The soil moisture can be considered as an appropriate indicator to investigate the level of ecological environment disturbance resulting from mining activities in semi-arid grassland. The main objective of this research is to explore the applicability of Chinese HJ-1B data for LST and soil moisture monitoring around mining-affected areas on the local scale. The JM&S, Qin and Artis methods for temperature retrieval were comparatively analyzed. The relationship space of NDVI-LST was used to generate temperature vegetation dryness index(TVDI). Furthermore, the reference data including in situ soil moisture and MODIS LST products were used for “dry edge” correcting of TVDI. Some conclusions have been reached: The Qin’s mono-window algorithm performs best in LST retrieval from HJ-1B data; there is a highest correlation between corrected TVDI value with C=0.3 and in situ soil moisture value; the feature of NDVI-LST space indicates that there is a linear relationship for “wet edge”, while the relationship for “dry edge” is conic; the TVDI imagery and LST imagery show different drought conditions of different features. The obvious geographical heterogeneity has been found from the TVDI and LST imagery in this area as well.

In order to overcome the disadvantages of the classification method based on spectral and texture features, the authors put forward a classification method based on the fusion of spatial relationship in this paper. Single object probability was built by G statistic after image object feature was extracted by histogram. The neighborhood object probability was described by land cover adjacency probability which was calculated by iterative statistics method. The joint probability of the object was built by the weighted combination of single object probability and neighborhood object probability. The classification result of the image was obtained according to the maximum a posteriori. The experimental results based on QuickBird image show that the proposed method can improve the classification accuracy compared with the traditional classifier using spectral and texture features. The overall classification accuracy and kappa coefficient are increased by 1.5% and 2.1%, respectively.

Residential area, as an important hazard-bearing body of earthquake disasters, usually constitutes the basis of earthquake emergency preparation. In this paper, 2 m resolution GF-1 satellite remote sensing data were used to extract the spatial distribution information of residential area, which could provide data support for the earthquake emergency preparation. The results reveal that more accurate residential area information of the high resolution GF-1 2 m image can be extracted based on gray level co-occurrence matrix, binarization and mathematical morphology. This proposed algorithm has high accuracy and good robustness. However, higher false alarm rate factor was shown in the extraction of sparse vegetation and non-residential buildings. Thus, in order to ensure data accuracy, the authors compared and analyzed the extraction results and the images, and extracted the ultimate data results semi-automatically by the artificial intervention.

A remote sensing image enhancement algorithm, which is based on the non-subsampled contourlet transform (NSCT)and the fuzzy theory, was proposed in this paper. Firstly, the low pass and high pass coefficients in different sizes of the image were acquired using the NSCT transform. Then, a membership function in fuzzy theory was defined to enhance the high pass coefficients. In the process of transforming the fuzzy domain to NSCT domain and reconstructing the image, the high pass sub-bands coefficients were added into low pass sub-bands step by step and the enhancement was realized finally. The results of the experiments show that the proposed method could enhance the remote sensing image perfectly in both subjective and objective aspects. The results obtained by the authors suggest that the high-pass coefficients of the NSCT transform of the image contain most of the details of the original image, and image enhancement task could be attained by fuzzy transformation of the high-pass coefficients. However, the proposed method has the disadvantages of large computation quantity and the requirement of manual adjustment of several parameters.

A novel segmentation method combining color histogram with LBP textural features for segmentation of high-resolution remote sensing images is presented. This method starts with an adaptive marker-based watershed algorithm to obtain an initial segmentation result, and the markers are constructed by dual-threshold joint segmentation of the gradient image. And then a regional similarity indicator combining color histogram with LBP textural features is adopted to guide regional merging procedure and obtain the final result. Comparative experiments on high-resolution remote sensing images have proved the effectiveness of the method.

In order to improve the effect of information extraction from high spatial resolution remote sensing images (HRI) of cropland, the authors put forward a new HRI segmentation algorithm. Due to the fact that the traditional Mean-Shift (MS) segmentation method only uses a global and single scale, and that some variable bandwidth MS only considers spectral information in their scale estimation process, and croplands with various sizes could be hardly extracted in one segmentation result, the authors improved a MS based approach to tackle this problem. The main consideration lies in two aspects: ① A local variable scale parameter estimation method is proposed; ② The model function for local variable scale is established for MS filtering. The proposed approach mainly consists of 3 parts: ① With the objective of comprehensively considering the response variation of different bands, the diagonal scale parameter matrix is adopted in the kernel function of MS filtering, and it is combined with sample point estimation model to derive the iterative function for variable scale MS filtering; ② For the purpose of increasing automation of the proposed method, local spectral variation and edge strength information are utilized to design a new local scale parameter estimation method; ③ For obtaining the final segmentation, the filtering result is used as input for the fractal net evolution approach (FNEA) which is a spatial clustering method. Two scenes of HRI acquired by RapidEye and OrbView3 were employed for experiment, and the results show that the proposed method can optimize the accuracy of cropland HRI segmentation.

To achieve better effect of remote sensing image fusion, the authors put forward a multi-scale sparse image decomposition method based on morphological component analysis (MCA). It combines curvelet transform basis and local discrete cosine transform(DCT)basis to form the decomposition dictionary and controls the entries of the dictionary so as to decompose the image into texture component and cartoon component. From the aspect of the amount of information, a remote sensing image(RSI)fusion method based on multi-scale sparse decomposition was proposed. By using sparse decomposition, the effective scale texture component of high resolution RSI and cartoon component of multi-spectral RSI were selected to be fused together. Compared with the classical fusion methods, the proposed fusion method gets higher spatial resolution and lower spectral distortion with a little computation load. Compared with sparse reconstruction fusion method, it achieves a higher algorithm speed and a better fusion result. Therefore,the proposed image fusion method based on multi-scale sparse decomposition has certain application value.

This paper presents a method for combining the color feature and target decomposition characteristics so as to study the classification of polarimetric SAR. It makes up decomposition feature vector by polarimetric target decomposition and then, through the pseudo color enhancement method, obtains the false color image of polarimetric SAR data representation; after that, it extracts color histogram from the pseudo color images to make up the color feature vector, thus providing additional information for further land classification. Classification experiments were performed at different feature vectors by using RadarSat-2 polarimetric SAR image. In addition, the quantitative and qualitative comparison analysis was conducted with classification results. The experimental results show that the addition of the color feature can effectively improve the classification accuracy of polarimetric SAR images.

For SPOT 6 satellite remote sensing image, a method based on rules was used to extract buildings. Firstly, the authors analyzed the extraction effects of every rule attribute and made the rule extract buildings based on the effect. Then the authors compared the methods of K-means clustering, K nearest neighbor (KNN), support vector machine (SVM) and neural network with the method used in this paper during the research. The precision evaluation of building extraction result shows that the accuracy of this method based on rules is higher than that of other methods. This method relieves the problems of the salt and pepper phenomenon and the same spectrum with foreign bodies, and provides some technical support for the wider application of SPOT 6 satellite images in the future.

Retrieving the required remote sensing images effectively and accurately is the kernel of a remote sensing retrieval system. In this paper, the authors proposed a direction based retrieval model based on context-sensitive Bayesian network(CSBN). In addition, an approach was also proposed that is suitable to retrieving urban area images according to the characteristics of urban areas. Initially, the proposed approach retrieved the candidate images based on CSBN. Then, the proposed approach obtained the final retrieval result containing the high level semantic concept “urban area” according to the average high frequency signal strength(AHFSS)of the candidate images. In order to make sure the direction relationships inside the image, the authors used the four directions of northeast, northwest, southeast and southwest to describe eight kinds of directions, which effectively reduced the time complexity of the algorithm. The experimental results show that the proposed approach can effectively describe the semantic concepts of the stored remote sensing images, and thus has higher retrieval precision and efficiency than the original context-sensitive Bayesian network based approach, thus proving that the proposed approach can meet the users’ requirements.

To develop a method for collecting spatial information of forest change to update forest resources database, the authors tested a forest change detection in an area in Shangsi County of Guangxi where the forest cover changed frequently and rapidly and had a lot of change parcels most of which were small patches. ZY-3 and GF-1 satellite remote sensing images and the thematic map of forest distribution composed of sub-compartments were used as the data sources, the length of change vector was measured by Mahalanobis distance, Euclidean distance and relative error distance, and the optimal threshold was determined by the objective function. In addition, the object-based change vector analysis (CVA)was used to detect the forest change based on the sub-compartment. The results show that the detection results based on the Mahalanobis distance and Euclidean distance are not ideal, for they have high omission rate and commission rate but low total accuracy and small kappa coefficient. The detection result based on the relative error distance is the best among the three detections, for its omission accuracy (21.0%) and the commission accuracy (32.5%) are the lowest in the three detection, and its total accuracy (89.6%) and its Kappa coefficient (0.664) are higher than the two other detections. False detections are usually found in the old forest land, construction area, road and some other places, and the commission objects are found in various land types.

Generally, the conservation of Wolong Giant Panda Habitat (Natural World Heritage site) is significant for the sustainability of the rare species of Giant Panda. As we know, forest coverage can be an essential impact on the suitability of the habitat. Owing to the all-weather, all-day operation capability of radar systems, in this study, the authors investigated the performance of Synthetic Aperture Radar (SAR) images in fine mapping of forests using multi-temporal/polarization PALSAR data. The authors firstly corrected the radiometric distortion of SAR data induced by the cliffy topography; then the authors selected 5 different temporal acquisitions for the forest mapping using the Support Vector Machine (SVM) approach. 5 multi-temporal/dual-polarization indexes, i.e., HH_m,HV_m,TSD,HH_m-HV_m and HH_m/HV_m, were applied for the training and classification. Experimental results demonstrated that the combination of HH_m, HV_m, TSD and HH_m-HV_m derived an optimal classification (e.g., total accuracy and user accuracy of forest/non-forest were 86.90%,82.34% and 92.83%, respectively), better than the single-temporal/polarization mode (total classification accuracy of 55.47%). This study shows the effectiveness of multi-temporal/polarization SAR data in forest fine mapping, particularly in the monitoring and evaluation of natural heritage sites located in cloudy and rainy environments.

Data field can describe the correlation between data objects, and is a simulation of interaction between particles in physical field. Potential value of a data object in data field can effectively represent the spatial interactions of its neighborhoods, and it can do so for pixels in high resolution remote-sensing image. In this paper, the authors propose a method for residential area extraction from high resolution remote-sensing image using data field and density clustering. The major steps are as follows: the generating of a high resolution remote-sensing image data field; the calculation if potential value for each pixel in this field to obtain a new feature image; the segmentation of the feature image via watershed segmentation, and the calculation of centroids of segmentation results; the clustering of all the centroids into different clusters based on the density, with the extracted residential area composed of target clusters. Compared with existing relative methods of residential areas extraction for high resolution remote-sensing images, the experimental results suggest that the presented method is robust and efficient.

This method was illustrated by applying it to simulating bottom-of-atmosphere(BOA) radiance for two 4.3 μm absorption bands of the SPIRIT-Ⅲ sensor by using MODIS data of band 23, which is close to 4.3 μm. First, surface emissivity images in these two 4.3 μm absorption bands were simulated using band translation models. Second, analytic model of BOA radiance was deduced based on an existing analytic model in mid-infrared bands, and then it was combined with simulations from radiative transfer model MODTRAN to calculate parameters of the atmospheric effects for these 4.3 μm absorption bands. Finally, based on the proposed analytic model, BOA radiance in SPIRIT-Ⅲ’s two absorption bands can be generated from surface emissivity, temperature, atmospheric effect parameters and SPIRIT-Ⅲ’s spectral response functions. Accuracy assessment on the simulation results shows that this method can produce surface emissivity and BOA radiance with errors less than 6% and 0.02%, respectively. Therefore, the method proposed in this paper can effectively and precisely simulate BOA radiance for the 4.3 μm absorption bands, and provide radiance datasets for the at-sensor radiance simulation and sensor imaging simulation.

The investigation of rare metal mine is very important due to its strategic position in recent years. In this paper, the investigation of the No. 414 superlarge rare metal mine (tantalum, niobium, lithium) in Yichun of Jiangxi Province was conducted using the satellite remote sensing data of ZY-3, IKONOS and Worldview-2. Firstly, the interpretation keys of each feature in the mine were built, and the mining situation was delineated with the help of the mining right data and geological rock data. Then, the difference between rare metal mining and kaolinite mining was studied using texture information on the image. Finally, the landslide of the mine was estimated using the three-dimensional representation of the ZY-3 satellite data, and the application comparison of the three kinds of satellite data was analyzed. The results show that high resolution remote sensing processing can provide a good method for quick and accurate mining status investigation and geological environmental analysis of rare metal mines,thus having further popularization and application value.

In this study, the authors used the data of TM and ETM remote sensing images, geographical condition census data and water conservation census data obtained in 2000, 2008 and 2013 and multi-year statistic data as the main data sources. Using remote sensing and GIS technology, combined with the field survey, the authors systematically studied the types, sizes, composition and distribution of wetland resources in Gansu section of Heihe River Basin. The results show that the area of wetland in the study area is decreasing, slowdown was faster before 2008, and slowdown is significantly slower since 2008. The change of wetland resources results from natural factors and human activities. Among them, the impact of the main factors on the changes of wetland resources in the area is precipitation, temperature, upstream water, discharge and permafrost. Human factors are becoming more and more obvious in the change of wetland resources as an external force. The change of population quantity and the intensity of the land reform and utilization are the most active and direct factors influencing the wetland landscape. Government decision making and the restoration project of wetland protection have played a role in promoting the evolution of wetland resources.

In order to analyze the spatial-temporal variations of construction land expansion in Wuhan City, the authors obtained construction land data by using object-oriented classification method based on Landsat data acquired in 1990, 2000, 2005 and 2010. Urban expansion indices, which included expansion speed index (ESI), expansion intensity index(EII) and integrated expansion degree index (EDI), were used to quantitatively analyze temporal variations of construction expansion. And standard deviational ellipse(SDE)method was employed to investigate the spatial dynamics. The results show that the area of construction land in Wuhan City increased continuously, with a total increase of 749.9 km² in two decades. The proportion of construction land expansion in earlier ten years, middle five years and later five years was 19.06%, 33.33% and 47.61%, respectively. The values of ESI and EII tended to increase, and the expansion degree was especially strong during 2005―2010 period. The spatial distribution of the construction land was discrete. Construction land partially extended in the earlier decade and tended to expand evenly in the middle five years. And the distribution of the construction land in the two periods exhibited a spatial pattern “spanning northeast to southwest”. In the later five years, construction land expansion was balanced in every direction with no directivity. This study may provide decision support for the rational expansion of construction land in Wuhan City.

Rural homestead ownership determination using unmanned aerial vehicle(UAV) has the advantages of high efficiency and low cost. Low-altitude UAV photogrammetry experiment was conducted to test its positioning accuracy. The simple UAV platform and ordinary digital camera effectively reduced the project cost. Through structure from motion, an approximation of the image exterior orientation elements was computed, and then a self-calibration bundle adjustment with additional parameters was undertaken, which significantly improves the accuracy of low-altitude UAV photography and thus has important practical value. By analyzing the main factors affecting the accuracy of UAV photogrammetry, the authors put forward some suggestions about the control of low altitude UAV photographic measurement accuracy.

Fractional vegetation cover (FVC) is a critical indicator for vegetation and eco-environment. It is frequently used as a basic input for hydrology, meteorology and water-soil protection studies at regional or global scales. Southwestern China is an important ecological barrier and the major water supplying area in China. It is important to carry out the study of changes of regional fractional vegetation cover for the protection of eco-environment. In this paper, based on the MODIS-NDVI data obtained from 2000 to 2014, the authors estimated fractional vegetation cover of southwestern China by using the method of dimidiate pixel model, and analyzed the spatial-temporal variation characteristics of the FVC. The results show that, in the past 15 years, the FVC of southwestern China has shown an increasing trend in general but decreased in some meadow areas over the northwest of the study area and the urban expanded areas. In different kinds of ecosystem types, the forest shows the largest average increase of the annual maximum FVC (0.096 2 a^-1, p<0.05), while the grassland shows the smallest increase (0.031 1 a^-1, p=0.582). Fractional vegetation cover has increased in different degrees in most seasons in the past 15 years in southwestern China, with the increase in autumn being most rapid (0.229 8 a^-1) and has most significant trend (p<0.01), followed by spring. For better understanding the effects of climate change on FVC, the correlation coefficients of climatic factors and the annual maximum FVC in different temporal durations were calculated. The results suggest that the annual maximum FVC is significantly related to accumulated precipitation of autumn and mean temperature in summer, showing correlation coefficients of 0.320 and 0.281. In addition, human activities are also important causes resulting in FVC change and the effect has increased in both positive and negative aspects.

At present, the extracting method for remote sensing alteration anomalies from principal component image relies mainly on the data’s normal distribution, without considering the nonlinear characteristics of geological anomaly. To tackle this problem, the authors have proposed the fractal dimension-change point method(FDCPM)in this paper. By calculating the self-similarity and mutability of alteration anomalies with fractal dimension-change point model, the critical threshold of an alteration anomaly was acquired quantitatively. The realization theory and access mechanism of the method were elaborated by an experiment with ASTER data in Fangshankou,Beishan,and the results of the proposed method and traditional method (de-interfered anomalous principal component thresholding technique,DIAPCTT) were compared with each other. The results show that the FDCPM has a relatively high extracting precision than the DIAPCTT for three alteration minerals in the experiment. In this experiment, the accuracy of three alteration minerals could reach over 83%. Moreover, the distribution of remote sensing alteration anomalies agrees well with a large amount of evidence from the geochemical anomaly and the heavy sand anomaly. What’s more, the known polymetallic ore spots and mineralized spots fall in the zone of remote sensing alteration anomaly or at its edge. All the results mentioned above show that the FDCPM is one of the effective distinguishing methods for the geological background and the remote sensing alteration anomaly in the future.

Retrievals of aerosol optical depth (AOD) is the key data sources to carry out the study of regional air quality. The study of a long-time series of AOD needs the temporal spread and spatial expansion of AOD stripe products, involving a series of complex and professional data processing. In order to help the non-remote sensing professional researchers to correctly use the data, NASA has developed a web service workflow-based data visualization and analysis system- GIOVANNI, but this system has not yet been widely used in China. In this paper, choosing Beijing, Shanghai, Guangzhou and the western region of China as the study areas, the authors interpolated the missing data based on the variation characteristics of multi-year time series. On such a basis, the authors carried out a comparative analysis of the monthly values of AOD in main cities and the western region of China from 2000 to 2014 based on GIOVANNI. The results show that, compared with things in the western region, the averages of multi-year values of AOD in Beijing, Shanghai and Guangzhou were significantly higher, of which Shanghai was the highest, followed by Beijing. AOD in Beijing and Shanghai had significant seasonal differences, and exhibited the high levels in spring and summer and the low levels in autumn and winter. Although the averages of annual values of AOD in Beijing, Shanghai and Guangzhou showed anon significant trend from 2000 to 2014, the time series of AOD in Shanghai and Guangzhou had similarity.

Ecological environment degradation and sustainable development of wetland has become the focus of the current wetland research, it is necessary to study on wetlands monitoring dynamic changes.The study investigated wetland type and dynamic changes in Hainan Dongzhai Reserve based on three high resolution remote sensing images recorded in 2002, 2008 and 2013, respectively. The results indicated that there was little change in wetland types during 2002 to 2013. Coastal water, mangrove and culture pond were the main types of wetland in this area. The area of natural wetland trended to decline, while the total area and the numbers of individual wetland to increase. The most increased wetland types were the area of tidal creek in proportion and the culture pond in individual numbers. The changes were discovered frequently in edges of landward edge in which human activities were extensive.In aspect of the area of wetland, the dynamic degrees of the wetland changes during 2002 to 2013 in this area were small. The types of coastal water mangrove and tidal creek mainly trended to be stable. While concentrated in the types of reservoir, sandy tidal wetlands and culture pond, the dynamic degrees were big, and they were the type of wetland with swift changes.

In this paper, the authors studied restoration and management situation in coal mine subsidence areas using remote sensing images. According to the research objective, the coal mine subsidence area in Shandong Province was chosen as the study area. Multi-stage optical remote sensing images and the technical method which included indoor research and field survey were used. Some conclusions have been reached: ① Coal mine subsidence in this area was very serious, by 2014, water area had reached 13.62 km²; ② Coal mine subsidence was developed rapidly in this area, from 2006 to 2014, coal mine subsidence water area increased by 7.78 km² totally, and the growth rate was 133%; ③ Restoration and management situation was good in study area, its area reached 11.70 km², and its rate was 85.90%; ④ Original restoration and management project could be destroyed by flow up coal mine subsidence, and the area needs restoration and management again; ⑤ In remote sensing images, restoration and management area is easily confused with other features, and hence it is necessary to use multi-temporal remote sensing data for careful comparison.

In this paper, the authors studied land use changes of the eastern route of the South to North Water Transfer Project in Jiangsu Province based on land use data from 2000 to 2010, and simulated and predicted land use patterns from 2010 to 2020 based on CA -Markov model. “Pressure-State-Response” (PSR) model was used to establish the evaluation index system of land ecological security in the hope of evaluating the ecological security of land use in different periods. Some conclusions have been reached: 1) The main land use type in the area was cultivated land and the area of forest land, water was enlarged, and a large amount of cultivated land was changed into urban and rural areas. 2) The simulation results of land use pattern in 2010 based on CA-Markov model indicated that the correct grid number reached 99.1% compared with the actual one and the Kappa coefficient reached 0.99, which could reflect change trend of land use pattern in the area. 3) The ecological safety of the region showed gradually improvement and, for the year 2000, 2010, 2020, the land ecological security index was 0.51, 0.68 and 0.73, respectively achieving the early warning, sensitivity and good grades, which indicated coordinate development of population, resources, environment and society.

Urban features classification is based on hyperspectral characteristics and high-resolution image segmentation objects. After the removal of bad lines and Smile effect, FLAASH atmospheric correction and 155 Hyperion bands were used in this study. Spectrum feature was used to determine objects recognition suitable spectral resolution, and after Hyperion dimensional reduction, 21 wide-bands were generated. Utility wavelet fusion was performed, and IKONOS high-resolution objects were generated by multi-resolution segmentation. On the basis of hierarchical analysis classification method for segmentation objects, fuzzy membership function of the vegetation red edge effect and the water absorption characteristics in the near infrared were used to complete first level classification. The larger distance of 10 Hyperion bands was used as feature bands, and the second level classification was completed by standard nearest neighbor classification. 9 types of urban features were separated. The classification results are better than the maximum likelihood classification and spectral angle mapper.

Radio-frequency interference (RFI) over eastern Asia land was detected and analyzed using one dimensional variational retrieval (1-DVAR) convergence metric method from AMSR-E (the advanced microwave scanning radiometer - earth observing system) Leval 2A measurements during July 1-16, 2011. And then its influence on the retrieval of surface parameters was studied. It is found that the RFI signals are detected both at C and X band channels of AMSR-E over eastern Asia, and the signals are most densely concentrated in industrial zones, scientific research centers, metropolises, airports and highways. Moreover, RFI signals at C and X band normally do not coincide with the same distribution area. AMSR-E RFI over eastern Asia land exists along both horizontal and vertical polarization channels. Furthermore, the intensity of AMSR-E RFI varies with the earth azimuth angle of the satellite; measurements are contaminated by RFI only when the spaceborne microwave radiometer is within some earth azimuth angle range. Lastly, it is also found that retrieved land parameters have large deviations from RFI contaminated microwave measurements. Therefore, it is expected to detect even weakened RFI effectively prior to retrieving land surface parameters from passive microwave remote sensing measurements.

Hyperspectral Data (CASI/SASI) play an important role in identification of alteration zones for its abundant spectral information and high spatial resolution. In this study, the analysis of metallogenic geological conditions in Fangshankou-Laojinchang area indicates a good potential prospecting of the intrusion-related gold deposits. Anomaly extracting results of CASI/SASI indicate that seven types of alteration minerals, which are Al-high sericite, Al-middle sericite, Al-low sericite, limonite, dolomite, chlorite, epidote and calcite, are mainly developed in this region. Alteration mineral distribution and abnormal composition analysis show that Al-high, Al-middle sericite and limonite are developed in the Jintanzi deposit, which constitute the indicating alteration mineral assemblage for gold mineralization. Then the authors sieved the hyperspectral abnormal information and delineated the Fuangshankou as the ore-prospecting region. Meanwhile, multiple quartz veins which have gold mineralization were found in the field verification process. Therefore the selected indicating alteration mineral anomaly combined with the CASI/SASI is effective for the mineral prospecting.

Using middle and high resolution remote sensing data such as WorldView2, IKONOS, QuickBird, ASTER and ETM⁺, and their processing methods such as de-relatedcalculation, ratio calculation, principal component analysis and image fusion, the authors delineated a siderite-hematite mineralization belt along Heiqia Pass in West Kunlun metallogenic belt on the basis of the field survey. The belt occurs in the Lower Silurian Wenquangou Group and stretches 120 km long northwestward, and has been eroded by the rock mass in the northwest part and truncated by Kalatage fault in the southeast part. Its ore-bearing layers remain stable, and its continuity in formation strike and dip direction is very good, so the belt is favorable for mineral resources investigation. The results of the survey show that geological survey with remote sensing technology is one of indispensable methods in regional geological and mineral resources survey, and will play an important role in the geological prospecting in western metallogenic belts, especially in the complex and dangerous regions.

In this study, GF-1 16 m multispectral images were used as data source, the spectral characteristics of each type of sandy land and its change characteristics of time series NDVI were analyzed, the sandy lands were classified by the GF-1 image at a single time, and time series NDVI data were compared with each other separately; on such a basis, the classification accuracy was evaluated. The results showed that the accuracy was 73.34% and Kappa coefficient was 0.7 by only using single time original data in growing season; however, the accuracy was increased to 92.04% by joining the time series NDVI data, with Kappa coefficient raised to 0.87; the accuracy was 81.44% and Kappa coefficient was 0.77 by using the time series NDVI data combined with non-growing season data, thus improving the classification accuracy obviously. It is indicated that GF-1 time series NDVI data have a huge application potential in the sandy lands classification.

Using high resolution remote sensing data obtained from 2013 to 2015, the authors monitored the subsidence disaster and its distribution range in the coal mining area of Shandong Province continuously, and it was found that the coal mining subsidence area decreased year by year. Based on the comparative analysis of the distribution characteristics and the mode of the subsidence disaster and the development history of coal mining, the authors established the dynamic monitoring technology of coal mining subsidence. According to the monitoring results, the distribution range, scale, severity and trend of the subsidence area in the province were determined, and 12 superlarge subsidence zones, 8 large subsidence zones, 50 medium-size subsidence zones and 12 small subsidence zones were delineated. The changing trend was predicted on the basis of the analysis of the above information for the past three years. It is found that 30 subsidence zones become better, 31 subsidence zones are stable, and 21 subsidence zones become worse. At the same time, this paper analyzes the influencing factors on coal mine subsidence disaster, discusses the management time, and puts forward some countermeasures and suggestions.

High geometric correction precision of GF-2 is a prerequisite for its widespread application. In this study, the authors selected Fuzhou as an experimental area. The ground control points (GCPs) were measured by Beidou satellite navigation system (BDS) and GPS respectively in the experimental area, which were used for geometric correction of the GF-2 panchromatic (PAN) image. The rational function model (RFM) was corrected by block adjustment with ground measurement point. The authors validated control point accuracy, distribution, as well as correction method for GF-2 panchromatic image correction, and analyzed the potential application to GF-2 PAN imagery geometric correction. The results show that a few control points can eliminate geometric error of GF-2 PAN imagery system. Affine transformation can reach the highest correction precision among three correction methods. Plane root mean square error (RMSE) of GPS check points using affine transformation is 1.49m and plane RMSE of Beidou RTK check points using affine transformation is 1.51m. In the two measuring modes of Beidou, Beidou RTK precision can satisfy the demand of GF-2 PAN imagery correction. The results show that, with a few high precision control points using GPS and Beidou RTK, GF-2 PAN imagery can reach high geometric correction accuracy and satisfy the demand of practical application.

In order to assess the fire-fighting effects in the Wuda coal field of Inner Mongolia, the authors adopted two Landsat 5 images(acquired in 2008 and 2011)and two Landsat 8 images(acquired in 2013 and 2015)as data sources which respectively represented before, during and last stage fire-fighting activities, generated four land surface temperature maps by using mono-window algorithm and recognized coal fire areas based on self-adaptive gradient-based thresholding(SAGBT)method. The results show that 75% overlapping between identified coal fire areas and in situ sampling coal fire areas. The coal fire evolutions on spatial distribution revealed that coal fire area presented remarkable downfall trend from 1.194 km² in 2008 to 0.873 km² in 2015. Overall, coal fire areas were reduced by 26.88% due to performing fire-fighting activities; however, more measures should be strengthened by local administrators because there remain residual 73.12% coal fire areas.

Cultivated land quality and planting suitability constitute a significant basis of cultivated land resource and agricultural sustainable development. Volumes of spatial data sets accumulated by crop suitability evaluation and land productivity surveying were dispersed over different organizations. Therefore, based on OGC standards, the authors investigated the sharing mechanism for spatial data sets of cultivated land quality and planting suitability. A regional sharing services platform for cultivated land integrated information was built. With Fujian Province as an example, an application system was built. The platform improves the sharing and availability of the result of cultivated land researches and provides significant scientific data support for agricultural sustainable development.