Oil spill is one of the main sources of pollution to the marine environment. Early monitoring of oil spill is very important for marine environment protecting. In this paper, the calculation of radar backscattering based on the wave spectrum was carried out, and a review of the study of the damping ratio of wave spectrum in consideration of the films characteristics, water molecular tension, elastic model and surface tension was carried out. The problem of insufficient research on the damping of the oil spill remote sensing monitoring with the wave spectrum and the quantitative calculation of the damping was discussed. The research on the damping of the oil spill for remote sensing monitoring in the future may be based on the backscattering characteristics of the real ocean wave spectrum under the cover with oil slicks. The research on radar coefficient calculation can provide support for quantitative analysis of the damping characteristics of oil spills, thus improving the accuracy of oil spill remote sensing monitoring.

Technical equipment development is an important part of natural resources monitoring. According to the different delivery platforms, the monitoring technology and equipment of natural resources are classified into three categories: satellite-borne, airborne and ground-based, and their details are described in this paper. The authors briefly analyzed the problems and challenges faced by the current development and, at the same time, put forward some suggestions for development which is based on the actual situation of China: To carry out inventory and assessment of natural resources and equipment; to emphasize the importance of infrastructure construction and resource integration of technological equipment, to make technical equipment development plan, and to adopt some other measures.

A method for remote sensing image semantic segmentation based on deep fusion convolutional networks and conditional random field is proposed. First, the fully convolutional networks framework with deconvolutional fusion structure is utilized to integrate the pooling-layer results at different levels. The low-level features with rich detail information are combined with the high-level features via deconvolutional fusion module. At the same time, the parameter inference process of conditional random field is embedded in the network architecture by adding recurrent neural networks iteration layers. In addition, the deep fusion convolutional networks and conditional random field model is established. Then, in the model training stage, the abundant detail information and context information in the image are introduced simultaneously to the positive and negative propagation. And lastly, the remote sensing image semantic segmentation is accomplished by the end-to-end framework. Semantic segmentation experiments were performed on the high-resolution optical remote sensing images, and the results show that, with the increase of the depth of deconvolution fusion layer in the model, semantic segmentation results are more refined, and the contour of terrestrial object is more accurate. The introduction of context information also improves the accuracy of image semantic segmentation to a certain extent. It is concluded that the proposed method can better maintain the consistency of internal areas of semantic object and effectively improve the accuracy of semantic segmentation.

The abandoned land has been spread all over the world and has become an important research direction for land use. Due to the lack of optical remote sensing images and serious mixed pixels in Karst rocky desertification land, it is difficult to accurately identify and extract the abandoned land. Based on short-term high precision image and long temporal resolution data features, taking Xifeng County in Guizhou Province as an example, using the high precision image accurately, and aided by maximum value composite (MVC) method, the authors calculated Landsat data for 2003—2018 time-series NDVI data, identified characteristics of abandoned land NDVI, and analyzed the relationship between the abandoned land and karst rocky desertification. The results are as follows: ①The combination of land parcels and time series NDVI can accurately identify and extract abandoned land, with an accuracy of 90.7% under the condition of 95.56% of cultivated land extraction. This method has a good application effect in cloudy and rainy mountains areas where optical remote sensing data are lacking and cultivated land is broken. ②The curve shape of the NDVI of the abandoned land is of “V” type, the continuous abandonment of arable land curve shape is of “U” type, and the cumulative NDVI curve shape is of asymmetric “W”, “M” or several combinations. ③The overall level of NDVI in non-karst abandoned lands is higher than that in rocky desertification abandoned lands. The level of rocky desertification is inversely proportional to the overall level of the curve and positively correlated with the degree of curve fluctuation. The number of abandoned lands is inversely proportional to the value of NDVI of the plot and positively correlated with the degree of dispersion of the NDVI curve. The results provide an efficient and feasible method for accurate identification and extraction of uninhabited land in karst cloudy and rainy mountain areas.

Forest coverage is a common variable in forest management, monitoring and planning. It is also an important reference for biodiversity impact assessment and carbon storage quantification. Due to its objective, fast and macroscopic advantages, remote sensing technology has gradually been widely used in forest resources monitoring. In order to get better application of forest cover products based on remote sensing technology, validation and quality evaluation are particularly important. Two widely used forest cover products (UMD and JAXA FNF) were selected in this paper. UMD (University of Maryland) forest cover product was derived by EDENext Data Management Team from original datasets produced by UMD et al. 2013/UMD/Google/USGS/NASA. The forest/non-forest (FNF) product from ALSO/PALSAR data in 2015 was available on the Earth Observation Research Center, Japan Aerospace Exploration Agency (JAXA). Validation and quality evaluation were carried out by comparison between products and ground-based survey data. UMD and JAXA FNF forest cover products were compared through qualitative analysis and statistical analysis. The results show that the forest coverage rate of JAXA FNF in western, southern and central eastern China is significantly higher than that of UMD. In southeastern provinces of China, the consistency of UMD and JAXA FNF is better, but both of them are higher than ground-based data. In general, the accuracy of UMD is higher than that of JAXA FNF. Although the forest coverage data of UMD and JAXA FNF and GFW and CFGA are different in all provinces, the difference of JAXA FNF is larger, and the average absolute error of JAXA FNF is about 3 times of UMD.

The saturation and overflow effects of DMSP/OLS data tend to misclassified the boundary pixels between urban built-up and non-urban built-up areas as urban built-up areas, which will affect the extraction accuracy of urban built-up areas. This paper proposed an urban built-up area extraction method based on the combination value of light brightness. By determining the accurate boundary location between built-up and non-built-up area, the authors tried to improve the extraction accuracy of built-up area. First, the combination of the brightness of the light is constructed to highlight the difference between the urban built-up and the non-urban built-up area. Second, the Mann-Kendall non-parametric test method is used to determine the change-point of the combination value of light brightness of adjacent pixels, which is used as edge pixel to identify urban built-up area. The relevant experiments were carried out in Guangzhou, Shenzhen, Wuhan and Nanjing. The results show that, the average area error is only 10.89%, the average coverage of spatial overlap is 84.93%, the kappa coefficient is 0.863 0, and the overall accuracy is 91.68%. Compared with threshold method, NSA and feature combination method, this method can reach the highest extraction accuracy, and hence the authors provide an effective and reliable method for extracting urban built-up area based on nighttime light data.

In order to solve the problem that there are large geometric deformation and gray difference between different remote sensing satellite image and it is difficult to match a large number of feature points, the authors put forward a multi-source remote sensing image matching method under multiple constraints in this paper. First, ASIFT algorithm is used to extract high-quality feature points and complete the initial matching, and the matching results are optimized by RANSAC algorithm. Secondly, affine transformation matrix of the two images is calculated by using the matching feature points set, and the remaining feature points are matched again by combining affine transformation and gray correlation coefficient. Finally, support vector regression (SVR) is used to check the matching results. Satellite images of ZY3-01, ZY3-02 and GF-1 were selected in the experiment. The experimental results show that, compared with SIFT and ASIFT algorithms, the proposed method can greatly increase the number of matching points between multi-source remote sensing images and improve the matching accuracy.

In the GF-1 image data applications, applying cloud layers influences accuracy of information extraction and image utilization rate, to tackling this problem, this paper proposes a support vector machine cloud detection method combining image spectral features and texture features. For GF-1 data, the method of the gray-level co-occurrence matrix is used to extract those texture features. Spectral characteristics of clouds and ground and texture characteristics serve as feature vector, and support vector machine (SVM) is used to conduct cloud detection to GF-1 data. Studies have shown that the precision and recall of this method for all kinds of cloud detection are above 99.2% and 93.9%, and the error rate is below 1.1%, which is obviously better than the cloud detection algorithm using traditional support vector machine and maximum likelihood value, and it combines the image texture and spectral characteristics, and thus it has certain universality in theory.

To tackle the problem of low inversion accuracy of regional forest crowns based on optical remote sensing data, in this study the authors used multiple stepwise regression (MSR), partial least squares regression (PLSR) and back-propagation (BP) network models to perform regional forest crown height inversion based on the texture, spectral and topographic characteristics of SPOT5 multispectral images. The inversion accuracy of the models was compared and analyzed to determine the optimal model for the study area. The results show that the correlation between the texture parameters of each forest type and the measured canopy height of the plot is better than other spectral parameters. The BP neural network model performs better than other models, and the determination coefficients R ² of the validation results for the broad-leaved, coniferous, and mixed forest were 0.76, 0.97 and 0.92, respectively, and the root mean square error (RMSE) were 1.6 m, 1.35 m and 2.29 m, respectively. Studies have shown that texture parameters can reflect the structural characteristics of forest canopy well, and the BP neural network model combining image texture, spectrum and terrain feature parameters has good application potential in forest canopy height inversion.

At present, the method of extracting the impervious surface area based on the impervious surface area according to the impervious surface spectral index has been widely used because of its concision and speed. However, the method of extracting impervious surface by spectral index has the disadvantage that bare land and impervious surface are easily confused. To tackle this problem, the authors created impervious surface and bareness area index (ISBAI) according to the spectral feature difference of impervious surface, bare land and water body as well as vegetation in the 4, 5 and 6 bands of Landsat8 OLI images. Based on ISBAI and bareness area index (BAI), the authors built a new type of impervious surface index, called the bareness - restrained impervious surface index (BRISI). Improved double-window flexible pace search (IDFPS) method was used to determine the optimal threshold, and impervious surface extraction was performed. Chongqing (a mountain city) and Xi’an (a plain city) were selected as the research area to evaluate the accuracy of BRISI extraction in comparison with other commonly used impervious surface indices. The experimental results show that the extraction accuracy of BRISI in the experimental area of Chongqing and Xi’an experimental area reach 86.8% and 86.8% respectively, in comparison with the accuracy of all other indices that took part in the contrast, BRISI extraction accuracy is the highest. Meanwhile, BRISI also eliminates the influence of bare land in the construction area extraction, and overcomes the problem that it is difficult for other impervious surface indices to distinguish bare land from impervious surface.

With the deepening and cross-fusion of modern remote sensing image research, the classification of high spatial resolution remote sensing image (referred to as “high-resolution image”) has become a research hotspot in the field of remote sensing. As the phenomenon of “homology spectrum” and “homology spectrum” of high-resolution image is more serious, the deep learning method that has emerged in recent years has proposed a new solution for high-resolution image classification. However, the lack of training samples of remote sensing images can easily lead to over-fitting of deep learning networks. In this paper, an improved Inception-V3 remote sensing image scene classification model is proposed by using deep learning method and transfer learning strategy. The model first adds Dropout layer before the full connection layer of the original Inception-V3 model in order to avoid over-fitting. In the training process, the transfer learning strategy is adopted to make full use of the existing model and knowledge and improve the training efficiency. The experimental results based on AID and NWPU-RESISC45 datasets show that the improved Inception-V3 has faster convergence speed and smoother training effect than the original Inception-V3 training. Compared with accuracy of other traditional methods and deep learning networks, the classification accuracy of the proposed model has been greatly improved and verified. The effectiveness of the model is verified.

The improvement of urbanization level and the imperfection of waste treatment infrastructure in China have caused the problem that informal garbage stacking becomes more and more prominent. With the development of high resolution remote sensing images, macro and efficient management of informal garbage dumps becomes possible. Most of the existing researches use visual interpretation and traditional supervised classification methods to identify informal garbage dumps from high resolution remote sensing images. Such methods are very time-consuming, and it is difficult for them to mine deep features from data, therefore the detection accuracy is limited. Therefore, the authors used convolutional neural network, proposed an informal garbage dumps detection framework SU-RetinaNet from high resolution remote sensing image based on sample updated and RetinaNet, analyzed the impact of different parameters and network structure on the model detection performance, and compared the results of using DPM, R-CNN , Faster R-CNN, RetinaNet and SU-RetinaNet 5 algorithms for the performance of informal garbage dumps detection. Experimental results show that the use of SU-RetinaNet for informal garbage dumps detection can achieve average precision of 85.92% and a detection speed of 0.097 s per image. Compared with traditional method, SU-RetinaNet greatly improves the detection efficiency of informal garbage dumps, and provides a feasible technical solution for effective monitoring and management of municipal waste.

This study observed relatively developed urbanized and industrialized urban agglomeration in central and southern Liaoning. The authors used remote sensing data as the information source to study the expansion of construction land and the characteristics of its occupation of cultivated land. The results show that, from 1990 to 2015, the total land use for construction of urban agglomeration in central and southern Liaoning increased by 1 942.07 km ², with an increasing rate of 30.52%, in which 64.68% of new construction land was derived from the occupation of cultivated land. The difference in the expansion of construction land among cities was obvious. The comparison of data in different periods shows that the expansion rate of construction land in central and southern Liaoning had a turning point. During the study period, the area of cultivated land in central and southern Liaoning continued to decrease, and 79.76% of its loss was converted into construction land. Due to the advancement of agricultural science and technology, the current reduction in cultivated land area has a small impact on grain production in central and southern Liaoning. However, some cities have caused food production reduction due to occupation of cultivated land, which requires further attention.

“Cold island effect” has obvious ecological significance to oasis water saving and agricultural development in arid areas. In this paper, the “cold island effect” of oasis was studied based on oasis surface temperature in Jinghe basin, which is both a national key development area and a typical ecologically fragile oasis in Xinjiang. The results are as follows: ① In the past 26 years, there was a great decrease of sub-heat island in Jinghe basin, and the oasis cold island effect showed an increasing trend; ② Relatively,the cold island effect of oasis in the research area was not obvious in 1990,but, with the increasing human activities, especially the increasing proportion of cultivated land, grassland and water since 2000, the cold island effect of oasis was more significant than the heat island effect in the study area. ③ The N-S and W-E trending profile analysis shows that the influence range of oasis cold island effect is dominated in the north-south direction; ④ The enhancement of oasis effect in the study area has a relatively consistent response with the increase of annual precipitation and temperature change,and the surface vegetation coverage is the main influencing factor of oasis cold island effect.

The active faults deformation, with distinct temporal and spatial variations, requires long-term and effective monitoring. InSAR (interferometry synthetic aperture Radar), with the advantages of high efficiency, wide coverage and low cost, is one of the main technical means for ground surface deformation survey in recent years. In this paper, the small differential deformation of the main active faults in Tangshan urban area was monitored by the interferometric point target analysis. Two kinds of radar data were used: the Wide strip mode of RADARSAT-2 C-band with 30m spatial resolution and the Strip mode of TerraSAR-X X-band with 3 m spatial resolution. The results show that the differential deformation of Tangshan-Guye active fault is obvious with the maximum differential velocity of 2 mm/a. In this case, the deformation results from RADARSAT-2 C-band medium resolution data can clearly show the tiny differential deformation between the two sides of the active faults. However, TerraSAR-X X-band data, with shorter wavelength, is more obviously affected by the change of surface cover. Compared with this deformation, the vertical differential deformation between the two sides of active faults is too small to be separated from the TerraSAR results.

The process of landscape fragmentation is accompanied by the decline of landscape function. Therefore, the study of landscape fragmentation is of great significance for timely monitoring of ecological security and adjustment of land cover structure. This research was designed to map the current status of land cover structure of Mata watershed using supervised classification algorithms in south of Yan’an City based on Landsat 8 satellite data and to describe its landscape fragmentation using six fragmentation indices at three organization levels: patch, class, and landscape. The results are as follows: ① The status of land cover in this area can be characterized accurately based on Landsat8 satellite image and the accuracy of supervised classification is 74% together with the kappa value of 0.68; ② Mata watershed can be classified into 6 land cover classes, i.e., forest land, shrub land, grass land, orchards land, farm land, and others land (road and village). The orchard land occupies the largest area in all land cover types; ③ The extent of landscape fragmentation for shrub, grass and farm lands is relatively more serious than that of forest, orchard and others lands, indicating that landscape functions of shrub, grass and farm lands have been weakened in capability of ecological protection and agricultural production; ④ Transforming small patches of shrub, grass and farm lands into adjacent land type with large patches should improve the integration level of landscape in the Mata watershed, which is conducive to the improvement of landscape function in Mata watershed.

In view of the low precision of existing water quality element retrieval models applied to the coastal waters of Golden Beach, the authors, based on the statistical retrieval models of water color for case Ⅱ water body in Yellow Sea and East China Sea by Tang Junwu, established the retrieval models of chlorophyll-a and total suspended matter concentration for coastal waters of Golden Beach by using the spectral data obtained from airborne marine hyper-spectrometer. The spatial distribution of chlorophyll-a and total suspended matter concentration in the study area was obtained and the influence of hyper-spectrometer gain on model retrieval accuracy was analyzed. After the models were improved, the determination coefficients and average relative errors between the retrieval results from spectrometer measurements and the sampling measurements were respectively chlorophyll-a 0.65, 4.41%, and total suspended matter 0.80, 3.55%. Retrieval results from the same spectrometer at the same coordinates and approximate time but under different gains were compared. It is proved that retrieval average relative errors and root mean square errors of improved models are all increased and the retrieval accuracy is reduced if gain changes. However, the error is in the allowable range and the model stability is good overall.

In order to solve the problem that the traditional geological survey has a large workload and low efficiency and that it is impossible to extract information from high slope geological points, the authors chose the Liujiang Basin in Qinhuangdao City, Hebei Province, as the research object. Oblique photogrammetry technology was used to obtain the image data of the study area. The generated image was used to construct a high-resolution 3D real-world model using the Context Capture software. The 3D model was identified and extracted in the Acute 3D viewer, including the extracted area’s latitude, longitude, and elevation. Several pieces of basic information were extracted, such as the distance between two points, the vertical distance, identification of the fault and calculation of the fault’s strike, tendency, and inclination, identification of the type and nature of the rock, and calculation of data such as perimeters, surface area, and volume value of the exposed rock. The results show that the extraction results are in good agreement with field measurements and can achieve the purpose of identifying geological phenomena and extracting geological quantitative information. It is shown that the method adopted in this paper has practical significance for geological information identification and extraction of UAV remote sensing technology.

The monitoring results in recent years show that land subsidence in Zhengzhou has been developed rapidly. In order to find out the evolution law of Zhengzhou’s land subsidence and serve the prevention and control work of urban land subsidence, the authors selected the synthetic aperture Radar(SAR) data in recent years (2007—2017), compiled the land subsidence distribution maps of Zhengzhou in four periods of 2007—2010, 2012—2013, 2013—2016 and 2016—2017 by combining with the bench-mark monitoring results, and analyzed the evolution law of Zhengzhou’s land subsidence from time and space. By the GIS room analysis method, the authors studied the space-time response relation between land subsidence and urban village evolution in recent years. Research results show that the urban village is the predominant factor in the land subsidence evolution of Zhengzhou, namely, the groundwater abstraction in urban village causes land subsidence; the relocation of urban village and reduction of groundwater abstraction cause slow land subsidence and even uplift of land subsidence; the relocation causes the floating population to move to outer suburbs, which forms a new floating population and industrial accumulation area, and further forms a new ground subsidence area.

Terminal lake, which is an important part of oasis ecosystem in the lower reaches of inland river in arid region, has the functions of water conservation, wind prevention and sand fixation. However, for nearly half a century, the terminal lake of inland river has been shrinking under the influence of climate change and human activities, and the ecological environment around the lake has been deteriorating, which has seriously threatened the regional social development. In order to explore the changes and driving factors of the water area of the terminal lake in the arid region, this paper took the Taitema Lake region as an example, selected nine Landsat TM/OLI remote sensing images from 1986 to 2019, and extracted and analyzed the change characteristics and driving factors of the water area of the Taitema Lake region. The results showed that the water area of the study region increased by 163.93 km ² in the past 34 years. In terms of spatial distribution, water area of the study region was scattered around the lower reaches of Qarqan River before the 21st century, and mainly distributed around Taitema Lake since the 21st century. The center of gravity of the water area in the study region experienced four processes: rapid westward shift in the late 1980s, slow eastward shift in the 1990s, rapid eastward shift in the late 1990s to early 2000s, and eastward shift in the fluctuation since the 21st century. Taking the water transport to the lower reaches of Tarim River as the boundary, the water area before water transport is mainly affected by the comprehensive effects of regional natural factors, with precipitation as the main driving factor (r ²=0.825). After water transport, the water transport process with human intervention is the leading factor (r ²=0.977) affecting the change of water area in the study region. With the continuous water transfer to the downstream of Tarim River, the ecological environment of Taitema Lake has been restored continuously. In order to prevent the lake from overflowing and spilling eastward into Lop Nor, it is necessary to explore the maximum water capacity of Taitema Lake so as to reasonably control the water transfer.

The rapid development of China’s industrialization and urbanization has brought about a series of ecological and environmental problems. China has proposed a new ecological redline policy (ERP), which plays an important role in protecting natural ecosystems and guaranteeing the national ecological safety. For accurate monitoring of ecological redline areas (ERAs), the high temporal-and-spatial resolution BJ-2 satellite imagery was used for land cover classification of the ERAs of Nanjing. Given the characteristics of BJ-2 satellite imagery, a workflow from data preprocessing to object-based land cover classification was established. The overall accuracy of the classification can reach to 91.65%. It is shown that the ERAs of Nanjing is mainly composed of three kinds of land cover types: forest, cultivated land and water, which occupy 33%, 21% and 25% of the study area respectively. In addition, buildings and artificial pile digging account for 6% and 2%, which can represent human influence to a certain extent. The experimental results show that the multi-temporal BJ-2 imagery can be used to detect the detailed changes of land cover that are difficult to identify in low- and medium-resolution images, and achieve the purpose of dynamic and accurate monitoring of ERAs.

Based on the remote sensing data, this paper aims to construct an urban environment livability evaluation system on the basis of remote sensing factors and further evaluate the livability of the urban environment. An object-oriented classification method was used to classify urban ecological land into five categories from GF-1 satellite data: water body, vegetation area, road, building land, and bare soil. In addition, the urban surface temperature was retrieved from Landsat8 thermal infrared data and then, the urban thermal environment was calculated. Finally, on the basis of constructing the urban ecological land and urban thermal environment factors, a weight method was used to calculate the ecological quality index (EQI) of the city in order to establish an overall evaluation system of urban environmental livability. Applications in Beijing City showed that, from 2013 to 2014, the urban heat island effect in Beijing was not obvious, and most areas were suitable or basically suitable for human habitation.

The ecological environment changes are related to social development. This paper analyzes the dynamic changes of the ecological environment quality of the Liuchi River Basin in Guizhou, in the hope of providing some reference that can promote the ecological civilization construction and environmental restoration of the watershed in the karst trough. The authors used remote sensing images which were from June to August, 2000, 2010 and 2017 Landsat satellites and GDEMDEM (30 m) data to extract vegetation cover, soil index, slope index, humidity and temperature index which served as the basis for constructing an ecological environment assessment model. In addition, spatial auto-correlation method was adopted to reveal the ecological environment that had different spatial and temporal characteristics in township scale. Some conclusions have been reached: ① From 2000 to 2017, the ecological environment quality showed a rapid restoration trend in the Liuchi River basin, but it varied differently in different counties. ② The quality of ecological environment showed significant differences in different towns. Meanwhile, the ecological improvement was slower in the county town and downstream towns, and was improved obviously in some towns of the central and southern area. ③ The ecological environment quality of towns had significant positive spatial agglomeration characteristics, and were dominated by high-high and low-low concentration in most years. In the trend analysis of towns’ ecological environment quality, the southeast basin was always the low-value zone, and the high-value zone is mainly distributed in the northern townships.

According to the fact that data missing problem often occurs during the study of long-term remote sensing dynamic monitoring of inland lake area, the authors tried to make up for the missing remote sensing images by utilizing the ESTARFM to combine with the MODIS data so as to simulate the missing Landsat images after 2000. On such a basis, the water index method was used to extract the lake area and shoreline so as to realize the long-term remote sensing dynamic monitoring of inland lake. Hongjiannao Lake in Inner Mongolia was selected as the study area. The method was tested by making up for the missing Landsat images from 1987 to 2018 with ESTARFM algorithm and extracting Hongjiannao Lake from all the images. Some conclusions have been reached: The MODIS and Landsat fusion images generated by the ESTARFM algorithm are ideal, which can effectively solve the problem of missing Landsat images after 2000. It is proved that the image obtained from the ESTARFM fusion can be applied to water extraction. In addition, time-series images with the fusion images are added to reflect the water changes more delicately when the water dynamic change is monitored, which contributes to the subsequent research. In addition, through the long-term remote sensing dynamic change monitoring of Hongjiannao, it is found that the lake changes generally show a shrinking stage, and the specifics can be divided into three stages: stability, sustained shrinkage and growth.

The deterioration of habitats caused by urbanization is an important reason for biodiversity loss. Based on Landsat TM/OLI images, DEM, digital maps and local climatic data from 1995 to 2015, the authors constructed the image interpretation system of Dongting Lake area, which was based on CART acquisition rules. According to the LUCC data obtained, the authors analyzed the spatio-temporal difference and correction of urbanization and avian habitat in Dongting Lake area, with the support of the software such as ENVI, ArcGIS, and MaxEnt. The results show that the classification method based on the CART algorithm for acquisition rule gained a higher precision than the ordinary classification method. The construction land in lake area increased by 808.61 km ² in 20 years, the high-intensity expansion area was spread from the urban center to the county-level city, and the expansion rate gradually decreased after 2005. In the past 20 years, the suitable habitat area of waterfowls has decreased by 195.19 km ², and the overall suitability of this area has been declined. However, the situation has gradually improved since 2000 with the government’s efforts to rectify the environment. The results of polynomial fitting and spatial comparison indicate that rapid and disorderly expansion of township construction land has reduced the suitability of avian habitats in lake area, and that, to protect avian habitats, people should take effective measures to protect the ecological land around there and rationally plan and control the expansion of construction land in towns and village areas.

Residents are the basic living places of human beings, and their livability is significant for human life experience. Based on the two aspects of natural environment and social environment and considering the factors of natural environment, environmental pollution, cultural environment, living convenience and social security, the authors constructed the method of regional livability evaluation. With Changzhi City as the study area, the authors established regional livability evaluation index system using analytic hierarchy process method and calculated the livability index based on Landsat OLI image, DEM, online map POI point and other data in 2018. In addition, the authors classified it by the mean standard deviation method. The livable situation in Changzhi City was analyzed, and the result can provide data support for urban planning management and sustainable development. The results show that the areas with high livability in the study area are mainly concentrated in the south-central as well as western and northwestern of central, whereas the areas with lower livability are concentrated in the northwest, southwest and northeast. There are more high-living areas and higher livable areas in Luzhou District and Shangdang District, and more moderately livable areas in other districts and counties.

Two crops with similar growth periods are likely to be confused during remote sensing recognition,which brings difficulties to crop identification and area extraction. For the purpose of solving this problem,Jianghan Plain was used as the study area and,according to the MODIS EVI timing curve of different features in combination with the spectral characteristics of the objects themselves and the phenological information of crops,the combination of decision tree and quadratic difference method was used to extract the planting area of rape and winter wheat in Jianghan Plain in 2010. The results showed that, compared with the statistical data,the overall accuracy of remote sensing extraction of rape and winter wheat planting area in the study area was 93.7% and 87.1%,respectively, and that, at the city (county, district) level,the correlation coefficient R ² between calculated data and statistical data is 0.88,and the winter wheat is 0.90. This study can accurately identify rape and winter wheat with similar growth periods so as to obtain better results of planting area extraction,which has certain universality and can provide technical support for monitoring and estimating yield of rape and winter wheat in Jianghan Plain.

Based on the two-year high-resolution satellite image data acquired in 2016 and 2017, the authors carried out remote sensing investigation and dynamic monitoring of mine geological environment restoration and management in Zhejiang Province. By using the method of interpretation and field investigation by domestic satellite remote sensing, the status quo of mine restoration and management in Zhejiang Province was clarified. According to the results of remote sensing monitoring and impact characteristics, six models were summarized and classified: ecological restoration model, industrial park model, agricultural land-making model, landscape transformation model, disaster management model and comprehensive management model. These six models can be used for reference and promotion of mine geological environment restoration and management in other areas.

In order to further study and explore the application effect and potential of hyperspectral remote sensing technology in geological prospecting, the authors used the aerial and ground hyperspectral remote sensing data of Hongshan region in Gansu Province to analyze the distribution characteristics of altered minerals or altered anomaly information, summarized the distribution law of altered anomaly of ore-forming geological bodies, and proposed the construction idea of ore-finding positioning model of ore-forming geological environment elements based on alteration information characteristics of known typical deposits, ore-controlling elements and alteration information law of ore-forming geological bodies. The authors gradually revealed the metallogenic geological environment expressed by “minerals-landmark minerals-altered minerals-prospecting anomalies”, and constructed the prospecting positioning model of typical deposits and the comprehensive prospecting positioning model of altered minerals and geological background. Guided by the ore-finding positioning model and combined with geological background and ore-forming laws, the authors applied different models to delineate 4 polymetallic prospective areas. After field investigation, better clues of polymetallic mineralization were found. Practice shows that, by analyzing the intrinsic relationship between hyperspectral alteration information and ore-forming laws, more accurate information can be provided for ore prospecting, and the same kind of polymetallic ore prospecting work can be effectively guided.

To tackle the problem whether mining will cause great changes in the types of surface features and environmental deterioration, the authors used Landsat (TM,OLI) images to classify the land in 2006, 2010, 2014 and 2018 in Hailiutu watershed, and revealed the temporal and spatial characteristics of land use changes in three stages (one stage every four years) from 2006 to 2018. Screening and comparing the classification methods MLE, SVM, RF and applying the statistical methods of features change and transfer matrix show that the accuracy of land classification map obtained by RF is better than that of the other classification methods, and the quantitative interpretation of land classification analysis was carried out for many years. In the three stages, the transformation of sandy land and grassland/shrub was frequent, the total area of sandy land decreased by 16.83%, the grassland/shrub increased by 12.68%, and the construction land increased steadily year by year. By 2018, the development of the mining area had not caused great damage to the ecological environment, and the change of the geological structure of the mine was consistent with the trend of the geological structure of Hailiutu basin.

In order to effectively carry out remote sensing monitoring work of mine geological environment restoration and management, the authors established a model for dynamic change detection of vector data, and proposed an automatic method for change detection of vector result data, which is suitable for the change of mine geological features. According to the spatial topological relationship and the attribute change situation of different temporal vector result data, the dynamic change of elements is automatically analyzed and detected, and the plug-in application framework based on ArcGIS Engine (AE) is used to develop the dynamic change detection software of result data. The results show that this method improves effectively the efficiency of dynamic change detection of vector result data, reduces the workload of restoring, harnessing and mining damaged data when they are put into storage, and provides strong support for relevant departments to further monitor the recovery and management of mine geological environment.

For the purpose of monitoring mine development status by using remote sensing technology, ArcGIS software is usually used to extract relevant information to produce illegal mine distribution maps, thus providing data support for relevant departments. At present, artificially produced illegal mine distribution maps have many problems, such as high production cost, high professionalism, large workload, inconsistent drawing format and low accuracy. Therefore, a method for automatically generating illegal mine distribution maps based on ArcPy is developed. The method automatically extracts the attribute information and spatial information of the remote sensing monitoring data of illegal mine mining areas, and automatically locates the image range and edits the text information according to the drawing template. The automatic generation of mine distribution maps greatly reduces the workload of technicians. It is of practical significance in improving the efficiency of mine remote sensing monitoring.