Remote sensing application has become an indispensable support for a number of industries, and the development of remote sensing systems in various countries is surging. Trusted remote sensing information for supporting law enforcement applications has become a new requirement from users. Based on the Group on Earth Observations (GEO) advanced transformation of “open data” to “open science” facing digital economy at Canberra GEO Ministerial Summit in 2019, it is held that Earth Observation Systems need to study new interoperability mode to supervise the construction of knowledge hub, adapt to the requirements of digital economy for data quality, and improve the ability to support scientific decision-making. The unique advantages of remote sensing information have attracted attention and participation of large international digital technology enterprises. New technical route of remote sensing image processing supported by new generation of digital technology is discussed, the whole parameter processing should be carried out under the whole chain from acquisition to information according to the concept of remote sensing science so as to realize quantitative remote sensing, and the quality assurance system of inspection and certification to fulfill reliable remote sensing should be built. It is further held that the construction of GEOSS under open science needs to process remote sensing information concerning the whole chain using cloud computing and big data with the consideration of data quality and result reappearance to promote new interoperability among remote sensing systems so as to ensure that trusted remote sensing information will be promoted to reassure decision-makers.

The key to improving the ability of independent innovation lies in talents and education. In order to cultivate high-quality innovative talents in the field of “quantitative remote sensing”, this summer course held an academic salon for graduate students in combination with the frontier issues of quantitative remote sensing that the trainees were concerned about. The academic salon held four academic salons aiming at the academic problems in the theory, method, technology and application of quantitative remote sensing. The mechanism of radiation transfer, the decomposition of hyperspectral remote sensing mixed pixels, the application and service of UAV quantitative remote sensing were discussed. Among them, the academic salon of “radiation transfer mechanism” mainly discussed the progress and limitations of radiation transfer theory from Maxwell’s equations to microcosmic physics. “Hyperspectral remote sensing mixed pixel decomposition” academic salon mainly focused on two aspects of endmember variability research, namely, the discussion of spectral variability within the same endmember category and the spectral similarity between different endmember categories. Participants deeply discussed the theory and method of how to eliminate the unmixing error of hyperspectral remote sensing mixed pixel. The “vegetation fluorescence remote sensing” academic salon mainly discussed the application progress of solar-induced chlorophyll fluorescence (SIF) remote sensing, the process and mechanism of SIF excitation from leaf to sensor and its mechanism. Participants discussed five main issues in depth. “UAV quantitative remote sensing application and service” academic salon focused on UAV quantitative remote sensing and multi aircraft cooperative networking earth observation and remote sensing application service in complex scenes. Participants believed that UAV quantitative remote sensing application and service has broad prospects in the future. In each academic salon, a graduate student made a keynote speech. The participants would apply for a speech around the topic, discuss and question, and elaborate personal views or supplement relevant research progress information on relevant progress. The host would make a summary speech. The online academic salon provided a new academic platform for graduate students to exchange and discuss the frontier progress of quantitative remote sensing. The academic salon attracted many interested graduate students and other personnel to participate through live broadcast of bilibili station, and expanded the dissemination of quantitative remote sensing knowledge.

Regional surface subsidence caused by the development and use of urban underground space is a major hazard endangering the safety of Beijing-Tianjin-Hebei city cluster. This paper briefly reviews the development history of interferometic synthetic aperture Radar (InSAR) technology, systematically summarizes the progress of applying gravity recovery and climate experiment (GRACE) satellite in underground water reserve, illustrates multiple factors containing subsidence, and finally ascribes the subsidence to multiple fields of underground space. Under the new hydrological background of the interaction between South-to-North Water Diversion and mining of underground water, InSAR-GRACE technology is a brand-new means for studying the impact of underground space evolution on land subsidence. Based on InSAR-GRACE technology, this paper rediscovers the regional water circulation laws, quantifies the contribution of multiple fields to subsidence evolution, proposes the surface response research framework for the evolution of underground space, and reveals the formation mechanism on the surface subsidence response model, thereby establishing an emerging risks prevention and control early warning mechanism for underground space security and realizing scientific regulation and control of the region.

Crop seeds are the most basic and original means of production in the planting industry. The selection of high-quality seeds directly determines the economic and production benefits in the agricultural production process. Hyperspectral imaging technology emerged in the 1980s, which has the characteristics of non-destruction, rapid imaging and “integration of atlas”. Previous studies of crop seeds using hyperspectral imaging technology mainly focused on the variety identification, vigor detection, and seed quality of crop seeds. In this paper, based on the previous research, the authors summarize and refine the data processing models, which include such methods as partial least square method, Ada-Boost algorithm, limit learning machine (ELM), random forest (RF), support vector machine (SVM), and artificial neural network (ANN). To sum up, the purpose of this paper is to provide the best spectral range, sample types, noise reduction methods, feature band extraction, model building and other aspects as the basis for various types of crop seed research, and to provide suggestions for future research direction.

Spatiotemporal fusion image can meet the needs of large-scale, high-precision and rapid change of surface cover monitoring, and hence has been widely used in environmental, hydrological and agricultural monitoring and other fields. In this paper, based on different types of remote sensing data, the authors propose a method of remote sensing image fusion based on particle swarm optimization (PSO) and linear mixed pixel decomposition. First of all, the PSO method is applied to the calculation of the end-point reflectance through the statistics of the different range of the end-point reflectance, and then the remote sensing image fusion is realized by considering the difference of the end-point reflectance between the high and low spatial resolution images and the space-time weight. Finally, the comparison with the existing methods shows that the proposed method can effectively improve the accuracy of the prediction image produced by data fusion. The root mean square error and spatial structure similarity index predicted in this paper are better than the results of the enhanced spatial and temporal data fusion model(ESTDFM). Therefore, the proposed method would be of great value for the study of land cover change monitoring.

In order to give full play to the advantages of hyperspectral spatial information and peak density clustering algorithm in dividing remote sensing image features, this paper proposes a hyperspectral image classification method based on the combination of hyperpixel and peak density features. Superpixel segmentation technology makes full use of the spatial and spectral information of hyperspectral images, dividing hyperspectral images into hyperpixels, extracting the gray value of hyperpixels as an important feature of peak density classification, selecting the spectrum with the highest peak density as the spectral cluster of the whole image, using the visual and hyperpixels as the basic units of classification, and then obtaining the pixels and hyperpixels respectively. The membership relation is obtained by the difference between spectral clusters. Finally, the image classification is completed by combining the membership degree. Experiments show that the proposed algorithm takes less time than other methods under the condition of ensuring the highest classification accuracy, and meets the requirements of hyperspectral image information extraction and analysis.

Remote sensing image plays an important role in ecological restoration, geological disaster monitoring and prevention of mining area; nevertheless, due to the complex environment of mining area, the obtained remote sensing images of the mining area contains different kinds of intensity noise, which affects the subsequent image interpretation and analysis to a great extent. In this paper, the study ideas of image edge detection and noise suppression are effectively fused together, and the improved denoising method (LWT-IEDPO) of remote sensing image in mining area based on the fusion of lifting wavelet thresholding (LWT) and improved edge dectction of Prewitt operator (IEDPO) is proposed. According to the basic principal of the new method proposed in this paper, firstly, lifting wavelet transform is done for the original remote sensing image; under the condition that the low-frequency wavelet sub-band is left untreated, a two-parameters thresholding function model is designed for adaptive noise suppression of high-frequency sub-bands, and the remote sensing image after initial denoising is obtained by the operation of inverse lifting wavelet transform. Secondly, for the purpose of effectively enhancing the details of the filtered remote sensing image, the detection template of the classical Prewitt operator is extended to 6 directions of 0°, 30°, 60°, 90°, 120° and 150°, and the corresponding detection results fusion rules are proposed. The improved Prewitt operator is put forward to extract the image edge information of the filtered image, and the edge and non-edge image are obtained. Then, the visual effect non-edge image is further improved by adopting the improved Pal-King fuzzy algorithm. Finally, the goal of high definition restoration of the original remote sensing image is realized by the superimposition of the enhanced non-edge image and edge image. Based on MATLAB language, the proposed remote sensing image processing method is compared with the classical hard thresholding model, soft thresholding model and two existing improved wavelet thresholding algorithms; in addition, the two indices of peak signal to noise ratio (PSNR) and edge protection index (EPI) are used to conduct quantitative analysis and comparison of the performance of the above algorithms. The study results show that the goal of effectively filtering of noise remote sensing image can be realized effectively, and the overall performance of the proposed algorithm is better than that of the other four algorithms.

With the concept of “smart city”, remote sensing target detection has gradually become an important way for town planning, construction and maintenance. In order to characterize the differentiating remote sensing features of different cities and solve the problem of uneven generalization ability in the model on objects of different scales, this paper proposes a pyramid structure search method based on hybrid separation and convolution. Firstly, this paper analyzes the spatial distribution characteristics of the remote sensing image dataset, also constructs a multi-receptive field hybrid convolution search space based on its characteristics, and then trains the weights in its sub-network. Secondly, the number and structure of feature extraction units are searched cyclically with the help of reinforcement learning algorithms for the convergent loss value sequence. Finally, when the architecture reward function is stable, the corresponding architecture parameters and weight matrix are fixed, so that the cross-scale information of the image can be adaptively fused on the test data to improve the positioning accuracy of similar targets at different resolutions. The average accuracy of the network searched by this method on the DIOR remote sensing dataset is 78.6%, which is 6 percentage points higher than that of CornerNet, 1.6 percentage points higher than that of Cascade R-CNN, and the accuracy of small objects is 2.1 percentage points higher than that of Cascade R-CNN. The optimization ability of multi-scale architecture search in remote sensing target detection was confirmed.

The detection and recognition of traffic signs is an important part of the intelligent driving navigation system. However, due to the shortcomings of low accuracy, high time complexity and poor robustness, the traditional method cannot meet the current needs of intelligent driving. Therefore, a method for detecting and recognizing road traffic signs of UAV images based on Mask R-CNN is proposed. Firstly, a set of high-quality UAV images road traffic sign data sets are produced. Then, based on the statistics of 200 labeled landmarks features, the region proposal network (RPN) structure anchor boxes width-to-height ratio and initial parameters in Mask R-CNN are improved to make it better applied to UAV images road sign scenes. Finally, the precision-recall (PR) curve and mean average precision (mAP) are used for accuracy evaluation. The experimental results show that the anchor boxes width-to-height ratio is better when the ratio is 1∶1, 1∶2, 1∶3; and that the average detection accuracy obtained by this method is 98.33%, which is higher than the accuracy of Faster R-CNN and YOLOv3, indicating better effectiveness.

Cooling tower emissions pollute the atmosphere. Using high-resolution remote sensing images to detect cooling towers can provide decision-making data for the treatment of exhaust emissions. Aiming at the problems such as low detection accuracy and slow detection speed of traditional algorithms in high-resolution remote sensing image object detection, the authors improved the RetinaNet by adopting a sampling-free mechanism to detect the cooling towers. First, Images in dataset were labeled as working cooling towers and resting cooling towers. Then, based on the number of object categories in the dataset and the proportion of positive samples in training, the bias term of the last layer in the classification subnetwork and class-adaptive threshold were determined. In addition, the regression loss was used to determine the adjustment ratio of the classification loss to avoid loss functions to be dominated by numerous background examples. Finally, ResNet50 was used to extract image features, and the FPN module was used to generate a multi-scale convolution feature pyramid. Detection boxes regression and category confidence calculations were performed for each layer of features. The results show that, for cooling tower detection on high-resolution remote sensing images, the proposed algorithm can improve the detection accuracy while ensuring the detection speed compared with RetinaNet, which proves the effectiveness of the proposed algorithm.

In order to accurately segment the building object of high-resolution remote sensing image, this paper proposes a multi-task learning method based on Unet network. Firstly, boundary distance map is generated from the ground-truth map of the building object remote sensing image; the boundary distance map, original remote sensing image and ground-truth map together are regarded as the input of Unet network. Then, based on the ResNet network, a multi-task network is built by adding the building object prediction layer and the boundary distance prediction layer at the end of the Unet network. Finally, the loss function of the multi-task network is defined, and the network is trained by using Adam optimization algorithm. Experiments on the Inria aerial remote sensing image building object dataset show that, compared with the full convolutional network combined with the multi-layer perceptron method, the intersection-over-unions of VGG16, VGG16+boundary prediction, ResNet50 and this method have been increased by 5.15, 6.94, 6.41, and 7.86 percentage points, and the accuracy has been increased to 94.71%, 95.39%, 95.30%, and 96.10% respectively,which ensures that the building object of high-resolution remote sensing image can be segmented effectively.

Using high-resolution satellite remote sensing images to extract the boundary of the built-up area is of great significance for urban expansion monitoring and urban development planning. In order to obtain high-precision and high-resolution built-up area data, this study uses the NDBI index and artificial visual interpretation methods to construct remote sensing image datasets of urban built-up areas and uses traditional machine learning methods and four deep learning methods including PSPNet semantic segmentation network to extract the built-up area of Sentinel-2 images. The training results show that the PSPNet network has the highest accuracy for the built-up area extraction (IOU of the training set is 79.5%). This paper employs Overlapsize method to optimize the extraction results of PSPNet, which further improves the accuracy of the built-up area extraction. The IOU on the training set reaches 80.5%, and the IOU on the test set reaches 83.1%. Compared with the traditional machine learning method, the method of PSPNet + Overlapsize has practical application significance in built-up area extracting.

The change detection of urban buildings through remote sensing images can help researchers grasp the planning and implementation of urban buildings comprehensively, and assist urban managers to find and investigate illegal buildings. This paper proposes a method for urban building change detection that combines Unet with IR-MAD. This method first uses weighted small Unet and IR-MAD to detect suspected change pixels in remote sensing images, and then fuses the suspected change pixels detection results based on voting to find out change pixels. For optimizing the change pixel areas, morphological operations are performed to remove speckle noise and fill holes in the changed pixel area. Finally, non-building change areas are removed based on the shadow characteristics of the building to obtain building change detection results. Experiments show that this method can detect building changes in remote sensing images more accurately than using only Unet or IR-MAD.

Rice is one of the most widely planted food crops in China. Therefore, timely and accurate rice identification and monitoring is of great significance to the national food security and the evolution of agricultural land spatial pattern. In this study, multi-temporal Sentinel-2A multispectral images, vegetation indices, vegetation abundance and Landsat 8 derived LST on the critical period of rice phenology were used. The CNN, SVM and RF classifiers were applied to extracting the paddy rice and finally the paddy rice map was obtained. The result shows that using multi-temporal and multi-source remote sensing data with the CNN algorithm can effectively extract rice information in high heterogeneity region. The overall accuracy of rice classification and Kappa coefficient are over 92% and 0.90 respectively. This study has demonstrated the potential of using moderate spatial resolution images combined with CNN to map the paddy rice in highly heterogeneous area.

In order to acquire the appropriate remote sensing data to obtain the plant growth information and identify the planting types of crops, the authors chose Quanjiao of Chuzhou in Anhui Province as the research area and the SAR (GF-3) data and optical remote sensing data as the data source to fuse optical data with the SAR data and make a comparative study of data classification results, optical and SAR data classification results and the data fusion results so as to conduct crop type identification. The comparison of the data of classification results reveals that SAR data can be used as a good auxiliary optical image for crop planting types in crop recognition. The fusion of SAR data and optical remote sensing data has a good identification effect on crops in the research area.

Wheat is a densely planted crop, and the planting volume per acre is nearly 20 kg. The plant density of winter wheat will directly affect the final yield. Therefore, real-time monitoring of wheat plant density is an important way to ensure wheat yield. At present, the main method for obtaining the plant density of wheat is mainly manual measurement, which is time-consuming and laborious. In this paper, the DJ inspire 2 UAV is equipped with a Zens X4S camera to obtain high-resolution visible light images of wheat planting areas, extract wheat coverage based on UAV images, and establish the relationship between plant density and plant density so as to achieve rapid acquisition of wheat plant density based on UAV image. Experiments show the following results: ① Using the improved HSI color model to extract wheat coverage improves accuracy and extraction efficiency compared with traditional visual estimation, manual counting and other classification methods, and overcomes differences in lighting conditions and shadows of different sorts of UAV images influences. ② There is a high correlation between wheat coverage and plant density at the seedling stage, overwintering stage and turning green stage. Among them, the correlation coefficient R² between the coverage based on drone image and the plant density of wheat are 0.737 9, 0.898 1 and 0.897 6 in three growth stages. The verification results of the relationship model using Niutengyu Village samples show that the inversion results based on the established relationship model also have a good correlation with the measured values, and R² reaches 0.919 8.

For the traditional remote sensing image data extraction, the accuracy of the cage information is low, and there exist the problems of “different object with the same spectrum” and “salt and salt” noise. Based on the Gaofen-2 satellite (“GF-2”) data, this paper proposes an improved double-branch network model cage information extraction method. The model uses the dense connection block to extract the spatial feature information of the cage on the spatial coding path, obtains the global context information of the cage quickly by using the global average pooling on the global coding path, and finally enriches the detailed information of the cage space through feature fusion. And deep discriminant feature information improves the extraction accuracy of the cage. The method has achieved scores of 87.37%, 72.56%, and 82.47% on the three evaluation indicators of precision, IOU, and F1, respectively, which are 7.82, 4.12, and 4.64 percentage points higher than the traditional method with the highest accuracy, respectively. The deep learning model has also achieved an increase of 8.43 and 8.69 percentage points in IOU and F1. Experiments show that the method can meet the extraction work of sea cage culture area, and can provide technical support for the regulation and regulation of offshore sea cage culture.

Since the 21st century, the outbreak of cyanobacteria in the Taihu Lake has seriously affected the development and utilization of local water resources. Based on Landsat8 imagery, this paper analyzes the spectral reflection characteristics of non-cyanobacteria water and cyanobacteria water. Cyanobacteria water shows strong reflectance characteristics in the near-infrared band, but the reflectance characteristics in the blue, green, red and shortwave-infrared bands are the same as those in non-cyanobacteria water. On such a basis, a method for extracting cyanobacteria water information, i.e., double infrared band water index (DIBWI), is proposed. On the basis of the Landsat8 imageries of 2014 and 2017 in Taihu Lake area, the comparison and analysis were made with the results of normalized difference water index (NDWI), modified normalized difference water index (MNDWI), new water index (NWI), multi-band water index (MBWI) and water index 2015 (WI₂₀₁₅), and the data of 2013, 2016 and 2018 were used for verification. The results show that DIBWI can extract the cyanobacteria water information, effectively eliminate the influence of cyanobacteria and better inhibit the background features. The overall accuracy is above 98%, and the Kappa coefficient is more than 0.95, which can provide technical support for the protection and reasonable development and utilization of water resources in Taihu Lake area.

The aerosol optical depth (AOD) derived via dark-target algorithm has been widely used as an effective tool for estimating PM_2.5concentrations. However, this algorithm cannot effectively retrieve AOD on the bright surface. Therefore, the authors used a random forest model incorporating meteorological parameters to predict the missing AOD values, and then employed a second-stage random forest model combining the retrieved AOD with meteorological parameters, vegetation cover and road density to estimate the PM_2.5concentrations in two districts of eastern coastal zone of China, i.e., YRD and PRD. The result shows that the proposed model performed very well, achieving R² of 0.94 for AOD predictions and MODIS AOD and an overall R² of 0.97 with RMSE being only 5.57 μg/m ³ between the estimated and observed PM_2.5 concentrations. The spatial distribution of PM_2.5concentrations suggests that the high values are mainly located in Jiangsu Province with low elevation (≥40 μg/m³). The results indicate that the proposed two-stage random forest model incorporated with satellite AOD and other variables could be effectively used for estimating the ground-level PM_2.5 concentrations.

The research on the seasonal spatial and temporal distribution of precipitation is of great significance to the ecological protection and agricultural production in northeast China. Based on the correlation between vegetation index, topographical factors and precipitation, this paper utilizes deep learning models to downscale TRMM_3B43 products to 0.01° (about 1 km) in January, April, July, and October during 2009—2018, and uses site measured data to make accuracy correction and fill areas above 50 ° N which are not covered by TRMM. The results show that the model is better than random forest and can effectively obtain the precipitation distribution in the study area with higher spatial resolution and accuracy in each season. The corrected global determination coefficient R² is between 0.881 and 0.952, the root mean square error (RMSE) is between 1.222 mm and 13.11 mm, and the mean relative error (MRE) is between 7.425% and 28.41%, among which the fitting degree is good in April and October, and relatively poor in January and July.

With the acceleration of urbanization process, the size of the city is growing, and hence it is of great importance to grasp the change of construction land quickly and accurately for the sustainable development of cities. Because SAR images are not affected by the weather, it is possible to use SAR time series to study the expansion of construction land. There are two kinds of time series structures in SAR, which are named “Z” structure and “V” structure in this paper. In view of the previous studies that only consider the “Z” structure but not the “V” structure, this study proposes a construction land extension method based on time series adaptive segmentation. The original time series is segmented in an adaptive manner, the average value of the segments is used as the characteristic value, and the extended area of construction land is extracted by the decision tree. The accuracy and completeness of the method are 89.60% and 92.73% respectively. The results are as follows: ① The method proposed in this paper can effectively monitor the expansion of construction land. Compared with that of the dynamic time warping(DTW) method, the accuracy is increased by 1.80 percentage points and the integrity rate is increased by 1.27 percentage points. ② From 2015 to 2019, construction land in Xinbei District of Changzhou increased by 557.96 hectares, mainly in the south and the southeast.

The study of the change of main urban construction land that is almost blank in the wide area space-time scale can make up for the blank in the wide area space-time scale in the study area. The construction land of six major cities was extracted by using SVM classification method based on the Landsat TM/ETM+/OLI data of long time series from 1990 to 2018 in the study area. The quantitative analysis was made on the landscape metric as well as annual increase and annual growth rate urban development mode. The results show that the SVM method can effectively extract the construction land, with the average of overall accuracy higher than 90% and Kappa more than 0.87. The area expansion of each urban area had reached 1.2~1.4 times and was growing continuously from 1990 to 2018. The annual growth that the largest among the six cities of Pyongyang has reached 1.15 km², while the growth rate of Wosan has a small fluctuation range. And the growth rate that the largest among the six cities of Humhang has reached 2.74% in the recent period. The expansion of six cities in the study area is concentrated in the flat terrain,and the main urban area is distributed along the river or the coast, with the expansion mode of filling type and filling type. In general, its urbanization process is on the rise. This study lays the foundation for the ecological environment protection and the urban expansion and provides reference for the relevant scientific research in the study area.

Chang-Zhu-Tan urban agglomeration is an important part of the economic belt in the middle reaches of the Yangtze River. The DMSP/OLS night light data was used to extract the space of the built-up area of Chang-Zhu-Tan urban agglomeration from 1993 to 2013. The spatial expansion analysis, standard deviation ellipsoid method, center of gravity change, landscape pattern index, and spatial autocorrelation method were used to quantitatively analyze the space-time characteristics of the spatial evolution of the built-up area of Chang-Zhu-Tan urban agglomeration in the past 20 years. The result can provide decision-making basis for the future urban spatial development of Chang-Zhu-Tan urban agglomeration. The results are as follows: ① From 1993 to 2013, the total amount of lights in the Chang-Zhu-Tan urban agglomeration showed a relatively dense core edge structure, the three cities of Changsha, Zhuzhou and Xiangtan were relatively dense, while the surrounding five cities were relatively sparse. The central urban areas of urban agglomerations were dominated by the planar development mode, showing irregular circular expansion outwards. ② The centripetal force of the Chang-Zhu-Tan urban agglomeration showed an increasing trend of “discrete-centripetal”. The center of gravity of the built-up area shifted from Yuelu District of Changsha City to Wangcheng District. ③ There existed the continuous concentration of the built-up areas and the continuous improvement of the degree of integration in the Chang-Zhu-Tan urban agglomeration. The number of patches in the built-up area decreased and the complexity of the shape also decreased. The built-up area grew from “patch quantity increase” to “patch scale expansion”. ④ The spatial distribution of the built-up area of the Chang-Zhu-Tan urban agglomeration presented cluster features. The hot spots of spatial agglomeration were mainly concentrated in the districts of the three cities of Changsha, Zhuzhou, Xiangtan and the surrounding counties of Changsha, Wangcheng, Xiangtan, Shaoshan, Liling, Ningxiang, Xiangyin and Miluo, while the cold spot areas were Yuanjiang and Nanxian.

In order to reduce the error in estimating urban electric power consumption by nighttime light images, it is necessary to consider the development status of sample areas and classify the samples before estimation. In this paper, the NPP-VIIRS nighttime light data from 263 prefecture-level cities in China’s mainland in 2015 were selected to estimate urban electric power consumption. A K-Means city classification method based on light structure rather than traditional statistical data is proposed. The authors used this method to divide the samples into 5 types and estimate the electric power consumption. A comparison of the estimated results with those from other classification methods shows the following regularity: The mean relative error and root mean square error of the estimated results are 32.02% and 57.04, decreasing by 25 and 3.39 percentage points compared with the estimated results without classification respectively. The proportion of high-precision cities in the estimation results is 53.99%, increasing by 13.59 percentage points compared with estimated result without classification, and is the highest proportion among values of all methods. Compared with the estimated results without classification, 152 cities have lower estimated errors. The performance of this method is similar to the optimal performance of other classification methods.

As one of the important mechanisms affecting urban thermal environment, industry accurately detects factories that cause thermal anomalies, and analyzes the impact of industrial thermal anomalies on local thermal environment, which is of great significance for scientific planning of industrial construction and improvement of urban thermal environment. Based on the Landsat8 data of different seasons, this paper uses the radiation transmission method to invert the surface temperature, compares the thermal anomaly detection method based on the thermal field variation index, and performs the local thermal environment effect analysis based on the higher precision detection results. The results are as follows: ① The four-stage method is more suitable for industrial thermal anomaly detection research. ②The scale of the factory production is directly proportional to the area of the corresponding thermal anomaly plaque. For every 5.8 square kilometers of factory production scale, the average thermal plaque area increases by 0.18 square kilometers. ③Industrial thermal anomalies have thermal environmental effects on local building and nonbuilding, the effect of warming on building is smaller with distance, and the effect of temperature increase on nonbuilding in the 1 km range is obvious. The research results can provide reference for industrial thermal anomaly detection and analysis of the effects of industrial thermal anomalies on the local environment.

In order to timely grasp the changes and future development of the ecological environment vulnerability, the authors selected 9 indicators, such as elevation, slope and land use type. Combining RS and GIS technology with AHP-PCA entropy weight model, the authors evaluated the vulnerability of the region in 2000—2018 dynamically, and introduced CA-Markov model for the development of 2021 simulation prediction. The following results have been achieved: ①The overall vulnerability of the region shows gradual decrease in gradient from north to south. ②The degree of vulnerability shows a gradual decrease in the proportion of micro, light, potential, moderate and severe grids. ③CA-Markov is not only suitable for predicting ecological vulnerability in the region, but also with high accuracy, and the proportion of potential, micro, mild, moderate and severe grids in 2021 is 20.18%, 38.02%, 25.71%, 10.96% and 5.13% respectively. ④Throughout the study period, the region’s vulnerability composite index is 2.539 2, 2.501 6, 2.485 6 , 2.460 7 and 2.436 6, respectively, and the 2021 value is 2.428 5. The continuous decrease of this value indicates that the overall ecological environment of the region has been effectively improved and will be in a state of change with sustained and good development. The study effectively reveals the law of dynamic change of ecological environment vulnerability in middle-upper reaches of the Yalong River Basin. With a more scientific analysis of its main drivers and future development, it can be used as an important theoretical reference for the formulation of ecological protection measures in the region.

At present, China has entered into the rapid development stage of urbanization, and urbanization is exerting its positive effect. Such a situation inevitably brings negative effects, but remote sensing technique can quickly, accurately, objectively and quantitatively reveal the present situation of the regional ecological environment quality. Bohu County is one of the most typical areas in arid oasis in Northwest China. Based on Tiangong-2 wide-band images spectrometer and Landsat8 images, the authors constructed a multi-factor comprehensive index RSEI evaluation model in combination with principal components, which was established to evaluate the ecological environment of Bohu County. Additionally, the authors explored the application of Tiangong-2 wide-band images to ecological environment monitoring. The result shows that the greenness (NDVI) and wetness (WET) have positive effects on promoting the ecological environment quality, while the heat (LST) and dryness (NDBI) have restraining effects on ecological environment quality. Greenness (NDVI) has a greater impact on the ecological environment than the other three indicators. In 2018, the RSEI poor grade (0 ~ 0.2) in Bohu County was mainly distributed over urban land and unused land in the north;the fair grade (0.2 ~ 0.4) was mainly distributed among mountains and sandy land. Quantitative and qualitative analysis shows that the driving factors included urban economic development, higher average temperature, greater evaporation, longer sunshine and other natural factors. The ecological environment of wetland and cultivated land was between good (0.6 ~ 0.8) and excellent (0.8 ~ 1.0), indicating that the ecological environment quality of this region was good. According to the spatial differentiation characteristics of ecological and environmental quality, the ecological and environmental quality of the research area in 2018 had a strong positive correlation and certain internal relations, and tended to cluster. This study offers important results and information for planning of regional ecological environment protection and development.

Carrying out territory virescence action and ecological barrier construction is one of the important measures of eco-civilization construction in China. At present, the ecological problems in the northwestern part of Changchun economic circle are relatively prominent, which have become the shortcoming of regional ecological security. Based on the ecological background characteristics and main ecological problems of the study area, the authors constructed a quantitative assessment method for landscape ecological risk including ecological barrier factor. The results are as follows: ① The key areas of ecological barrier construction in the study area include two categories: areas where landscape ecological risk is relatively high, and areas where ecological barrier construction is relatively weak; ② the areas with relatively high landscape ecological risk are mainly distributed in the central part of Nong’an County, accounting for 1.25% and 1.74% of the total area of the study area; ③ the relatively weak areas of ecological barrier construction are the areas where the landscape ecological risk increases after the introduction of ecological barrier factor, mainly distributed in the border zone between Changlin County and Nong’an County, accounting for about 1/4 of the total area. Based on the comprehensive problem-oriented and target-oriented approach, the authors put forward strategies from two aspects of strengthening the control of landscape ecological risk and promoting ecological construction in relatively weak areas, so as to promote the high-quality and sustainable development of Changchun economic circle. This study provides not only a feasible idea for the study of regional ecological barrier construction and related ecological planning but also a useful reference for the optimization of land space spatial pattern.

In order to grasp the law of sea surface temperature (SST) change in the East China Sea from 2003 to 2018, the authors analyzed the relationship between SST changes and climate anomalies, and used remote sensing data to monitor the temporal and spatial evolution of SST in the East China Sea for 16 years. With the 2003—2018 MODIS SST product as the data source, the data were first repaired by the nearest neighbor point value replacement method, and the measured data were used to verify the accuracy. The least square method and Pearson correlation coefficient were used to analyze the SST change trend. Through cross-correlation analysis, the correlation between sea surface temperature anomaly (SSTA) and southern oscillation index (SOI) was studied. The results are as follows: ① SST in the East China Sea generally showed an upward trend from 2003 to 2018, and the temperature rise in summer was more obvious. The temperature rise rate in the Yangtze River estuary could reach above 0.042 ℃/a; ② SST in the East China Sea showed a SE—NW distribution, and at the same latitude, SST near the mainland was usually lower than the eastern sea area, but the SST of Hangzhou Bay area from April to September was higher than that of the eastern area; ③ SOI was basically not related to the East China Sea SSTA that was 15 months behind it, but it had a strong negative correlation with the East China Sea SSTA that was 21~39 months behind with correlation coefficient exceeding -0.2. The research results can provide a reference for grasping the laws of climate change and predicting extreme weather.

Land cover information in farming areas is the basis of land resource management and planning, which plays an important role in the rational development of land resources, adjustment of land use structure, and dynamic monitoring of land. Due to the complex land types and high heterogeneity in farming areas, the accuracy of land cover information extraction has been facing challenges. Therefore, this study used Sentinel-2A/B remote sensing data as the data source. Firstly, a normalized difference vegetation index (NDVI) time series data set and tasseled cap wetness (TCW) time series data set were constructed; Secondly, the J-M (Jeffries-Matusita) distance was used to analyze the separability of the surface features and select the best time series data combination of NDVI and TCW; Finally, combined with random forest (RF), support vector machine (SVM), maximum likelihood classification (MLC) and single phase remote sensing data, the classification of typical features in farming areas was studied, and the accuracy of classification results was evaluated and compared. The research results show that the classification accuracy of the time series data combined with the random forest classification algorithm is relatively high. The overall classification accuracy reaches 88.87%, and the Kappa coefficient reaches 0.855 7, which improves the classification accuracy by 10.05 percentage points and 0.209 3 respectively compared with that of the single remote sensing data. This fully demonstrates that the combination of time series data and random forest classification algorithm can effectively improve the classification accuracy of typical features in farming areas.

Based on “Easy Interpretation” image processing software and using BJ-2 satellite remote sensing images, the method of “object-oriented+deep learning” was introduced into the intelligent information extraction and automatic classification of the 200 km ² test plot in Shihe District, Xinyang City, which included forestland, tea garden, paddy land, water area, construction land and some other land. By the method of ratio vegetation index (RVI) in combination with the real-time selection of the boundary index threshold and eigenvalues, the forestland, tea garden and paddy land information of the test plot was classified intelligently. The water area information was extracted intelligently by the green and near-infrared band normalized difference vegetation index (NDVI). The information of construction land was extracted by using standard deviation of band1 as the eigenvalues. Based on the above methods and field geological survey, the results show that the intelligent information extraction in the test plot has a high accuracy of over 90%. The efficiency is 19 times higher than the traditional method. The study shows that “Easy Interpretation” image processing software is effective and highly accurate and can do half the work with twice the results, which has good value for extension and application in the intelligentized interpretation of natural resources and environment.

The traditional remote sensing survey method of debris flow is mainly orthophoto aerial photography, which has limitations in the accuracy and dimension of data acquisition. UAV tilt photography technology can simultaneously acquire images from different angles, such as vertical and tilt, obtain more complete and accurate information of ground objects, establish more intuitive three-dimensional model, and provide new technical means for geological disaster investigation. Taking the Caojiafang debris flow gully in Shijiaying as an example, the authors carried out the feature recognition analysis of debris flow disaster based on UAV tilt photogrammetry. It is believed that the high-precision three-dimensional model and texture details of debris flow gully can be obtained by incline photography, which can truly reflect the high-resolution information of the top and side of the real surface, and accurately obtain the distribution of debris flow material source and the estimation of material source volume; the data can be used to calculate the maximum amount of debris flow once washed out. This method can provide a more powerful means for the investigation of debris flow geological disasters and the assessment of current situation. The remote sensing technology can be fully used in the investigation and evaluation of debris flow disaster.