This PDF file contains the front matter associated with SPIE Proceedings Volume 12710, including the Title Page, Copyright information, Table of Contents and Conference Committee lists.

Aiming at the lack of current forestry data sharing platform, combining the characteristics of regional forestry ecological data, this paper makes use of big data technology to make statistics and integrate different types of forestry data in various departments and platforms, and combines the front-end visualization technology to establish a data sharing visualization system for regional forestry ecological resources data. The regional forestry ecological resources protection and sharing platform adopts mapreduce under Hadoop framework for clustering analysis of C4.5 decision tree algorithm model, and the application system function development adopts Java language combined with SSM framework of spring+springmvc+mybatis. Besides charts, visualization tools also provide map display, and the tools used are Echarts and dojo.

This paper takes the geological disaster-prone area of Mayang County of Hunan Province as the test area.The results of the GNSS station time series calculation show that the vertical shape variable of the regional overall crustal plate ranges from 3.39mm to 4.85mm, indicating the overall upward trend;According to the calculation based on the InSAR long time series, the absolute vertical shape variable of the whole surface ranges from -41mm to 31mm, indicating the overall subsidence and local uplift trend.This paper explains the difference between the two methods and the reasons for the difference.It determines the difference is the difference between the bedrock plate movement and the bare surface absolute movement and forms a set of regional pure deformation field calculation methods. In the study area, the GNSS bedrock observation and InSAR surface observation results were respectively fitted with the global spatial difference. Then the difference was calculated to obtain the relative shape variable of the exposed surface movement compared with the bedrock, and the regional pure settlement field was -34.46mm~34.89mm. In this paper, it is proved by experiments that there are significant differences between the monitoring results of bedrock GNSS station and InSAR technology in the geological disaster research area. If the absolute surface shape variable observed by the InSAR method is not deducted from the relative shape variable, it is easy to misjudge the landslide hidden danger trend. To better reflect the surface subsidence trend, when studying the subsidence of geological disaster areas, it is necessary to subtract the bedrock shape variable from the analysis and to use the surface’s movement trend (pure deformation field) relative to the crustal bedrock as a variable.

The current SAR and optical image fusion methods still face the challenge of detail loss and low contrast. Therefore, this paper proposes a Sentinel-1/2 image fusion framework based on multiscale transform (MST) and generative adversarial network (GAN). The fusion framework based on MST and GAN is called MST-GAN, which can fully synthesize the information of source images and improve the contrast of fusion results. In this framework, a GAN low-frequency fusion strategy is proposed to fuse the low-frequency coefficients of multiscale decomposition, and the high-frequency coefficients are fused using the traditional “max-absolute” rule. In this paper, three multiscale decomposition methods are selected to verify the effectiveness of the proposed fusion framework. The qualitative and quantitative evaluation of the experimental results shows that the fused images under this framework fully integrate spectral, scattering and texture information, and have better ground objects differentiation.

With the continuous acceleration of urbanization and the rapid development of modern architectural technology, there are increasingly large public buildings and more complex indoor space patterns. At present, there are two main challenges in the analysis of interior floor plans: the first is that the objects in the image show scale diversity; the second is that due to the limitations of the convolution layer itself, ordinary convolutional neural networks have some difficulties in capturing global semantic information. In this paper, we propose a multi-task convolutional neural network model based on the multi-scale and polarized self-attention mechanism to improve the multi-scale information aggregation and global semantic information capture capabilities of semantic segmentation networks. Experiments on the public data set CubiCasa5K show that the proposed model can more accurately complete the identification of building components and the extraction of spatial area information in the floor plan.

The problem of parameter estimation widely exists in the field of surveying and mapping. In view of the special form of linear combination of nonlinear functions as the parameter model, the linear parameters are eliminated by variable projection operators, which reduces the dimension of the parameters to be solved and increases the possibility of convergence. Then, the Jacobian matrix of the residual vector is numerically approximated by the finite difference method, and the nonlinear function matrix in the iterative objective function is decomposed by QR and SVD. It can simplify the difficulty of matrix calculation and ensure the stability and efficiency of pseudo inverse matrix solution. The experiment is carried out through the full waveform decomposition of the height measurement LiDAR of ICEsat-1 satellite. The results show that under the same optimal solution, the method of approximating Jacobian matrix with finite difference effectively reduces the calculation time, improves the calculation efficiency, and provides a new idea for improving the separable nonlinear least squares variable projection algorithm.

During active and passive radar seeker tracking target, in order to effectively fuse information detected by two seekers for precise tracking，in this paper, BP neural network was used to classify the information into two different kinds, then fuzzy system was applied to adjust BP neural network’ learning speed based on target’s maneuver extent, at the same time, adjust the fusion center’s parameter based on target’s distance. Computer simulation shows that compared with traditional method, this method can conduct target tracking perfectly, it is proved to be ascendant．

Recent years have seen an increase in the scope of applications for airborne cameras as well as a diversification in their design. One of the most important steps in airborne camera imaging is the design of the optical system, which has a wide range of requirements. This paper describes the design of an anti-telepresence aspheric optical system that allows for a greater working distance. Optimize the optical system design to reduce the effects of factors such as aberration and distortion, resulting in a better imaging optical system.

As an active member of the earth ecosystem, vegetation is the center of material circulation and energy flow in the ecosystem, and plays an indicator role in environmental changes. Any changes in the terrestrial ecosystem will inevitably lead to certain changes in its type, quantity or quality. Vegetation cover is the most direct indicator of vegetation changes, and studying its distribution can provide a reliable scientific basis for vegetation planting and protection. With the continuous development of technology, a relatively complete landsat monitoring system has been formed, and abundant landsat image database has been established. Based on Landsat 8 time series remote sensing data from 2016 to 2021, normalized difference vegetation index was used to study vegetation growth and monitor the dynamic changes of vegetation growth in Jining City in recent 6 years. The results showed that the vegetation coverage was good in Jining except Weishan County. The vegetation growth trend of Jining City was increasing in Liangshan County, Wenshang County, Jiaxiang County, Rencheng District, Yanzhou District and Qufu City. Sishui County, Zoucheng City, Weishan County and Yutai County showed a partial growth trend. The greenness of vegetation in Jinxiang County showed a decreasing trend. The growth trend of Wenshang County was the strongest, while the decline trend of Jinxiang County was the strongest. The results showed that remote sensing monitoring can greatly improve the efficiency of monitoring vegetation growth. The vegetation growth changes have a certain relationship with the urban construction level and the urban ecological construction policy.

Real-time fire detection plays an important role in disaster prevention and mitigation, but we still know little about the development of fire occurrence. To this end, we propose a satellite hotspot monitoring method that combines polar-orbiting meteorological satellites and geostationary meteorological satellites to analyze satellite hot spots monitored in real time from July to November 2022 and the same period in 2021 in Hunan Province, China. The results show that the satellite hot spots from July to November 2022 have typical temporal and spatial distribution characteristics, and there is a clear correlation with the continuous development of the extreme drought area.

Remote sensing image segmentation is an important problem in the processing of remote sensing images. Existing methods for remote sensing image segmentation include supervised, weakly supervised and unsupervised approaches. Supervised and weakly supervised approaches require certain prior statistical knowledge on different regions in remote sensing images, while unsupervised approaches are able to accomplish the task of segmentation to a certain extent in the absence of such knowledge. The purpose of this paper is to realize an unsupervised image segmentation method that can be applied to remote sensing images. The approach utilizes a hidden Markov model to accurately describe the statistical distributions of the R, G and B components of pixels and the correlations among those of different pixels. The labels in a segmentation result are described by the states in the hidden Markov model and the segmentation with the maximum likelihood is obtained with a dynamic programming approach based on the Viterbi’s algorithm. Experimental results prove the feasibility of the proposed approach for segmentation. A comparison with state-of-the-art segmentation methods show that the proposed approach can lead to segmentation results with improved accuracy. The proposed approach is thus potentially useful for improving the accuracy of remote sensing applications that require segmentations of remote sensing images.

Water use efficiency (WUE) is an important parameter of interaction between ecosystem and hydrological cycle. Based on the land evapotranspiration and total primary productivity of MODIS products, the water use efficiency of the ecosystem was calculated, and the changes of vegetation and meteorological elements of the regional grid point were analyzed by using the method of one-dimensional linear regression and person correlation coefficient. The results show that WUE and gross primary productivity (GPP) in most regions showed an increasing trend. The regional distribution of increase and decrease accounted for 92% and 8% of the whole basin, and evapotranspiration (ET) showed a slight upward trend. Vegetation types have different effects on WUE and GPP changes, WUE and GPP of forest and shrub were the largest. In addition, WUE and GPP vary with altitude, and WUE has a maximum in the range of 1600-2200m altitude. WUE, GPP and ET have different responses to meteorological elements. Especially, temperature, precipitation, relative humidity and sunshine hours have effects on WUE by influencing GPP and ET. Under the climate scenarios of RCP2.6, RCP4.5 and RCP8.5, WUE and GPP will increase in the future. These studies provide scientific basis for exploring and protecting the carbon water cycle process of ecosystems.

In order to obtain the real ground point cloud, it is necessary to filter the original point cloud data to remove the non ground points. In the subtropical evergreen broad-leaved forest belt, the vegetation is dense, and the ground points are seriously sheltered. The ground points and non ground points are staggered. According to the characteristics of point clouds in the natural belt of the study area, this paper proposes a slope filtering method for dense vegetation terrain point clouds of airborne LiDAR. First, the virtual grid is introduced to segment the point cloud data, and the grid index of each point is calculated; Then calculate the elevation of the lowest point in each grid to calculate the height difference between each point and the lowest point in each grid. Then, the slope values of all points in each grid are obtained, and the local slope statistics are carried out to calculate the slope threshold; Finally, ground point filtering is carried out according to the adaptive slope threshold obtained by preset height difference threshold, k-means clustering and normal distribution. The flat terrain area and undulating terrain area are used for experimental analysis respectively. The results show that this method solves the problem of dense filtering of non surface points, and can not only remove dense vegetation but also retain terrain details.

UAV remote sensing technology is more and more used in topographic map mapping because of its efficient and flexible operation mode, and most of the current remote sensing UAV mapping adopts a single or the same type of UAV joint operation. Taking the Gongqing City Topographic Map Planning Project of Jiangxi Province as an example, this paper adopts two types of UAVs, fixed-wing and multi-rotor, to take aerial photography of the planned survey area according to the oblique and ortho ways, and summarizes the process of image data collection, image data processing, model accuracy verification, digital line drawing production and accuracy verification, and verifies that the image data obtained by fixed-wing and multi-rotor joint remote sensing mapping fully meets the requirements of 1:500 topographic map mapping through accuracy comparison. It shows that UAV remote sensing technology can be applied to the production of large-scale topographic maps.

GPS technology in the military, telecommunications, meteorology, surveying, navigation, remote sensing, geodesy, geodynamics, and astronomy, and many other fields to obtain a very wide range of applications, and promote the rapid development of science and technology, but also enrich the human science and culture life. GPS The exterior design has been gradually to the portable development, and gradually into the lives of people in. Therefore, the study of global positioning system has very important significance.This paper studies the principle of GPS positioning, design a microcontroller-based GPS positioning system, a total is divided into four modules, each module is a single-chip system, GPS positioning module, LCD12864 liquid crystal display module, key circuit module. Which is the use of single-chip system module STC89C52 microcontroller core module as an external peripheral circuits, GPS positioning module UBLOX-6M GPS positioning module. The use of MCU and GPS positioning module for data exchange to obtain the current location information and displays it on LCD12864. Eventually made into a portable GPS positioning device.

Aiming at the problems of low accuracy and missing of some details in road extraction by existing road extraction methods, this paper proposes a road extraction method integrating improved PSPNet. This method mainly improves the original PSPNet model by embedding SGE attention mechanism in the convolutional neural network stage to enhance the extraction accuracy. In pyramid pooling, Hybrid Dilated Convolution is used to replace global mean pooling to enhance the attention to details. In order to verify the effectiveness and feasibility of the proposed method, the improved model in this paper was compared with other common road extraction models. The experimental results show that compared with other methods, the proposed model has higher accuracy and integrity for road segmentation, and has certain practical application value.

The space star camera and altimeter set strict temperature rage and stability target, overmuch temperature fluctuate will make optical-mechanical deformation, decrease modulation transfer function (MTF), result in hot noise of Complementary Metal-Oxide-Semiconductors (CMOS) detector, decrease signal-to-noise ratio. The optical systems of the star camera and the laser altimeter are transmission lens and RC two-mirror lens respectively. The detectors of a star cameras are two CMOS; the laser altimeter includes six laser transmitters. To satisfy the requirements of thermal control index, thermal design multi-stage insulation technology for lens of star camera. Design thermal management for the laser altimeter’s heat sources. The thermal control system has been running for more than 2 years in space, under the extremely conditions in orbit, the temperature data show that the temperature fluctuation of lens is minimized to ±0.2℃, the temperature variation of CMOS and laser transmitters are less than ±0.1℃and ±0.2℃ respectively, which verify that the thermal control is appropriate.

Geospatial data shows an explosive growth trend, which leads to "massive data and difficult knowledge". Therefore, the transformation from data to spatial knowledge has become a scientific problem to be solved urgently. Firstly, this paper puts forward the technical process and basic ideas of Knowledge graph construction; Secondly, as geospatial knowledge has specific spatio-temporal characteristics, the representation method of geospatial data knowledge is described, and its spatio-temporal relationship is expounded; Finally, this paper takes some geospatial data in Sichuan Province as an example to study the problem of knowledge extraction and multi-source data fusion, and generate the corresponding knowledge graph. This research can promote the integration of geospatial data and semantic web technology, realize the spatialization of web text and the semantics of spatial data, and further improve the intelligent service level.

With the continuous development of remote sensing technology, the types of remote sensing image data are more diversified. More spatial information of images can be obtained by multisource fusion. In addition, the complementarity between sensors can effectively overcome the limitations of a single sensor in complex environments. The registration of optical image and SAR image is the key point in multisource image registration. Optical images have high resolution. But are vulnerable to the impact of harsh environments, resulting in the loss of spectral details. SAR images have strong penetrability to vegetation, cloud, ice and snow. But they are interfered by speckle noise. The imaging mechanism of optical image and SAR image is different, and the difference of gray information is large, which may lead to the performance failure of two kinds of image registration. To solve the above problems, in this paper, we propose a method of optical image and SAR image registration based on geometric constraints, which optimizes the feature descriptor locally through the spatial geometric structure characteristics between similar feature points. The experimental results show that the proposed method improves the matching performance compared with several state-of-the-art methods in terms of the matching accuracy.

The 3D photo-realistic model is characterized by high accuracy and large data volume. When visualizing these data in GIS applications, the whole model will be tiled and pyramid for efficient network transferring and loading. Then the tile loading strategy based on the distance from the viewpoint (hereinafter referred to as “the distance-based strategy”) will lead to the tiles appearing at the edge of the screen own better visual quality and large volume compared with those in the center of the screen, which mostly been the primary area that attracts human attention. To address this data redundancy problem, this paper proposes a novel tile loading strategy method based on the human eye's area of interest which provides better visualization quality on the central area of the screen while keeping the overall data volume loading on the web. Compared to the distance-based methods, the proposed method considering the angular deflection of the direction from the camera position to the center of the tile and the LOS (light of sight) direction, and recalculating the screen space error (SSE) determines which tiles in the pyramid will be selected. This method can choose higher-resolution tiles for the center region and choose lower-resolution tiles for the marginal region simultaneously. The experimental results show it increases the screen resolution of the center area while keeping the memory occupation steady. This research can facilitate the applications of 3D photo-realistic models in digital cities.

Residual fuselage deformation measurement is vital for comprehensive evaluation of aircraft structural strength and loads to ensure flight safety. However, the existing measurement methods have problems such as low automation and long time spent on data interpretation. This paper proposes a digital rapid measurement method for residual deformation of aircraft structures, which can quickly generate a measurement network construction scheme through digital station deployment and virtual simulation. Further, the automatic measurement and real-time data processing based on the remote control technology. The results show that the method can effectively assist the field decision and significantly improve the efficiency of flight test.

As a kind of transportation hub, bridge has important social and military value. This paper constructs bridge data sets of multi-source satellite remote sensing images, and realize the automatic bridge detection based on deep learning method to solve the problem that traditional method has bad results and poor stability. The effect of Faster RCNN and YOLO-v3 in remote sensing image detection of different types of bridges are mainly studied, and the advantages and disadvantages of two typical deep learning methods in bridge detection are analyzed through experiments. Through the experiment, the two methods can achieve the automatic bridge detection and the accuracy is all over 90%. In terms of detection accuracy, YOLO-v3 is higher than Faster RCNN(AP50), but its detection stability is lower than Faster RCNN. In addition, the training time and single image detection time of YOLO-v3 are both better than that of Faster RCNN.

The land around the Arctic is covered by extensive high-latitude permafrost. Affected by global warming, the Arctic permafrost is degrading. As a precise space-to-Earth observation technology, InSAR can reconstruct permafrost surface deformation with high spatial resolution and high temporal resolution. In this study, SBAS-InSAR was adopted to obtain time series of surface deformation due to permafrost degradation from 2017 to 2021 in the local area of northwest Banks Island. The results indicate that almost the whole study area is under the influence of permafrost degradation in continuous surface subsidence, and the water includes soil moisture and surface water system distribution is closely related to surface subsidence. The area with high soil moisture and dense surface water system distribution is more prone to significant surface deformation than other areas. This study provides a new reference for the vulnerability of the Arctic tundra under the impact of global warming and helps to make more scientific decisions on key issues e.g., global climate change, carbon neutrality, Arctic protection and development, and sustainable development.

In order to promote the ecological protection and the high-quality development of the Yellow River basin, and promote the coordinated protection of the resource-exhausted cities and the cultivated land resources in the basin, this paper took Jiaozuo City as the studied area, and carried out a comprehensive analysis of its urbanization level and the trend of the change in the cultivated land resources. The coordination degree model was used to calculate the coordination degree and the type of urbanization and the change of the cultivated land resources. The results of the research can provide suggestions for the urban transformation and development, the resource protection and the utilization of the cities with exhausted river basin resources.

Flooding is one of the most frequent, widely distributed, and serious natural disasters in the world, Spaceborne SAR is an important tool for flood disaster monitoring at present. In this paper, we designed a dynamic monitoring scheme of the reservoir area using multi-source spaceborne SAR images and used SDWI to extract the reservoir water area to monitor the water changes in the Changzhuang Reservoir and surrounding areas during the "7-20" heavy rainstorm in Zhengzhou. The monitoring results show that compared with July 15, the area of water body in Changzhuang Reservoir area increased massively on July 20, with the growth area reaching 1647.54m2, and the area of water body in the reservoir area decreased by 672m2 on July 22. The advantages of multi-source spaceborne SAR for monitoring reservoir water area are verified.

Compared to traditional multi-spectral information, nighttime light data has more intuitive features and is more sensitive to feedback on human-acquired trends such as urban sprawl and industrial development. Therefore, this paper uses Shanghai as the study area and combines the 30m resolution land use raster data with the monthly NPP-VIIRS data to determine the threshold value using the spatial comparison method of higher resolution image data. The threshold dichotomy was used to extract the urban built-up areas from the nighttime light images, and the error was only 0.85 when compared with the statistical data.

To obtain land surface temperatures data with high spatiotemporal resolution, A Geographically Weighted Durbin Model (GWDM) for Spatial Downscaling of Land Surface Temperatures is newly proposed in this study. The normalized difference water index (NDWI), the normalized difference built-up index (NDBI), and the normalized difference vegetation index (NDVI) were selected as scale factors to conduct downscaling experiments. Beijing and Zhangye were taken as the study area. Compared with the thermal data sharpening (TsHARP), the geographically weighted regression (GWR), the geographically weighted autoregressive (GWAR). The results indicate that the GWDM-based algorithm has better spatial texture and is closer to the real image. The determination coefficient (Beijing: 0.88, Zhangye: 0.91), mean absolute error (Beijing: 0.85℃, Zhangye: 1.06℃) and root mean square error (Beijing: 1.22℃, Zhangye: 1.57℃) are better than the other three methods.

It studies the correlation between the amount of cultivated land resources and the level of urbanization in the paper, explores the change relationship between the level of urbanization and the amount of cultivated land resources, carries out the research on the synergic relationship between the level of urbanization and the amount of cultivated land resources, and explores the internal law of urbanization deeply, so that the countermeasures and suggestions put forward in this paper are conducive to promote the new urban planning, improve the urbanization process and protect the cultivated land in the city of Zhuhai.

Accurate classification of wetland vegetation types is the basis for monitoring and management of wetland ecosystems. At present, the recognition accuracy of wetland vegetation classification based on UAV remote sensing images is not high. In this paper, to address this problem, 17 feature variables contained in spectrum, spatial geometry and texture were firstly selected; then the improved JM-Relief F algorithm was used to rank, judge and select the weights of the constructed feature variables, and remove the redundant variables; finally, the preferred 8 feature variables were classified by random forest, support vector machine and K-nearest neighbor models, respectively. The results showed that the R, G, and B band averages, asymmetry, GLCM homogeneity, GLCM contrast, GLCM correlation, and GLCM standard deviation obtained by feature optimization were the best features for classifying wetland vegetation, which could adequately represent the image features while improving the classification accuracy. The highest classification accuracy was achieved using the random forest model, with an overall classification accuracy of 88.80% and a Kappa coefficient of 84.3%, followed by the support vector machine and K-nearest neighbor models. The improved JM-Relief F feature selection algorithm combined with random forest classification model method can be effectively applied to the application of wetland vegetation classification and identification in the intertidal zone.

Using unmanned aerial vehicles (UAVs) to take aerial images and detect ships in harbors provides a new way of harbor monitoring. However, due to the limitation of computing resources and storage resources, ship detection on UAVs is very challenging, which puts a higher demand on model lightweighting. In this paper, we propose an efficient model lightweighting scheme based on knowledge distillation. We use two advanced large-scale models YOLOv7 and PP-YOLO as teacher models, and transfer the excellent detection ability of these two models to small-scale student models YOLOv7-tiny through knowledge distillation. This scheme not only greatly reduces the parameter scale and computation, but also retains the ship detection performance equivalent to that of large models.

The study selects three most representative megacity regions in China: Beijing-Tianjin-Hebei, Yangtze River Delta and Guangdong-Hong Kong-Macao greater bay area to explore the spatio-temporal characteristics of the development of urban agglomerations in China. To do so, we employ the built-up area data from 1990 to 2020, and calculate three types of urban expansion indicators such as urban expansion dynamics, landscape expansion index, and urban ranking clock theory. The comparison of these indicators in the past three decades is further conducted. Moreover, the correlation analysis helps to find the driving factors of the growth of the three major urban agglomerations. To be specific, (1) The Beijing-Tianjin- Hebei urban agglomeration has entered a period of rapid urbanization, the Yangtze River Delta urban agglomeration has entered a period of steady expansion, and the Guangdong-Hong Kong-Macao urban agglomeration has entered a period of slow maturation of expansion. (2) The expansion patterns of the three major urban agglomerations are mainly marginal expansions, while the enclave expansions are gradually decreasing; the expansion pattern is starting to move from diffusion to agglomeration, but the degree of agglomeration is still low, in the meantime, the degree of agglomeration in the Beijing- Tianjin-Hebei urban agglomeration is higher than that in the Guangdong-Hong Kong-Macao and Yangtze River Delta urban agglomerations. (3) The changes of city order scale of Beijing-Tianjin-Hebei city cluster and Guangdong-Hong Kong-Macao city cluster show a decreasing trend, while the changes of city order scale of Yangtze River Delta city cluster show an obvious increasing trend. (4) Socio-economic factors are significantly and positively correlated with urban expansion, among which the increase of the proportion of tertiary industry in GDP is the most relevant to urban expansion, while factors such as policy and geographic location are also important factors .

Although optical image has high resolution, it is extremely susceptible to the impact of harsh environments which will cause the loss of spectral details. SAR image has strong penetrating power to vegetation, clouds and snow, but they will be disturbed by coherent speckle noise. In addition, the imaging mechanism of optical image and SAR image is different, and the gray information is vastly different, which may lead to the failure of the performance of the two image registrations. In this letter, we propose an improved SIFT-Like image registration algorithm. Different gradient operators are used to calculate the gradient of optical image and SAR image respectively, and the corresponding scale space is constructed, and then the feature descriptor is constructed. Experimental results on optical and SAR image pairs show that the proposed algorithm in this letter has a significant improvement in the correct matching rate and registration accuracy compared with other algorithms.

The influence of urban heat island (UHI) on the environment, human health, and urban livability has been and will continue to be a source of concern for humanity. We examined the spatial distribution and intensity of UHI in a metropolitan city Shanghai utilizing long time series of satellite images and land use data. Then we properly and objectively monitored the UHI phenomenon and evaluated its driving variables. Using Landsat 5, 7, and 8 satellite data with China's Land-Use/Cover Datasets (CLCD), we retrieved land surface temperature (LST) and computed the Urban Heat Island Intensity Index (UHII) to measure UHI intensity. When investigating the correlation between impervious surface area and UHI area, we also made use of the Pearson correlation coefficient. The study found that the LST in Shanghai fluctuated upwards from 1990 to 2020, with the downtown, suburbs, and outer suburbs increasing by 5.394 degrees, 9.187 degrees, and 5.211 degrees, respectively, with the high-temperature zone expanding to the city's periphery year by year. Meanwhile, the UHII varies dramatically, the average increase of UHI in suburbs is 4.522 degrees, while the average growth of UHI in downtown and outer suburbs is only 0.728 degrees and 0.546 degrees, respectively. The Pearson correlation coefficient between UHI area and impervious surface area is 0.8698331. Shanghai's UHI has been rising year after year, owing to an increase in impervious surface area as a result of urbanization. The causes of the large changes in surface temperature and UHI have yet to be discovered and studied in depth. Overall, these findings show that urbanization has an impact on the establishment and intensity of UHI diffusion. To balance urban development with the urban thermal environment, more measures should be explored.