With the fast development and wide application of remote sensing technology, remote sensing images with higher quality are needed. However, it is difficult to directly acquire high-resolution, multispectral remote sensing images. To obtain high-quality images by integrating the information from different imaging sensors, pansharpening technology emerged. Pansharpening is an effective method used to obtain multispectral images with high spatial resolution. Many scholars have studied this method and obtained fruitful achievements. In recent years, deep learning theory has developed rapidly and has been widely applied in pansharpening. This study aims to systematically introduce the progress in pansharpening and promote its development. To this end, this study first introduced the traditional, classical pansharpening methods, followed by commonly used remote sensing satellites. Then, this study elaborated on the pansharpening methods based on deep learning from the perspective of supervised learning, unsupervised learning, and semi-supervised learning. After that, it described and analyzed loss functions. To demonstrate the superiority of the pansharpening methods based on deep learning and analyze the effects of loss functions, this study conducted remote sensing image fusion experiments. Finally, this study presented the future prospects of the pansharpening methods based on deep learning.

Water color represents the most intuitive visible perception of the color of water bodies that is jointly affected by substances such as suspended particulate matter, chlorophyll, and soluble organic matter. Water color is a water environmental parameter with a long history and plays a critical role in research on the ecosystem of inland and nearshore water bodies. With the progress made in colorimetric research, as well as hyperspectral imaging and satellite remote sensing techniques, the colorimetric method of water color has developed. This study systematically reviewed the colorimetric research progress of inland and nearshore water bodies and elaborated on the theories and practical applications of the colorimetric method from the angles of apparent optical properties (AOP) and inherent optical properties (IOP). Moreover, it presented the colorimetric processing method of satellite remote sensing data. The colorimetric method is a technical method for the quantitative expression of water color. It is also an important branch of water color research and an extension and supplement to the study of water color components, with a broad application prospect. To further improve the application of the colorimetric methods in inland and nearshore water bodies, it is necessary to enhance the construction of bio-optical datasets of water bodies in the future. Moreover, colorimetric studies should be conducted in two dimensions, namely AOP and IOP, and it is necessary to intensify research on domestic satellite-based colorimetric methods and increase the types of relevant water color products.

Targeting the subtropical climate characteristics of the Guangdong-Hong Kong-Macao Greater Bay Area, this study acquired the images of the experimental area from the TerraSAR-X Radar remote sensing satellite. Given the varying scale of the surface feature targets in the Radar satellite observation scenes, this study proposed an ENet convolution spatial pyramid pooling module (ENet-CSPP) model for surface feature classification. Since ordinary convolution can more effectively maintain domain information than atrous convolution, this study proposed a multi-scale feature fusion module based on convolution spatial pyramid pooling. Since there were a few training samples in the SAR remote sensing image dataset, this study combined the multi-scale feature fusion module with the lightweight convolutional neural network. The encoder of the ENet-CSPP network consisted of an improved ENet network and the convolution spatial pyramid pooling module. The decoder output surface feature classification images after the fusion of deep and shallow features. The quantitative comparison experiments were conducted on the GDUT-Nansha dataset. The ENet-CSPP model outperformed other models in three performance indices, namely pixel accuracy, average pixel accuracy, and mean intersection over union. This result indicates that the multi-scale lightweight model effectively improved the accuracy of surface feature classification.

Multi-scale segmentation is a key step in the information extraction of high-resolution remote sensing images. However, the evaluation of segmentation quality and the quantification of segmentation errors are still challenging. Based on boundary strength information, this study developed an unsupervised segmentation evaluation method of selecting the optimal scale parameter and elevating the local segmentation quality for multi-scale remote sensing image segmentation. Segmentation errors include over-segmentation and under-segmentation. This study modeled the two types of errors using normalized boundary gradient characteristics. The gradient information of patch edges was considered in the estimation of over-segmentation errors, while the intra-patch gradients were employed for the assessment of under-segmentation errors. To validate the proposed method, this study conducted an experiment on the evaluation of multi-scale segmentation results using two scenes of high-resolution remote sensing images. The segmentation evaluation results of the method coincided perfectly with the actual segmentation effects. The results indicate that the method proposed in this study can effectively reflect over- and under-segmentation errors.

Mangrove forests are periodically inundated by tidal water. This characteristic opens up an opportunity but also poses a challenge for the information extraction of mangrove forests using remote sensing technology. To explore the contribution of the red-edge band of GF-6 satellite data in information extraction of mangrove forests under the condition of random tides, this study investigated the southeastern Dongzhaigang area-the largest mangrove forest area in Hainan Province and obtained standard samples using the GF-2 satellite data. The reflectance spectral curves of typical surface features were constructed based on the standard samples and the GF-6 satellite data. Then, a baseline was established based on the bands strongly absorbed by vegetation, and the intertidal mangrove forest index (IMFI) applicable to the GF-6 satellite data was defined using the average reflectance of bands above the baseline. Meanwhile, the red-edge normalized difference vegetation index (RENDVI) was also established. The two indices were compared with commonly used indices, such as the normalized difference vegetation index (NDVI) and the normalized difference water index (NDWI), using box-whisker plots. Then, using the decision tree model constructed based on IMFI and RENDVI, information on typical mangrove forest in the study area were extracted. The precision of the extraction results was verified through comparison with visual interpretation results of the samples extracted from the GF-2 satellite data. The results show that: ① Because mangrove forests are periodically inundated by tidal water, the reflectance spectral curves of intertidal mangrove forests were relatively scattered between the standard spectral curves of water bodies and mangrove forests; ② IMFI and RENDVI can reflect the differences in the reflectance spectra of the red-edge and near-infrared bands and thus effectively separated the intertidal mangrove forests, mangrove forests, and water bodies; ③ The decision tree model constructed based on IMFI and RENDVI can effectively extract the distribution information of the mangrove forests, with an overall accuracy of 0.95 and a Kappa coefficient of 0.90. The introduction of the red-edge band plays an important role in the information extraction of mangrove forests and has great potential for application. This study can be used as a reference for the ecological applications of red-edge data from domestic satellites.

With 14 types of multi-feature information, such as spectrum, index, and texture, of remote sensing images from satellite Sentinel-2 as input and using the Bayesian optimization algorithm, this study designed the BO-XGBoost method used to automatically obtain the optimal hyperparameter combination. This method was successfully applied to the information extraction of cyanobacteria in Yangcheng Lake in 2021. The results show that: ① The optimal hyperparameter combination was obtained using the Bayesian optimization algorithm, and then the BO-XGBoost cyanobacteria classification model was established through obtaining. The training results performed well on the test and training sets, with an accuracy rate of up to 96.07%; ② The BO-XGBoost method was applied to the images used in the sample set. The comparison between the cyanobacteria identification results and the manual interpretation results shows that the two methods yielded roughly the same spatial distribution of cyanobacteria, with a lowest intersection over union (IoU) of 41.31%; ③ To evaluate the applicability of the BO-XGBoost method in other periods, images of other periods were selected for the information extraction of cyanobacteria. As a result, both BO-XGBoost and manual interpretation also yielded roughly the same spatial distribution of cyanobacteria, with a lowest IoU of 43.85%.

High-resolution meteorological data serve as an important data basis for fine-scale meteorological services. Using the hourly 2-meter air temperature grid data from January 2020 to March 2021 and the terrain factors such as altitude, longitude, and latitude, this study aimed to enhance the resolution of 2-meter air temperature grid data with a resolution of 1 km to 100 m through downscaling based on four machine learning methods, namely LightGBM (LGB), XGBoost (XGB), gradient boosting tree (GBT), and random forest (RF). Then, this study conducted the weighted fusion of downscaling results of different models. Finally, the downscaling results of different models were compared with the bilinear interpolation results, and the results are as follows. The results of each downscaling model were relatively consistent with the observational data. Compared with the bilinear interpolation results, the results of the LGB, XGB, and RF models had similar spatial structures but were more detailed. All downscaling models yielded the same spatio-temporal distribution characteristics of errors. Compared with the bilinear interpolation results, the data of the LGB, XGB, and GBT models showed significantly higher precision, and their root mean square errors (RMSEs) decreased by 5.2%, 4.1%, and 4.6%, respectively. Meanwhile, the RMSE after weighted fusion decreased by 5.9%, which was higher than that of any single machine learning model. The downscaling results of the LGB, XGB, and GBT models were improved to a certain degree compared with the bilinear interpolation results under different terrain conditions, especially in high-altitude areas (above 600 m). The correlation coefficients of results of the LGB, XGB, and BGT models and model based on weighted fusion increased by 0.45%, 0.40%, 0.63%, and 0.66%, respectively, and their RMSEs decreased by 9.1%, 8.0%, 12.7%, and 13.1%, respectively. These results indicate that the downscaling model based on the weighted fusion of different machine learning methods can both improve spatial resolution and maintain data precision and, thus, is suitable for downscaling research on 2-meter air temperature data in the study area. This study can be used as a reference for developing high-resolution data products.

To achieve accurate remote sensing scene classification, this study proposed a classification algorithm based on DenseNet feature hashing. First, dimension reduction was conducted for high-level semantic features output by a DenseNet through a fully connected layer. Then, normalized feature vectors were generated as the input of the classification layer using an activation function, and an end-to-end classification network was formed. Using the trained network as a feature extractor, the features of the activation layer of test data were mapped into binary hash codes. Finally, the remote sensing scene classification was conducted using support vector machine. The new algorithm was validated on public data sets UC Merced, WHU, and NWPU-RESISC45, and its classification effect was compared with that of multiple algorithms at three levels, namely the conventional local feature descriptor, transfer learning, and depth feature coding. The experimental results are as follows. The new algorithm had significantly higher classification accuracy than conventional algorithms based on mid- and low-level semantic features. Compared with the algorithm based on transfer learning, the proposed algorithm has fine-scale DenseNet feature mapping and accumulates elements used to determine core categories of images and, thus, is more suitable for the feature distribution of remote sensing images. Compared with the depth feature coding algorithm, the new algorithm has a simple feature structure, high classification accuracy, and strong transferability and extensibility and, thus, can meet the classification requirements of different remote sensing scenarios.

Earthquake-induced landslides are unnegligible secondary earthquake disasters and tend to cause severe casualties and property loss. Remote sensing identification of earthquake-induced landslides is an important means of the investigation and assessment of post-earthquake disasters. With GF-1 remote sensing images as a data source, this study identified the earthquake-induced landslides in the Xiongmaohai area in Jiuzhaigou using the object-oriented classification method. Specifically, the rule set for hierarchical identification of earthquake-induced landslides was constructed based on multi-scale segmentation and multi-conditional threshold classification. The aim is to fully utilize the features of ground objects, reduce the mixing of ground objects with similar spectra, and improve the identification precision of landslides. The identification results show that about 2.18 km² of landslide area was extracted near the Xiongmaohai scenic spot, with a general identification accuracy of up to 98.11%. Therefore, the method proposed in this study can quickly identify earthquake-induced landslides, with high identification accuracy and applicable identification rules, and, thus, can be used as a reference and basis for the emergency investigation and rapid loss assessment of post-earthquake disasters.

A rational land use plan is of great significance for avoiding high carbon emissions. The simulations of land use optimization from the perspective of low-carbon economy are conducive to the development of green economy and the scientific allocation of land resources. Taking Beijing as an example, this study incorporated the points of interest (POI) into the BP-ANN algorithm module of the FLUS model and verified the simulation accuracy of the improved model through comparison using the land use data of 2010 and 2020. On this basis, by coupling the Markov method and the order preference by similarity ideal solution (TOPSIS) method, this study simulated and analyzed the structure and spatial layout of land quantity in the study area in 2030 under the natural evolution scenario and the low-carbon economy scenario. The results show that: ① Compared with those of the original FLUS model, the Kappa coefficient and the overall accuracy of the improved model by incorporating POI data increased by 4.85% and 3.42%, respectively. These results indicate that the improved model had higher simulation accuracy. ② The simulation results verified that, under the natural evolution scenario, the carbon emission and the land for construction would increase by 7.70% and 7.68%, respectively, and the areas of farmland and grassland would continue to decline. ③ Under the low-carbon economy scenario, the carbon emissions would be reduced by 198.49×10⁴ t, the continuous expansion trend of construction land would be curbed, the occupation of grassland in low mountainous areas would be mitigated, and the area of forest land in the north would increase significantly. The results show that the simulation accuracy of the land use model would change with urban development elements and that the incorporation of POI data helped to provide more effective decision support for land planning. The low-carbon economy-oriented land structure adjustment and spatial layout optimization can be used as a reference for the rational use, planning, and layout of regional lands.

The current dead tree detection primarily relies on manual field surveys and, thus, is limited by forest topography, suffers a low detection efficiency, and is dangerous. Given these problems, this study proposed a YOLOv4-tiny dead tree detection algorithm based on the attention mechanism and spatial pyramid pooling (SPP) and improved the original detection model. First, the SPP structure was introduced after the Backbone part of the model to combine local and global features and enrich the feature representation capability of the model. Then, the original activation function LeakyReLU in the model was replaced with ELU, which made the activation function saturate unilaterally, thus improving the convergence and robustness of the model. Finally, the attention mechanism ECANet was introduced into the model to enhance the capacity of the network to learn important information in images, thus improving the performance of the network. The images of trees in a mountain forest of a scenic area in southern Liaoning were collected using an unmanned aerial vehicle (UAV). Then, dead trees in these images were detected using different models. The detection results show that the improved algorithm had a detection accuracy of 93.25%, which was improved by 9.58%, 12.57%, 10.54%, and 4.87% than that of the YOLOv4-tiny, YOLOv4, and SSD algorithms and an algorithm stated in literature [8], respectively, and achieved the effective detection of dead trees.

Currently, the high-quality fusion of SAR and optical images is a hot research topic. However, the significant radiation difference and weak gray correlation between SAR and optical images greatly reduce the fusion quality. In this regard, this study proposed a SAR and optical remote sensing image fusion algorithm that coupled non-local self-similarity and divergence. First, images were decomposed in the frequency domain. Then, the non-local directional entropy and divergence were used as characteristic parameters to guide the fusion of low- and high-frequency components, respectively. Finally, the fusion components were reconstructed to obtain fusion images with clear structural features and rich spectral information. The comparative experiments verified the effectiveness of the proposed algorithm in fusing SAR with optical images and its superiority in maintaining structural features and reducing spectral distortion.

Aiming at the problems of poor extraction effect and slow extraction speed of traditional road extraction methods in the information extraction of roads from high-resolution remote sensing images, this study proposed a new information extraction model based on improved Deeplabv3+. In the new model, the combination of the MobileNetv2 backbone feature extraction network with the Dice Loss function effectively balanced the contradiction between the precision and speed of road information extraction from high-resolution remote sensing images. As a result, high extraction precision was achieved while meeting timeliness requirements by reducing model parameters. The experimental results based on the open-source road information extraction dataset show that: ① The road information extraction model proposed in this study was feasible for high-resolution remote sensing images, with overall accuracy of up to 98.71%; ② In terms of the information extraction speed, the new model had an average frame number of 120.05 and parameter amount of only 5.81 M. Therefore, the new model was more lightweight lighter than original models, meeting the timeliness requirements. Therefore, the model proposed in this study meets the timeliness requirements by greatly reducing the parameter amount while ensuring high extraction accuracy. This study provides a new philosophy and method for improving the accuracy and speed of road information extraction from high-resolution images.

The soil roughness of cultivated land is an important element affecting the monitoring of agricultural information, such as soil moisture, microwave remote sensing observation, and plant growth. Soil roughness is generally interpreted according to field photos. However, such interpretation suffers some shortcomings such as low efficiency and anthropogenic effects on processing results. UAV low-altitude remote sensing is sensitive to surface relief. To explore the precision of the soil roughness determined using UAV data, this study employed UAV photogrammetry to photograph the surface and then compared the photogrammetry results with the data obtained using a gauging plate for soil roughness. The results show that the close-range photogrammetry had mean absolute errors of mainly 0.4~1.2 cm, a mean relative error of 6.16%, and a root mean square error of 0.40 cm. Therefore, UAV-based point cloud photogrammetry could be effectively applied to the measurement of surface roughness, and a smaller sampling area is associated with more accurate soil roughness.

The Hadamengou gold deposit in Baotou City, Inner Mongolia is an important large gold deposit in the Wulashan-Daqingshan metallogenic belt, with great prospecting potential. To give full play to remote sensing technology in geological prospecting, this study extracted the mineralization alteration information of the Hadamengou gold deposit from remote sensing data of different satellites. Based on the spectral characteristics of alteration minerals in the mining area, this study proposed a comprehensive processing method that extracted iron staining information from the Landsat8 OLI and WorldView-3 data, hydroxyl information from the ASTER and WorldView-3 data, and carbonation information from the ASTER data through principal component analysis. As a result, two alteration zones were delineated based on the distribution patterns of alteration anomalies and the geological map analysis of the mining area. By combining the study results of the ore-controlling structures, it is believed that the metallogenic hydrothermal processes of the Hadamengou gold deposit were closely related to structures. This study can provide a reference for the prospecting for the same type of gold deposits in the Wulashan-Daqingshan metallogenic belt and can guide the peripheral prospecting of the Hadamengou gold deposit.

Xiong’an New Area is a national new area. It has a low groundwater level and close water exchange between the zone of aeration and the saturated zone, with the upward recharge of groundwater increasing the water content in soil. On this basis, with remote sensing images as the data source, this study carried out object-oriented land classification for the study area, extracted the vegetation information by mask, and further extracted the soil moisture information of the vegetation area using the temperature vegetation dryness index (TVDI). Then, by combining the geological and geomorphic characteristics of the palaeochannels in the area, as well as visual interpretation, this study identified the palaeochannels in the study area and verified them in the field. Finally, it reconstructed the paleodrainage system of the study area. The results are as follows: ① The method proposed in this study can effectively extract information on the paleodrainage system in the study area; ② The distribution of the current surface water bodies in the study area is quite different from that of the paleodrainage system; ③ The comparison between the land classification results and the paleodrainage system interpretation results shows that the paleodrainage system was mostly distributed in present construction land, which is present as rural residential areas in remote sensing images. 50 m, 100 m, and 200 m buffer zones were set in the paleodrainage system areas, and then a intersection analysis was made for the buffer zones and the land classification results. The results show that the proportion of construction land in the buffer zones is significantly higher than that of construction land in the whole region. This result indicates that there exists a certain correlation between the distribution of the paleodrainage system and villages.

In 2020, a flood disaster occurred throughout Anhui Province due to the persistent heavy rainfall during the super-long plum rain period. To quickly and accurately extract the flood inundation ranges and provide scientific support for flood prevention and disaster relief, this study selected the pre-disaster and mid-disaster Sentinel-1A/SAR data of the Chaohu Lake and Huaihe River basin in Anhui Province. After rapid data preprocessing, this study extracted information about water bodies in the plains and mountainous areas using the Sentinel-1 dual-polarized water index (SDWI) method and topographic factors. Then, it established a monitoring process for flooded areas. Using this process, this study extracted the flood inundation ranges of the Chaohu Lake and Huaihe River basins on July 27, 2020 using the pre-disaster and mid-disaster synthetic aperture Radar (SAR) data. The results are as follows. The SDWI was superior to the backscattering coefficient in the extraction of information about water bodies. The Chaohu basin had a flood inundation area of 524.8 km² on July 27, and the Baishitian River subbasin was the most severely inundated, followed by the Xihe River subbasin. In the flood flowing and storage areas of the Huaihe River basin within Anhui Province, the flood inundation area of four cities along the Huaihe River basin decreased in the order of Huainan City, Fuyang City, Lu’an City, and Bengbu City. The results of this study show that the Sentinel-1A-based monitoring process of flood inundation areas established using SDWI and topographic factors has high accuracy, applicability, and timeliness for plains and mountainous areas and is convenient for the timely monitoring of flood disasters in these areas.

The Zhada earth forest, located in Zhada and Pulan Counties, Tibet, is composed primarily of weakly consolidated to semi-consolidated clastics s of the Tuolin and Xiangzi formations. This area forms a unique geological landscape consisting of peaks and ravines due to the long-term erosion by rivers and rain. To further explore the tourism resources in the Zhada earth forest distribution area and fully reveal the scientific and aesthetic values of the study area, this study carried out the geological interpretation of the study area mainly based on the GF-1 satellite remote sensing images, with the interpretation focusing on the Xiangzi and Tuolin formations constituting the earth forest landscape, as well as ophiolites and tectonic melanges reflecting plate subduction. Based on the interpretation results and the 3D interpretation environment of the Aerial Geophysical Remote Sensing Multivariate Data Processing and Product Display Platform, this study extracted information on typical geological landscapes in the study area, including earth forests, various rocks, and fault structures. The remote sensing technology helped delineate the distribution range of the earth forest more accurately. The 3D display platform enabled the more vivid display of the geological relics that represented the dramatic changes in the regional evolution history, such as earth forests, oceanic crust remnants, and unconformities. The application of modern information technology can provide strong support for the landscape planning of the Zhada Earth Forest National Geopark.

With the large-scale exploitation and utilization of coal resources, the geological environmental problems of coal mines have been increasingly severe, thus restricting social and economic development. This study aims to ascertain the disaster status of the collapse of the mined-out areas in Anhui Province, analyze the changing trend of the collapse areas, and summarize the countermeasures and methods for the collapse areas. With 2016—2017 remote sensing images of Anhui Province obtained from domestic GF satellites as an information source, this study conducted the processing, interpretation, and analysis of the remote sensing images and field surveys using the 3S technology (the collective term of remote sensing, global position system, and geographical information system). The results are as follows: ① The total area of the collapse areas in 2017 was 396.62 km², accounting for 0.28% of the land area of the province; ② The growth rate of the area and quantity of collapse areas decreased compared with those in previous years; ③ A set of countermeasures and four treatment methods were proposed. As revealed by the results, the 3S technology-based remote sensing monitoring of the mine environment in Anhui Province can be used to produce high-quality data and extract relevant data information macroscopically, efficiently, and accurately, thus greatly improving the treatment efficiency of mine geological disasters. This study will provide technical support for the restoration, treatment, and sustainable development of the collapse areas of coal mines in the future.

On September 5, 2022, a Ms 6.8 earthquake occurred in Luding County, Ganzi Prefecture, Sichuan Province, inducing numerous landslides. This study collected the pre- and post-earthquake images from the GF-2 and GF-6 satellites, as well as the DEM data of Luding. Then, using the object-oriented method, the stepwise optimization multi-scale segmentation method, and the nearest neighbor classification method, this study extracted the landslide information according to the spectrum, thematic index, geometric texture, and topographic features of the objects in the experimental area. The overall identification accuracy of pre- and post-earthquake landslides was 92.3% and 95.4%, respectively. The comprehensive analysis of the distribution of pre- and post-earthquake landslide landslides shows that 23.91 km² of new landslides were induced by the earthquake. This study summarized the distribution characteristics of post-earthquake landslides through the spatial statistical analysis of seven topographic factors. The results are as follows: ① The post-earthquake landslides were mainly affected by the Xianshuihe fault zone, and they show a banded distribution along rivers and a lamellar, dense distribution along the hillsides and valleys near the fault zone; ② Compared with the historical landslides, the new landslides have a relatively stable elevation range and a large slope range. Moreover, there is a significantly negative correlation between the area of the post-earthquake landslides and the surface roughness.

The Axi mining area in Xinjiang has a complex geographical environment. The long-term exploitation of mineral resources has caused severe ground subsidence and deformation in the mining area, as well as safety hazards of mining and production and the destruction of the surrounding ecological environment. This study aims to further investigate and analyze the spatial-temporal variation characteristics of the ground subsidence and the patterns of surface deformation in the Axi mining area. To this end, this study first calculated the land subsidence using the small baseline subset-interferometric synthetic aperture Radar (SBAS-InSAR) technique based on the 127 scenes descending Sentinel-1A images acquired from February 9, 2017 to April 25, 2021. Then, it compared the subsidence monitoring results obtained using the InSAR technique with the leveling results for verification. Finally, this study analyzed the spatial-temporal variation characteristics of land subsidence in the Axi mining area in recent five years and investigated the driving factors for the land subsidence. The results show that the surface deformation of the Axi mining area showed a roughly stable trend and significant local subsidence throughout the monitoring period. The main factors affecting the ground subsidence included mineral exploitation, geological structure, precipitation, and the impoundment of open-pit mines. This study will provide a scientific basis for ground subsidence monitoring and the future proper exploitation of underground minerals in the Axi mining area.

The forest area of Tibet ranks among the top in China, and the forest resources in Tibet play an important role in water conservation and ecological service. Therefore, it is of great significance to assess the assets of forest natural resources in this region. However, existing products and statistical data related to forest cover fail to meet the demands for the assessment of forest natural resource assets in this region, and it is necessary to explore a fine-scale forest classification method suitable for this region. Based on the cloud computing platform Google Earth Engine (GEE), this study constructed the temporal, spatial, spectral, and auxiliary feature sets of the forest coverage in Motuo County using the Landsat8 remote sensing images of 2015 and 2020, as well as field survey data, and the basic geographic data. Then, it conducted forest classification using the random forest (RF) and classification and regression tree (CART) algorithms. As indicated by the accuracy evaluation of the assessment results obtained using the two algorithms, the forest classification results of 2015 and 2020 obtained using the RF algorithm had relatively high accuracy, with overall classification accuracy of 0.88 and 0.87, respectively and Kappa coefficients of both greater than 0.8. The analyses of the areal and spatio-temporal characteristics of forest classification results show that: ① Motuo County had a total forest area of 34 000 km² in 2015, with a forest cover rate of up to 84.63%, which was 2% less than that in 2020; ② The forest resources in Motuo County are dominated by broadleaved forests, which are mainly distributed in Yarlung Zangbo Grand Canyon and low-altitude areas and accounted for 72.27% and 75.37% of the total forest area in 2015 and 2020, respectively. Coniferous forests accounted for 25.96% and 23.19% of the total forest area in 2015 and 2020, respectively and are concentrated in high-altitude areas, such as the Namcha Barwa and Gyala Peri peaks. This study determined the spatio-temporal distribution of the forests in Motuo County in 2015 and 2020 by developing a spatio-temporal-spectral classification method. It can provide a reference method for calculating specific forest cover indices SDGs and fill the gap of forest data of small zones. The obtained monitoring data will provide data support for the natural asset assessment and ecological function evaluation in Motuo County.

The rapid detection of soil salinity using remote sensing technology can scientifically guide the soil salinization control and the rational development of oasis agriculture. Based on 95 soil samples from the oasis of the Weigan-Kuqa River delta, this study established four soil salinity estimation models of multiple linear regression, partial least squares regression (PLSR), support vector machine regression (SVR), and random forest regression using the spectral index, band reflectance, and the measured soil salinity. Then, it conducted the remote sensing inversion for the spatial distribution pattern of the soil salinity in the study area using the optimal estimation results. The results are as follows: ① Nine spectral factors that were significantly related to soil salinity were screened using the all-subsets regression method, with correlation coefficients of all above 0.5 (P < 0.01). Among them, the correlation coefficient between salinity index SI-T and the soil salinity was the highest (0.648); ② The comparison of estimation precision show that the fitting effect of the four inversion models was in the order of random forest regression > SVR > PLSR > multiple linear regression. Among these models, the random forest model had the best fitting precision. Its training and validation sets had coefficients of determination(R²) of 0.870 and 0.766, respectively, with relative percent deviation (RPD) of 2.792 and 2.105, respectively, both of which were greater than 2. These results indicate that the random forest model had a good inversion effect and stable estimation capacity; ③ According to the inversion results of the random forest model, grade I and II zones account for 41.62% and are distributed in the cultivated land area inside the oasis; grade III, IV, and V zones account for 56.41% and are primarily distributed in the desert and the desert-oasis ecotones. Therefore, compared with conventional statistical models, the random forest modeling method can yield significantly better estimation effects in the inversion of soil salinity. This study can be used as a reference for the monitoring of soil salinization in oases in arid areas.

Given the land surface types and atmospheric features of the Heihe River basin, this study calculated the surface emissivity of the study area using the ASTER Global Emissivity Database and the vegetation cover method (VCM) and estimated the atmospheric water vapor content using the improved multilayer feed-forward neural network (MFNN). Moreover, by establishing the coefficient lookup table of input parameter groups, this study developed an ASTER data-based split-window algorithm for the remote sensing inversion of land surface temperature. To validate the applicability and accuracy of the split-window algorithm, this study elevated the algorithm using the measured site data on the land surface temperature of the Heihe River basin in 2019 and MODIS instruments. Compared with the site data, the results of the split-window algorithm had root mean square errors of 1.81~3.01 K. In the cross-validation using the MODIS instruments, the split-window algorithm had relatively small errors and deviations, with root mean square errors of 1.11~1.75 K. Overall, the accuracy of the land surface temperature obtained from the inversion using the split-window algorithm can meet the needs of meteorological and climatological studies. Moreover, the development philosophy of the split-window algorithm can be used as a reference for similar thermal infrared sensors.

This study aims to explore the optimal remote sensing salinization detection index (SDI) model for the inversion of soil salinization in the Alar reclamation area. Based on Landsat8 OLI remote sensing images and field measured data, this study built the remote sensing SDI models using the salinity index (SI), the normalized difference vegetation index (NDVI), the modified soil adjusted vegetation index (MSAVI), and the surface albedo. Then, using these models, this study extracted the soil salinization information on the Alar reclamation area and verified the model precision. Finally, this study determined the optimal remote sensing-based SDI model through comparative analysis. The results are as follows. The four types of remote sensing-based SDI models SDI1 (SI-NDVI), SDI2 (SI-MSAVI), SDI3 (SI-Albedo), and SDI4 (Albedo-MSAVI)had general classification precision of 83.45%, 69.78%, 53.23%, and 71.94%, respectively. Model SDI1 was the most suitable for the inversion of the degree of soil salinization in the Alar reclamation area. Models SDI2 and SDI4 can be utilized as a reference for soil salinization monitoring of the Alar reclamation area. As revealed by the inversion results of the SDI model, the reclamation area is dominated by non-saline and lightly saline soils, with heavily saline soil and saline soil primarily distributed in the northeast and southeast. Model SDI1 established based on SI and NDVI has high accuracy in extracting the soil salinization information of the Alar reclamation area and can be used as the remote sensing-based SDI model for the inversion of soil salinization in reclamation areas. This study can provide an effective technical reference for the control and prevention of soil salinization.

Areas with power transmission lines have been frequently struck by flood disasters in recent years. Therefore, forecasting the water table depths in these areas is critical to the safety of these areas. This study forecasted the water table depth using remote sensing satellite products and observed meteorological and hydrological data. Based on the meteorological and hydrological data, this study forecast the daily and monthly water table depths using the long short-term memory (LSTM), gated recurrent unit (GRU), long short-term memory-seq2seq (LSTM-S2S), and feedforward neural network (FFNN) models. The results indicate that the LSTM-S2S and FFNN models delivered the best and the worst performances, respectively. Meanwhile, the LSTM, GRU, and LSTM-S2S models performed well in forecasting both daily and monthly water table depths, with their forecasts of daily water table depths having a higher coefficient of determination (R²) and a Nash-Sutcliffe efficiency coefficient (NSE) than those of monthly water table depths. Therefore, the method presented in this study can be used to forecast the future daily and monthly water table depths in areas with power transmission lines.

Ecological carrying capacity is an important indicator used to measure the stability of an ecosystem. The spatial-temporal change analysis of the ecological carrying capacity can help understand the changing trend of a regional ecological environment and serve as a comprehensive reference for the evaluation of ecological management and restoration, research on the overall sustainable development of an environment, and the optimization of land resources. Targeting the arid and semi-arid regions at the northeastern margin of the Ulan Buh Desert, this study constructed a comprehensive index evaluation system of ecological carrying capacity based on the actual ecological conditions of the regions and Landsat remote sensing images as the data source. Then, this study determined the spatial-temporal distribution and evolution pattern of regional ecological carrying capacity and made a driver analysis of the change in the ecological carrying capacity from the angles of rainfall, temperature, and land use changes. The results show that the ecological carrying capacity of the northeastern margin of the Ulan Buh Desert showed a first decreasing and then increasing trend from 1990 to 2020. The irrigated areas north of the Yellow River continued to expand to the desertification areas in the southwest. As a result, the percentage of the area with relatively high ecological carrying capacity increased greatly, while the area with high ecological carrying capacity decreased. The change in the ecological carrying capacity of the irrigated areas was mainly affected by land development and utilization, followed by temperature and rainfall. In contrast, the ecological carrying capacity of the desertification areas south of the Yellow River was mainly at a moderate level, which was shifted to a low level in large areas before 2010 and was restored to a moderate level in 2020. The change in the ecological carrying capacity of the desertification areas was greatly affected by temperature, followed by rainfall and changes in shrub and grass vegetation cover.

Net ecosystem productivity (NEP) represents the carbon sequestration capacity of a regional ecosystem. Based on the Google Earth Engine (GEE) platform, this study analyzed the temporal and spatial variations in the NEP of the Three-River Headwaters Region (TRHR) from 2001 to 2020 based on the Moderate Resolution Imaging Spectrometer (MODIS) and meteorological data and revealed their relationships with climate factors. The results are as follows: ① The TRHR had an important carbon sink function, with carbon sink areas accounting for 99.89%; The carbon source areas in the TRHR were primarily distributed in the northwest, accounting for only 0.11%. The NEP of the TRHR decreased gradually from the southeast to the northwest and differed significantly among different ecological areas; ② The NEP of the TRHR showed an upward trend overall in the past 20 years, with an annual increasing rate of 1.13 gC/(m²·a), indicating huge carbon sequestration potential; ③ The area of zones whose NEP showed an upward trend accounted for 95.05% of the total area. Ecological engineering construction significantly improved the NEP of vegetation. As a result, the carbon sink function gradually increased and was highly stable; ④ The TRHR had an annual average NEP of 120.93 gC/(m²·a), and the NEP was positively correlated with the annual precipitation but negatively correlated with average annual temperature and annual solar radiation. The warm, humid climate and the ecological engineering construction contributed to the carbon sink function of vegetation in the TRHR. This is of great significance for improving the carbon sink value of the terrestrial ecosystem and achieving the peak carbon dioxide emissions and carbon neutrality of China.

Accurate and efficient quality inspection and database updates of cadastral data are essential for natural resource management. The current cadastral data management faces problems such as the low efficiency of quality inspection and updates, difficulty in meeting the demand for dynamic supervision, and small application scopes of relevant methods. To solve these problems, this study proposed a method framework based on spatio-temporal knowledge graphs. Moreover, with cadastral data and remote sensing images as data sources, this study constructed a spatio-temporal knowledge graph targeting the quality inspection and update workflow of cadastral data by designing conceptual and data layers and inference rules. Finally, experiments on the method proposed in this study were conducted using seven parcels of land in Changsha. As a result, the common errors in the process of quality inspection and updates were solved, and the method proposed in this study was proven to be more efficient than common methods.

The inconsistency of multi-source geographic data in scale, geometric position, and attribute cause difficult data fusion and update. This study proposed a fusion and update method for geographic data based on geometric and attribute matching. First, the candidate set was acquired using the generalized Voronoi diagram, thus effectively improving the acquisition efficiency and reducing the impact of unrelated targets on the candidate set. Then, the matching analysis of point, line, and plane data was made using key techniques such as geometric and attribute matching. Finally, based on the matching results, the incremental data were extracted from the reference geographic information data, followed by fusion and update of target data. The experimental results show that the method proposed in this study can efficiently identify and extract incremental data and serves as a reference for the innovative exploration into the update mode of monitoring data.

With the vigorous development and in-depth application of civilian and commercial satellites, the space-based remote sensing application demands of different users become increasingly complex. However, the current space-based remote sensing services face problems such as single application mode, weak pertinence, and insufficient flexibility. Based on the analysis of the major application demands of various remote sensing users, this study proposed a space-based remote sensing application service mode based on cloud + terminals. This mode covers eight subcategories in three categories, whose characteristics and application process were analyzed and formulated individually. Last, this study presented the potential applications under two typical scenarios, namely single-person independent application and multi-person collaborative application. The results of this study will lay a foundation for the development, construction, and optimization of various space-based remote sensing ground systems and further improve the space-based remote sensing service capabilities for different users, different application demands, and different application scenarios.