sum }_{i=1}^n{\left(w_i\widetilde{x_i}sin\mathsf{\theta }-w_i\widetilde{y_i}cos\mathsf{\theta }\right)}^2}{{\sum }_{i=1}^n{w_i}^2}}\end{array}\right. \end{array}$$

(9)

In the standard deviation ellipse, the long semi-axis reflects the direction with a higher degree of dispersion, while the short semi-axis represents the direction with a higher degree of clustering [60]. The greater the difference between the two, the more pronounced the directional tendency of the data.

3.2.3. Compactness Index

To further analyze the spatial distribution orientation of steel structure buildings in Lanzhou City, the standard deviation ellipse method [61] is employed. Through this method, the spatial characteristics of steel structure buildings, including central tendency, dispersion trend, and diffusion direction, are summarized. This method calculates the standard deviation of X and Y coordinates from the average center as the starting point. The formula for calculating the error ellipse is as follows:

The compactness index is a measure of the shape characteristics of a region [62]. There are various methods for calculating urban compactness, and one of the most widely used formulas was proposed by Batty in 2001 [63]. The calculation formula is as follows:

$$\begin{array}{c}C=\frac{2\sqrt{\pi A_i}}{P_i} \end{array}$$

(10)

where $C$ is the compactness index, with a range of 0 to 1, and its value is positively correlated with the morphological compactness; $A_i$ represents the area (km²) of steel structure buildings during period $i$; $P_i$ is the perimeter (km) of the boundary outline of steel structure buildings during period $i$.

3.2.4. Patch Density and Landscape Percentage

Landscape percentage refers to the proportion of a specific patch type to the total area of the study region [64]. It is a measurement indicator for landscape composition and patch-type dominance [65]. Patch density (PD) is the number of specific patches per unit area [66,67], which can reflect to some extent the clustering degree of color steel plate buildings [68].

3.2.5. Weighted Kernel Density Analysis

Kernel density analysis is a method used to depict the clustering intensity of features across an entire area by calculating the density of point features around each output raster pixel [69]. The kernel density function effectively reflects the distance decay effect, where the values gradually decrease as the distance from point features increases [70]. In conventional kernel density estimation, the same density distribution is applied to all local regions containing feature points [71]. However, in this study, there is a significant variation in the area of color steel plate buildings in Lanzhou, and different-sized buildings have varying degrees of impact. To enhance the density distribution of larger buildings, a weighted kernel density estimation method is employed [72], with the formula being as follows:

$$\begin{array}{c}{p}_i=\frac{1}{n\pi R^2}{\displaystyle \sum _{j=1}^n}{k}_j{\left(1-\frac{D_{ij}^2}{R^2}\right)}^2 \end{array}$$

(11)

Here, $p_i$ represents the predicted density at point $i$; $R$ is the search radius; $n$ is the number of points within the search radius; $k_j$ is the spatial weight of point $j$; and $D_{ij}$ is the distance between $i$ and $j$.

The search radius R has a significant impact on the final results of the kernel density analysis [73]. If the value of $R$ is too small, it will result in density estimates distributed within a small range around the central point. This can lead to higher-density areas that appear scattered and lack overall coherence. Conversely, if the $R$ value is too large, it may fail to highlight local variations.

3.2.6. The Extreme Point Detection Model

In order to analyze the distribution and clustering areas of color steel plate buildings in Lanzhou more accurately, the method of extracting peak points using a digital elevation model (DEM) was employed to detect extreme points on the kernel density surface [74,75]. The specific approach is as follows: Firstly, through neighborhood statistics, the maximum density surface raster within the specified neighborhood range is obtained; then, utilizing raster calculation [76], the difference between the maximum value surface and the initial raster yields a non-negative value surface, where cells with a value of 0 indicate the locations of extreme points; subsequently, through a reclassification method, cells with a value of 0 are distinguished from those with non-zero values; finally, the cells with a value of 0 are extracted and converted into vector points. Subsequently, using graduated symbols, these extreme points can be visualized and analyzed.

3.2.7. Kappa Coefficient

To quantitatively evaluate the performance of the color steel-clad building boundary extraction model and better demonstrate the strengths and weaknesses of the self-attention mechanism U-net neural network model used in the experiments, the Kappa coefficient is employed to evaluate the extraction results obtained by the model [77]. The formula for the Kappa coefficient is as follows:

$$\begin{array}{c}Kappa=\frac{N{\sum }_{i=1}^n{x}_{ii}-{\sum }_{i=1}^n\left(x_{i+}{x}_{+i}\right)}{N^2-{\sum }_{i=1}^n\left(x_{i+}{x}_{+i}\right)}\end{array}$$

(12)

where $n$ is the total number of columns (total number of classes) in the confusion matrix; $x_{ii}$ is the number of samples correctly classified in the $i$-th row and $i$-th column of the confusion matrix; $x_{i+}$ and $x_{+i}$ are the total number of samples in the $i$-th row and $i$-th column, respectively; $N$ is the total number of samples used for accuracy assessment.

3.2.8. FLUS Model

The FLUS model (Future Land Use Simulation Model) integrates human activities, the natural environment, climate change, and other influencing factors to simulate future land use changes and scenarios, providing a basis for land change and management [78]. Color steel plate buildings are a minor classification of land use types, and the FLUS model can be extended to the field of color steel plate building clusters for predictive analysis [79]. The FLUS model is based on a Cellular Automaton (CA) model and an Artificial Neural Network (ANN) algorithm for simulation and prediction. The ANN algorithm can learn and predict change patterns, and ANN includes prediction and pre-training stages. It is composed of an input layer, hidden layers, and an output layer, with the calculation formula as follows:

$$\begin{array}{c}p\left(p,k,t\right)={\sum }_j{w}_{j,k}\times \frac{1}{1+e^{{-net}_j\left(p,t\right)}}\end{array}$$

(13)

where $p(p,k,t)$ represents the suitability probability of land use type $k$ on grid $g$ at time $t$; $w$ is the weight between the hidden layer and output layer, and ${net}_j(p,t)$ represents the signal received from grid $g$ at time $t$ at the $j$-th hidden layer.

3.3. Work Framework

In this study, as shown in Figure 4, the spatiotemporal data of color steel plate buildings in Lanzhou City are obtained by extracting the information on color steel plate buildings from 2013 to 2021. Due to the launch and operation of the GF-2 satellite since 2014, Google imagery needs to be utilized for 2013 data on color steel buildings to avoid errors and omissions, adopting manual vectorization methods for acquisition. From 2014 onwards, data on color steel buildings are extracted using GF-2 satellite imagery and neural network models. The extracted color steel plate building data were used to analyze the spatial distribution and extension direction of the color steel plate building cluster in Lanzhou City by using the methods of center of gravity distribution and transfer and standard deviation ellipse, to analyze the degree of aggregation of the color steel plate buildings in Lanzhou City based on the index of compactness, the density of patches and the percentage of the landscape, and the characteristics of temporal changes, and to calculate the hotspot areas of the spatial distribution of the color steel plate building groups in Lanzhou City by using the methods of kernel density analysis and extraction of extreme value points.

4. Results

4.1. Accuracy Analysis of Extraction

Based on high-resolution satellite imagery from the GF-2 satellite, the boundary information of color steel-clad buildings in the urban area of Lanzhou was extracted using a self-attention mechanism U-Net neural network model. During the model training process, the epoch was set to 64. With the increase in iterations, the accuracy of the self-attention mechanism U-Net neural network model gradually improved and stabilized, and the model converged. Adam optimizer was chosen, and the model was trained using augmented data samples. The performance of the model was evaluated using a confusion matrix. Figure 5 shows the training and testing curves of the self-attention mechanism U-Net neural network model.

As seen from

Table 1. Accuracy of the self-attention mechanism U-Net neural network model in color steel-clad building extraction.

Indicator	White Color Steel Buildings	Red Color Steel Buildings	Gray Color Steel Buildings	Blue Color Steel Buildings	Mean
Kappa	81.48	82.61	80.94	82.95	82.00

, the accuracy indicators of the extraction results for the four colors of color steel-clad buildings based on the self-attention mechanism U-Net neural network model are all above 80%. This demonstrates that using deep learning methods for obtaining remote sensing information of color steel-clad building clusters is feasible and efficient, and it saves a considerable amount of manpower and time costs.

4.2. The Overall Spatiotemporal Distribution Characteristics of Color Steel Plate Buildings

Color steel plate buildings are primarily concentrated in Chengguan District and Qilihe District of Lanzhou City; color steel plate buildings in Honggu District are mainly distributed in stripes along the Beijing-Tibet Expressway and Lanyong Highway. From the point of view of temporal and spatial changes, the number of color steel plate buildings in Lanzhou City has a certain decreasing trend, especially near Cuijia Datan, around Mataan Xincun, and Tumendun Village in Qilihe District, and there is an obvious decrease in Liujiabao Street, Yintan Street, and Kongjiaya Street in Anning District. However, there is a significant increase in the number of color steel plate buildings in the foothills of the northern part of the city. From the scale of each municipal district, the distribution of color steel plate buildings in Qilihe District gradually shows the trend of distribution along the traffic line, especially along Gonghu Highway, Lanhai Highway, and Gonglin Road 3 lines, and this distribution, in space, is mainly dependent on the villages, middle schools, and hospitals; the number of color steel plate buildings in Xigu District has little change in the whole of the period of 2013–2021, and there is a small amount of increase; Honggu District only increases a little in Hualong Street and Mining Street. In addition, far away from the urban area in the suburbs, the color steel plate buildings show sporadic distribution characteristics and are mainly attached to villages and highways without forming a large-scale gathering, as shown in Figure 6.

For the distribution number and area of color steel plate buildings in Lanzhou city as a whole, statistics are carried out. From

Table 2. The number and area of color steel plate buildings in different periods in Lanzhou.

Year	Quantity	Area
		Minimum/m2	Maximum/m2	Sum/km2	Average/m2	Standard Deviation/m2
2013	20,497	4.58	34,500.28	11.71	571.07	1285.28
2014	22,724	4.58	48,804.17	13.38	588.75	1479.34
2015	25,528	4.58	48,804.17	14.45	565.96	1425.79
2016	25,638	4.58	58,962.08	15.02	585.99	1513.06
2017	27,318	4.58	58,962.08	15.39	563.52	1474.15
2019	26,996	4.58	58,962.08	14.73	545.62	1455.52
2020	26,551	4.58	59,059.43	14.67	552.35	1549.01
2021	26,425	4.58	59,059.43	14.72	557.20	1564.28

, it can be seen that the number of color steel plate buildings in the study area first increases and then decreases, first from 20,497 buildings in 2013 to 27,318 buildings in 2017, an increase of 6821 buildings and an increase in area of 3.68 km², and then decreases to 26,425 buildings in 2021 and an area of another decrease of 0.67 km². It shows that after 2017, the government gradually paid attention to the improvement of shantytowns and urban-rural combinations within the urban area and began to gradually clean up the color steel plate buildings.

Table 3. The quantity and area of color steel plate buildings in various urban districts of Lanzhou.

District	Year	Quantity	Area
			Minimum/m2	Maximum/m2	Sum/km2	Average/m2	Standard Deviation/m2
Anning	2013	3397	11.61	29,078.83	1.56	457.53	1185.07
	2014	3498	13.48	29,078.83	1.67	476.42	1228.85
	2015	3545	13.48	29,078.83	1.65	465.43	1218.06
	2016	2886	19.54	29,078.83	1.48	511.20	1389.40
	2017	2752	19.54	29,078.83	1.46	528.72	1427.87
	2019	2434	16.58	29,078.83	1.28	527.78	1514.71
	2020	2349	16.58	29,078.83	1.26	535.36	1539.80
	2021	2196	13.16	29,078.83	1.23	559.74	1627.33
Chengguan	2013	3524	11.21	23,468.54	2.85	807.87	1552.97
	2014	4446	11.21	31,689.51	3.54	797.25	1549.21
	2015	5512	11.21	31,689.51	4.08	739.34	1459.27
	2016	6045	11.21	31,689.51	4.32	714.84	1423.64
	2017	7179	11.21	31,689.51	4.57	637.16	1320.72
	2019	7227	11.21	31,689.51	4.53	626.47	1299.61
	2020	7147	11.79	31,689.51	4.43	620.47	1289.78
	2021	7153	11.79	31,689.51	4.45	621.70	1289.25
Honggu	2013	3059	4.58	25,545.80	1.38	449.89	1472.78
	2014	3673	4.58	31,265.57	1.64	446.75	1484.82
	2015	4288	4.58	31,265.57	1.87	435.34	1441.76
	2016	4348	4.58	58,962.08	2.05	471.40	1755.40
	2017	4455	4.58	58,962.08	2.17	487.96	1762.36
	2019	4526	4.58	58,962.08	2.36	522.40	1897.96
	2020	4595	4.58	59,059.43	2.51	546.96	2105.28
	2021	4595	4.58	59,059.43	2.51	546.96	2105.28
Qilihe	2013	6715	7.86	33,086.66	3.89	578.87	1346.25
	2014	6869	7.86	48,804.17	4.13	601.43	1504.99
	2015	7637	7.86	48,804.17	4.36	571.31	1437.76
	2016	7601	7.86	48,804.17	4.44	584.66	1484.07
	2017	7955	7.86	48,804.17	4.37	549.81	1449.64
	2019	7727	7.21	33,086.66	3.72	480.93	1284.41
	2020	7368	7.21	33,086.66	3.51	476.02	1301.30
	2021	7453	7.86	33,086.66	3.46	464.23	1282.26
Xigu	2013	3834	10.31	34,500.28	2.08	543.48	1401.46
	2014	4268	10.31	48,804.17	2.50	584.91	1706.66
	2015	4579	10.31	48,804.17	2.60	567.68	1657.41
	2016	4790	10.31	48,804.17	2.84	592.85	1660.71
	2017	5016	10.31	48,804.17	2.93	583.54	1626.16
	2019	5125	10.31	38,241.70	2.92	569.37	1446.96
	2020	5134	10.31	48,276.52	3.02	588.90	1632.85
	2021	5069	10.31	48,276.52	3.14	620.43	1696.38

shows the number and area distribution of color steel plate buildings in each municipal district of Lanzhou City. First, comparing the number of color steel plate buildings in each area, it is found that the largest number of color steel plate buildings are distributed in Chengguan District and Qilihe District, and from the point of view of the total area, the largest year for the distribution of color steel plate buildings in Chengguan District is 2017, which reaches 4.57 km². Qilihe District reached the peak of the color steel plate building area in 2016, which is 4.44 km².

From the long-term trend, the overall trend of color steel plate building area in Chengguan District shows an increase, while the color steel plate building area in Qilihe District increases first and then decreases; from the point of view of the average area, the average area of color steel plate building in Chengguan District is larger than that in Qilihe District. On the whole, Chengguan District is the region with the largest distribution of color steel plate buildings. Anning District is the region with the least number of color steel plate building distribution and the smallest area, and the number of color steel plate buildings in Anning District also shows a decreasing trend, and the total area is also decreasing. The number and area of color steel plate buildings in Xigu and Honggu districts are both have increasing trends, and during 2013–2021, the total area of color steel plate buildings in Honggu district increased by 81.88%, and the total area of color steel plate buildings in Xigu district increased by 50.96%. Also, comparing the number of color steel plate buildings with the average area, the number of color steel plate buildings and the average area of color steel plate buildings in Xigu district are larger than that of the Honggu district in every period. Overall the number and area of color steel plate buildings distributed in Xigu District are larger than those in the Honggu District, but the growth trend of color steel plate buildings in Honggu District is more obvious.

4.3. Characterization of the Shift in the Center of Gravity of the Color Steel Plate Building Group

By calculating the weighted center of gravity of color steel plate buildings in Lanzhou, the shift distance and direction of the center of gravity can be obtained (see

Table 4. The shift distance and direction of the center of gravity for color steel plate buildings.

Year	X Coordinate	Y Coordinate	Center of Gravity Shift Distance/m	Clockwise and North Direction Angle/°	Expansion Direction
2013	379,729.00	3,995,964.85
2014	379,628.20	3,996,097.74	166.80	322.82	NW
2015	379,502.30	3,996,112.00	126.70	276.46	NW
2016	379,285.49	3,996,109.99	216.82	269.47	SW
2017	379,148.44	3,996,137.97	139.88	281.54	NW
2018	378,412.58	3,996,350.12	765.82	286.08	NW
2020	377,742.56	3,996,518.83	690.93	284.13	NW
2021	377,640.82	3,996,545.52	105.18	284.70	NW

). This analysis allows for an understanding of the spatial development patterns of color steel plate buildings. From 2013 to 2021, the cumulative shift of the center of gravity for color steel plate buildings amounted to 2005.20 m, with an average annual shift of 250.65 m, primarily expanding in the northwest direction. During the period from 2013 to 2014, the center of gravity shift for color steel plate buildings was at its maximum distance. Subsequently, from 2014 to 2016, the shift distance gradually decreased. Between 2017 and 2021, the shift distance showed a slight increase, maintaining around 280 m. In general, the shift distance of color steel plate buildings in Lanzhou City has been decreasing, indicating a tendency for the spatial distribution of color steel plate buildings to stabilize.

Further, the standard deviation ellipse analysis of color steel plate buildings within the whole of Lanzhou city is carried out, and the results of the error ellipse analysis for the 8-year period of 2013–2021 are obtained as shown in Figure 7. From the figure, it can be seen that the center of gravity of the spatial distribution pattern of the development of the color steel plate building moves along the path of the center of gravity in general, showing a tendency to move in the northwest direction. The expansion of color steel plate buildings in Lanzhou City has a clear direction, and the long axis of the ellipse is growing during 2013–2021, indicating that overall, the color steel plate buildings in Lanzhou City are expanding in both directions of the long axis. The flatness indicates the clarity of the expansion direction and the degree of centripetal force. From the ellipse flatness, the flatness of the standard deviation ellipse of the color steel plate building in Lanzhou City is generally maintained at about 0.8.

At the district scale, the strength of directional expansion for color steel plate buildings in different districts of Lanzhou varies, as shown in Figure 8. Among them, Honggu exhibits the strongest directionality, with the aspect ratio of the standard deviation ellipses around 0.84 for all eight periods. Following closely is Anning, with aspect ratios consistently above 0.72, reaching a maximum of 0.76 and an average of 0.75. Xigu also has relatively high aspect ratios, maintaining around 0.70. In comparison between Qilihe and Chengguan, Qilihe has a higher aspect ratio, with an average of 0.56, while Chengguan has the smallest aspect ratio, with an average of only 0.34. Therefore, in descending order of directional expansion among the five districts of Lanzhou City, it is Honggu, Anning, Xigu, Qilihe, and Chengguan. In terms of the standard deviation ellipse areas for each district, Honggu has the largest area, with an average of 384.97 km². This indicates a relatively extensive distribution of color steel plate buildings in Honggu, which is also sparsely populated due to a lower quantity of color steel plate buildings in that area. Looking at the change in orientation of the ellipses, the overall directional expansion of color steel plate buildings in Lanzhou City has remained relatively stable, as shown in Figure 8.

Overall, as shown in Figure 9, from 2013 to 2021, steel plate buildings in Lanzhou City were primarily distributed in the east-west direction, exhibiting a slight overall expansion trend. In the eastern region, the area of color steel buildings initially increased and then decreased. From 2013 to 2017, it increased by 3.08 km², but by 2020, it had decreased by 1.56 km². Looking at the expansion of steel plate buildings in each administrative district, Chengguan showed a noticeable expansion trend, especially in the southeast direction, where the area demonstrated an increasing trend. Over the 8-year period, the area increased by a total of 0.74 km². In the eastern, western, and northeastern regions, the area of steel plate buildings showed an initial increase followed by a decrease. The areas of steel plate buildings in Anning and Qilihe exhibited a trend of increase followed by a decrease. In Qilihe, the most significant variation in steel plate building areas occurred in the northwest direction. Honggu showed an expansion trend in the northwest, west, and southeast directions, while Xigu experienced the most noticeable increase in steel plate building areas in the eastern and northwest regions. In summary, steel plate buildings in Lanzhou exhibit significant spatial variations, with a substantial difference between the east and west. Moreover, the area of color steel buildings in the eastern region is much larger than that in the western region. This is attributed to the city’s development center being primarily located in the east, leading to the concentration of numerous urban villages and informal settlements in that area.

4.4. The Analysis Results of Fragmentation and Aggregation for Color Steel Plate Buildings

As can be seen from

Table 5. The relevant parameters of fragmentation and aggregation.

Year	Study Area Area/km2	Color Steel Plate Building Area/km2	Color Steel Plate Building Perimeter/km	Color Steel Plate Building Quantity	Area Proportion/%	PD/unit·km−2	C
2013	1625.92	11.71	1865.32	20,497	0.72	1750	0.0065
2014	1625.92	13.38	2105.08	22,724	0.82	1698	0.0062
2015	1625.92	14.45	2323.17	25,528	0.89	1767	0.0058
2016	1625.92	15.02	2366.41	25,638	0.92	1707	0.0058
2017	1625.92	15.39	2467.11	27,318	0.95	1775	0.0056
2019	1625.92	14.73	2389.32	26,996	0.91	1832	0.0057
2020	1625.92	14.67	2352.89	26,551	0.90	1810	0.0057
2021	1625.92	14.72	2346.03	26,425	0.91	1795	0.0058

, the color steel plate building area accounted for 0.72% of the total area of the study area in 2013, and then, with an overall trend of increasing, the proportion reached 0.91% by 2021. During that period, the highest proportion of color steel plate buildings was 0.95% in 2017, and then there was a slight decline, but the overall is maintained at more than 0.90%. The overall distribution of color steel plate buildings in Lanzhou City has increased in aggregation and decreased in fragmentation. As can be seen in Table 5, the C-index was 0.0065 in 2013 and decreased to 0.0058 in 2021. As can be seen from the definition of the compactness index, the larger the value of compactness, the more compact the morphology of the landscape patches. Also, the compactness of the color steel buildings in Lanzhou City from 2013 to 2021 is relatively small and has a decreasing trend, which indicates that the degree of aggregation of the color steel buildings in the whole range of Lanzhou City is relatively low but the extensibility is gradually increasing.

The results of fragmentation and aggregation analysis for various administrative districts in Lanzhou are presented in

Table 6. The relevant parameters of fragmentation and aggregation in different districts.

District	Year	Area Proportion/%	PD/unit·km−2	C
Anning	2013	1.80	2178	0.016
	2014	1.93	2095	0.016
	2015	1.91	2148	0.015
	2016	1.71	1950	0.018
	2017	1.69	1885	0.018
	2019	1.48	1902	0.020
	2020	1.46	1864	0.020
	2021	1.42	1785	0.021
Chengguan	2013	1.31	1236	0.015
	2014	1.62	1256	0.013
	2015	1.87	1351	0.012
	2016	1.98	1399	0.012
	2017	2.10	1571	0.011
	2019	2.08	1595	0.011
	2020	2.03	1613	0.011
	2021	2.04	1607	0.011
Honggu	2013	0.25	2217	0.018
	2014	0.30	2240	0.016
	2015	0.34	2293	0.015
	2016	0.37	2121	0.015
	2017	0.39	2053	0.015
	2019	0.43	1918	0.015
	2020	0.46	1831	0.015
	2021	0.46	1831	0.015
Qilihe	2013	0.98	1726	0.011
	2014	1.04	1663	0.011
	2015	1.09	1752	0.011
	2016	1.11	1712	0.011
	2017	1.10	1820	0.011
	2019	0.93	2077	0.011
	2020	0.88	2099	0.011
	2021	0.87	2154	0.011
Xigu	2013	0.56	1843	0.015
	2014	0.67	1707	0.014
	2015	0.70	1761	0.014
	2016	0.76	1687	0.013
	2017	0.79	1712	0.013
	2019	0.79	1755	0.013
	2020	0.81	1700	0.013
	2021	0.84	1614	0.013

. From the table, it is evident that Chengguan has the highest proportion of color steel plate building area, reaching 2.04% by 2021. In contrast, Honggu has the lowest proportion, accounting for only 0.46%. Analyzing the trend in area proportions, Anning shows an overall decreasing trend, while Chengguan, Honggu, and Xigu exhibit an increasing trend. Qilihe shows a fluctuating trend, with an initial increase followed by a decrease, reaching its highest proportion in 2016 at 1.11% and the lowest at 0.87%. Observing the changes in the number of color steel plate buildings per square kilometer, we found that Anning and Honggu show a decreasing trend in the PD index. In Honggu, the total area of color steel plate buildings increases, and the overall integrity of patches increases while fragmentation decreases. This indicates that the increase in color steel plate buildings in Honggu is mainly due to large-area buildings.

Examining the temporal changes in compactness for each administrative district, it is observed that the compactness of color steel plate buildings in Anning is gradually increasing, suggesting a higher degree of overall aggregation and lower expansiveness. In contrast, Chengguan, Honggu, and Xigu districts all exhibit a certain degree of decreasing compactness, indicating a more scattered distribution of color steel plate buildings and higher expansiveness in these three districts.

The results of kernel density analysis for color steel plate buildings in Lanzhou at different periods are shown in Figure 10. Overall, the areas with a dense distribution of color steel plate buildings in Lanzhou City are Qilihe and Anning, with the intersection between Anning and Qilihe being the most concentrated area for color steel plate buildings. The next densely distributed area is Chengguan, particularly in the northeast region. In contrast, the distribution of color steel plate buildings in Xigu and Honggu is relatively sparse, with no highly concentrated areas.

In terms of temporal changes, the overall kernel density maximum in Lanzhou city exhibits a developmental pattern of initially increasing and then decreasing. Upon comparison, the highest density value was around 500.079 in 2013, peaked at 561.291 in 2017, and decreased to 511.922 in 2021. Generally, the clustering trend of color steel sheet buildings in Lanzhou city is less pronounced than before. Simultaneously, examining the changes in hotspots of color steel sheet buildings in various districts reveals a noticeable aggregation trend in Chengguan, with Honggu also showing a similar trend, though less pronounced than Chengguan.

4.5. Extraction of Kernel Density Extremum Points for Color Steel Plate Buildings

Through detection, extreme points of kernel density for color steel plate buildings in Lanzhou were obtained, as shown in Figure 11. Xigu also exhibits some significant extremum points, indicating the presence of hotspots for color steel plate building distribution. However, due to large distances between points, there is limited connectivity, and the extremum points are relatively isolated, suggesting a loose distribution of color steel plate buildings in this area. In contrast, for Anning and Qilihe, the extremum points cluster together, with close distances between points, forming strong connectivity, especially in the northwest direction of Qilihe and the border area with Anning. In summary, color steel plate buildings in Lanzhou are mainly distributed in urban villages, surrounding towns, and areas around primary and secondary schools.

Through temporal comparisons, it is observed that the extremum points that were originally clustered in Anning and Qilihe show a trend of separation. In the northeast direction of Chengguan, extremum points are starting to exhibit clustering, and connectivity is also increasing. In Honggu, the position of extremum points has not significantly changed from 2013 to 2021, with only minor numerical variations, indicating that the aggregation of color steel plate buildings in Honggu is not apparent and there are no distinct hotspot areas. Notably, in the northern part of Qilihe, along Langongping Road and Gonglin Road, extremum points are gradually showing a tendency to cluster. By observing the kernel density map, it is found that these two roads are forming increasingly prominent bands. Considering the clear trend of the discrete distribution of color steel plate buildings in Qilihe, it is evident that color steel plate buildings in Qilihe are shifting from the urban area to the suburbs. Analyzing the reasons for this phenomenon, two main factors are identified. First, there are policy restrictions, as Lanzhou has gradually tightened policies on the construction of color steel plate buildings within the city, leading to a reduction or shift of such buildings within the urban area. Second, there are limitations related to land resources. In recent years, Lanzhou has undergone extensive urban development, resulting in significant consumption of land resources in the original main urban area. The rise in land lease prices has further driven the transfer of some color steel plate buildings, especially factories, to the surrounding suburbs. In recent years, areas with a gradual decrease in the distribution of color steel plate buildings include the vicinity of Yintan Street and Anningbao Street in Anning, where the density of color steel plate building distribution has noticeably decreased. In Qilihe, areas near Matan, Wanda Yuan, and Xiuchuan Street have seen a reduction. In Xigu, areas near Lintao Street, Chenguanying, and Xiping Village also experience a decrease in the distribution of color steel plate buildings.

4.6. Predictive Analysis of Color Steel Plate Building Clusters

The study utilizes data from two phases of color steel plate buildings within the built-up area of Lanzhou City, selecting six factors as driving forces affecting the distribution of color steel plates: vegetation coverage, population density, road distance, elevation, water system distance, and nighttime lighting. These factors cover a wide range of human and natural influences. Based on this, the suitability distribution probability of color steel plates is calculated (Figure 12). The closer the value is to 1, the stronger the suitability, and vice versa. Areas with strong suitability are mainly distributed near main roads, leaning towards urban areas. Conversely, regions with higher elevations and close to hilly areas have lower suitability, indicating the spatial distribution pattern of color steel plate buildings in terms of urban layout and topographical features.

Using 5-year intervals, data from 2016 and 2021 serve as the basis for prediction. The interactions and diffusion between spatial units are simulated using the Cellular Automaton model to predict the changes in color steel plates by 2026. To evaluate the simulation accuracy of the model, we use the Kappa coefficient. A Kappa coefficient greater than or equal to 0.75 indicates high simulation accuracy, while a value between 0.5 and 0.75 indicates moderate accuracy. In the experiment, the spatial distribution of color steel plate buildings in 2021 was predicted using data from 2016. The prediction results closely matched the actual 2021 results, with a Kappa coefficient of 0.797189, which is greater than 0.75, indicating that the FLUS model is suitable for simulating changes in color steel plates in this region. Furthermore, combining the actual results from 2016 with the 2021 prediction results, we used the Markov chain model to predict the future number of color steel plate buildings. According to the prediction results, the number of pixels of color steel plate buildings in the built-up area of Lanzhou City in 2026 will be 213,024. Using this pixel number as the prediction baseline and considering the six influencing factors, the distribution of color steel plates in the built-up area of Lanzhou City in 2026 is predicted (Figure 13). Compared to the actual number of 214,915 in 2021, the difference ratio in the predicted number of pixels is 0.88%. As shown in the figure, the number of color steel plate buildings in the built-up area of Lanzhou City slightly decreased over the years, remaining generally stable.

5. Discussion

Lanzhou, as an important provincial capital in the northwest region of China, faces challenges such as low economic development levels, weak industrial clustering, and limited financial capacity. From the current distribution characteristics of color steel-clad buildings, there appears to be a certain correlation between these structures and the socioeconomic development of the city. The clustering pattern of color steel-clad buildings is a product of specific stages of regional urban development. Delving deeper into urban issues, research and analysis provides reliable spatiotemporal data for urban governance. Through the analysis of spatial and temporal characteristics of the evolution of color steel plate buildings in Lanzhou City, it was found that the number of color steel plate buildings in Lanzhou City increases first and then decreases, while compared with other regions, the distribution of the number in each region has a significant difference. For instance, in Chengguan District, Honggu District, and Xigu District, there is a significant increase in the trend. Moreover, the expansion of the color steel plate building in Lanzhou city has obvious directionality, and the overall expansion direction is northwest and southeast. At the same time, the degree of aggregation of color steel plate buildings in the scope of the whole of Lanzhou City is low, and the expandability is also gradually increasing. From the spatial aggregation degree point of view, in Lanzhou city of Anning district and Chengguan district, the color steel plate building is obviously more aggregate, but from Lanzhou city as a whole point of view, there is a clear tendency to discrete, indicating that Lanzhou city for the rectification of the color steel plate building has a preliminary effect. The continuous expansion of large-scale color steel-clad buildings in Chengguan and Xigu districts within the main urban area of Lanzhou from 2013 to 2021 indicates rapid economic development in the city and a gradual transformation and upgrading of industrial structures. Meanwhile, the decreasing number of small-scale color steel-clad buildings in the Anning and Qilihe districts suggests that the urban management system in Lanzhou is gradually improving and the level of urban governance is continuously enhancing. This reflects the progression of high-quality urbanization in Lanzhou.

Existing research is more from the economy, market, policy, planning, environmental pollution, and other perspectives on the existence of the problem of analysis and the evolution of the empty pattern of the color steel plate building group [80,81]. The aggregation of the systematic research is insufficient [82], so researchers cannot quickly, effectively, and objectively grasp the development of the current situation of the color steel plate building group. And the existence of the problem of color steel plate building information can be obtained through the real-time high precision of high-frequency satellite imagery [83]. If we can clarify the correlation relationship and quantitative expression method of color steel plate building group [84], it will provide dynamic observation data, new methods, and new indexes for the rapid monitoring, scientific evaluation, and sustainable development research of similar urban color steel plate building groups.

Color steel plate building is a product of the development of the times. With the completion of rapid urbanization and the improvement of the urban environment, the color steel plate buildings will be gradually reduced. However, at present, the spatial and temporal correlation between color steel plate buildings and urban development is still an urban spatial problem that is worth studying, and future research needs to analyze its spatial distribution by using a more detailed remote sensing data system, especially to explore the impact of urban villages and shantytowns with color steel plate buildings on the lives and safety of the residents.

6. Conclusions

From the perspectives of time and space, as well as overall and local scopes, a comprehensive analysis of the spatiotemporal evolution patterns of the color steel plate buildings in Lanzhou was conducted. The results show that from the overall trend, the number of color steel plate buildings in the study area increased first and then decreased during 2013–2021, reaching its peak in 2017 and then showing a downward trend. The trends in area, density values, and corresponding peak points are consistent with the trend in number. This indicates that the government has strengthened the renovation of color steel plate buildings in urban villages and shanty towns. In terms of direction and centroid, the overall distribution shows a northwest-southeast distribution pattern. The annual centroid shift distances are all within 800 m, with stable migration amplitude and a smooth overall spatial distribution pattern. The fragmentation degree of the whole of Lanzhou city is decreasing, but some small areas show a slight increasing trend, indicating that the increase in color steel plate buildings in this area is mainly dominated by large buildings.

From a regional perspective, the two districts with the most color steel plate buildings are Chengguan and Qilihe. Overall, Chengguan has the most color steel plate buildings, while Anning District has the fewest, and the number in Anning District shows a decreasing trend over time. As for Honggu and Xigu, the number and area of color steel plate buildings in these two districts are increasing. The total area of color steel plate buildings in Honggu increased by 81.88%, and that in Xigu increased by 50.96%. In general, the number and area of color steel plate buildings in Xigu are greater than those in Honggu, but the growth trend in Honggu is more obvious. In addition, the areas with a dense distribution of color steel plate buildings in Lanzhou are Qilihe and Anning, especially the border area between these two districts. Next is Chengguan, which is mainly in the northeast. Since Lanzhou was named a National Civilized City in 2017, the distribution density of color steel plate buildings in its jurisdictions has been gradually decreasing, and the aggregation trend is not as obvious as before.

Specific conclusions are as follows: (1) In terms of quantity, color steel buildings in downtown Lanzhou show a trend of initially increasing and then decreasing, with the peak reached in 2017. However, there are significant differences in quantity among different regions. (2) In terms of direction, color steel buildings as a whole tend to shift towards the northwest, but the extent of this shift decreases with increasing years, gradually stabilizing. (3) In terms of distribution, the clustering degree of color steel buildings has slightly increased but remains relatively low. The fragmentation has slightly decreased, indicating a gradual enhancement in expansibility. This study utilizes multi-temporal data and focuses on spatial analysis to analyze the spatial patterns and evolutionary characteristics of color steel buildings in Lanzhou. The research results provide a reference value for the facility layout in Lanzhou, aiming to strengthen the planning and management of color steel buildings, guide reasonable layouts, and provide theoretical support for further research on sustainable urban development, promoting a low-carbon, green, and healthy development model.