The target motion information extraction technology described in this paper uses satellite remote sensing to detect ground moving targets and estimate its motion parameters. It is one of the important application directions of remote sensing images and has been widely used in traffic monitoring and military remote sensing. As an excellent tool for the study of large-scale target motion characteristics, the high-resolution optical satellite image has more obvious texture features and richer information. After summarizing the research progress of moving targets in optical satellite imagery, this paper describes the methods of moving target detection and motion parameter estimation according to the process of target motion information extraction from high-resolution optical satellite image. Meanwhile, the principle and ideas of a novel method which is based on sequence panchromatic satellite images to detect moving target are introduced. In the end, based on analyzing the weaknesses of existing target motion information extraction research in data source and algorithm, it is pointed out that the target motion information extraction is developing towards automation, intellectualization and real-time.
Population spatial distribution information is the basic information in the study of geography, resources, sociology and other disciplines, and hence is of great significance in practical applications such as urban planning and emergency rescue. The population distribution can be well simulated by using the auxiliary data of physical geography and social economy data. Nighttime light data reflect the distribution of population comprehensively. Compared with traditional remote sensing data, it has the advantages of convenient data acquisition, small data volume, wide coverage and fast data update. With the development of DMSP-OLS, NPP-VIIRS and other platforms, the study of population spatial distribution based on continuous archiving nighttime light data has attracted the attention of scholars, and a rich research result has been formed at regional-scale population estimates and grid-scale simulations of population distribution. Nevertheless, there are also problems in data correction, data fusion, scale selection and precision verification. Therefore, with the expectation of providing references for other researchers, this paper elaborates the nighttime light data characteristics and access platforms, summarizes the methods and models of population spatial distribution based on night lighting data, and analyzes the problems and solutions in the research. Finally, the important development directions in the future are discussed.
Urban expansion caused by urbanization brings many problems to the social and ecological environment. Monitoring urban change is an entry point to solve these problems. As an important indicator of urban expansion, the impervious surface has become a hotspot of research. The acquisition of impervious surface data and its time series variation analysis are the core of the current study. Compared with early planar map-based impervious surface extraction, remote sensing has been widely used in impervious surface research due to its continuous, rapid and extensive observation of the ground. Multi-source data fusion and multiple retrieval method make constant progress in remote sensing-based retrieval of ground impervious surface percentage, and the focus of the study is gradually shifted from the classification map of impervious surface to the quantitative retrieval of impervious surface percentage. In this paper, the authors summarized the methods of remote sensing retrieval of sub-pixel impervious surface percentage from the perspectives of single period and time series, analyzed the advantages and disadvantages of the retrieval methods in detail, and briefly described and compared the common precision validation method. Finally, the authors summarized the problems existing in the current remote sensing retrieval methods, proposed the corresponding solutions, and pointed out the trend of development in the future.
Sky view factor (SVF) is a numerical value that describes the three-dimensional characteristics of the city, and hence it is widely used in urban heat island effect, urban energy balance and some other fields. In this study, taking the national sport stadium area as the study area, the authors calculated the SVFs based on the digital surface model (DSM) in the urban areas. Furthermore, fisheye photos were used to extract the in-situ SVFs. Finally, the authors compared and analyzed the results between these two methods. The main conclusions are as follows: ①The SVF value calculated based on the DSM is affected by the search radius and the search direction number. SVFs are decreasing with the increasing of the search radius and the search direction number; ②Specifically, when the search direction is 32 and the search radius is 80 pixels, the value of RMSE is 0.064 (noting that it is the lowest value). The SVF values calculated using the DSM are most similar to those calculated using fisheye photos; ③A certain correlation is observed between the SVF values calculated using the fisheye photos and those calculated based on the DSM, indicating the feasibility of using DSM to calculate the SVF in large urban areas.
Some existing remote sensing image segmentation methods do not take the edge feature into consideration, therefore, an edge-incorporated multi-scale segmentation algorithm based on weighted aggregation (EIMSSWA) is proposed. Firstly, the edge features of adjacent primitives are generated by counting the gradient strength and gradient direction on the common edges. Secondly, these features are infused into the similarity measurement of the adjacent primitives in segmentation by weighted aggregation, so as to improve the segmentation. Finally, the segmentation of the proposed method is compared with segmentations of eCognition as well as segmentation by weighted aggregation (SWA) a. The results demonstrate that the EIMSSWA method is capable of gaining more accurate and more reasonable segmentation.
In order to accurately extract a large range of plastic greenhouse distribution information, the authors took Changzhou City, which is located in the Taihu Lake basin, as the study area, used Landsat8 imagery, employed plastic greenhouses spectral analysis and spectral separability analysis, selected seven multi-spectral data and one thermal infrared datum of Landsat8 image and three remote sensing indexes(NDVI, NDBaI and MNDWI)and, based on Logistic regression analysis, constructed a new plastic greenhouse index (NewPGI). Accuracy verification results show that, in the sample area, the high-resolution image of the plastic greenhouse reference map shows that NewPGI’s overall classification accuracy is 94.9%, and Kappa coefficient is 0.74. Throughout Changzhou, the verification sample points were selected based on the Google Earth image. The overall accuracy of NewPGI is 91.28%, and the Kappa coefficient is 0.78. Compared with the existing plastic greenhouse index, NewPGI can better extract plastic greenhouses under complex surface coverage.
At present, most of the bands or spectral indices used in the high temperature target remote sensing recognition researches only involve qualitative analysis, with the lack of quantitative evaluation indicators and screening methods. In order to establish a universal band screening principle and screening method for judging indicators so as to achieve effective identification of high temperature targets, the authors, according to the idea of variance analysis, constructed a separability metric to screen characteristic bands between high temperature targets and various types of normal temperature objects respectively, determined some effective bands for high temperature targets identification, constructed identification indices with the spectral characteristics of the ground objects, and screened the optimal one. The result shows that the optimal spectral index determined by quantitative screening can effectively distinguish high temperature targets from most normal temperature objects. On such a basis, by using the other optimal identification index between high temperature targets and their confusing color steel objects to identify once again, the recognition accuracy could be improved further, and recognition accuracy of high temperature targets reached 95.4% and 97.6% , respectively.
In this paper, the multi-dimensional adaptive weighted filter (AWF) is used to filter the hyperspectral image with a certain dimension which are reduced by the feature extraction method based on spectral dimension. Then, the filter results obtained on all scales are hierarchical fusion into a new image, and the hierarchical fusion framework is designed. These treatments make the essential spatial and spectral features in hyperspectral images extracted effectively, so the classification accuracy is improved. The principal component analysis (PCA) algorithm is integrated into the framework, and a hierarchical fusion-principal component analysis (HF-PCA) algorithm is proposed. This method not only reduces the redundancy between bands, but also weakens the internal differences of the samples and improves the classification accuracy of hyperspectral images. Experimental results on the Indian Pines and Salinas databases demonstrate that the classification accuracy obtained by the HF-PCA algorithm is significantly higher than that of other algorithms, even when the number of training samples is small, and the maximum value of the overall classification accuracy is 86.73% and 95.01%, respectively. The classification accuracy of hyperspectral images is improved effectively.
In view of the lack of theoretical research on the extraction and inversion of hyperspectral features in ground-sky synchronization, the authors, in combination with the principle of “maximum density within class and maximum distance between classes”, studied the separability and importance selection of the spectra for different ground objects in different spectral regions, proposed an improved projection pursuit classification method, and realized the projection pursuit method based on weighted feature band. In the case study, the spectral and PHI hyperspectral images of different ground objects in the experimental area were collected synchronously and, combined with the measured spectra on the ground, constructed different classification rules around the strategy of overall optimization and local optimization. It was applied to the classification of PHI hyperspectral image to extract the information of vegetation and non-vegetation and to subdivide more than ten different kinds of ground objects. The results show that 8 spectral areas are important separability bands of different vegetation: 420 nm, 520 nm, 570 nm , 610 nm, 660 nm, 690 nm, 715 nm, and 810 nm. The classification results have the advantages of strong stratification, clear outline of vegetation and avoiding shadow influence.
Superpixel segmentation has become a new hotspot in remote sensing image preprocessing, but it has the problem of over segmentation. To solve this problem, the authors propose a high resolution remote sensing image segmentation method combining superpixel and graph theory. First, the simple linear iterative clustering (SLIC) algorithm is used to divide the image into superpixels, then the superpixels are merged by the graph theory algorithm, the local variance corresponding to the combined number of the merged numbers are calculated, and the appropriate image segmentation number is determined. Finally, according to the appropriate image segmentation number, the graph theory algorithm is used to recluster and merge the superpixels. Four groups of remote sensing images of different scenes and different spatial resolutions were selected as experimental data. The qualitative and quantitative analysis of experimental results was evaluated. Experimental results show that the proposed method can effectively overcome the effect of over segmentation results and achieve good segmentation results.
In this study, domestic GF-1 WFV data were used as the data source, SiB2 model was used to estimate the LAI of forest vegetation in Mohe County of Heilongjiang Province and the value was compared with the estimation result of the enhanced vegetation index (EVI) linear model. Estimation results of the two models were combined with the synchronous ground LAI data for accuracy evaluation. The results show that the coefficient of determination (R 2) of the LAI estimated by the EVI linear model is 0.582, and its root mean square error (RMSE) is 0.701. The R 2 of the LAI estimated by the SiB2 model is 0.798, and its RMSE is 0.358. Compared with the performance of the EVI linear model, the results estimated by the SiB2 model are improved on both R 2 and RMSE. The results show that the SiB2 model is more suitable for LAI inversion of forest vegetation in the study area, in combination with the high spatial resolution GF-1 WFV data.
According to the spectral features of domestic ZY-3 remote sensing images, the formula of LBV transformation for ZY-3 is proposed and deduced, and the feasibility of improving the quality of ZY-3 remote sensing images is testified. At first, based on the characteristics of ZY-3 remote sensing images, the spectral information of nine types of typical ground features were selected, and regression coefficients were used to calculate regression coefficients. Then, the three components of L, B, V of ZY-3 satellite images were calculated according to the characteristics of the typical ground features space (bare land, water body, vegetation), color space (red, green, blue) and the space of LBV variables (the general radiance level of the ground objects, the visiable - infrared radiation balance, the band radiance variation vector). Finally, the experiments of ZY-3 remote sensing image in Ningde City of Fujian Province were carried out, and quantitative analysis was conducted to evaluate the experimental results. Firstly, the results show that, in the aspect of the visual effects, compared with the original image, the transformed image is more clear, and the details are more abundant, and thus can contribute more to the determination and identification of subsequent features. Secondly, through the LBV transformation, the image information entropy is 6.21, the average gradient is 4.71, the deviation coefficient is 0.46, and the quality of the remote sensing image is better than other transformation methods. Thirdly, by classifying the LBV image, the overall accuracy is up to 89.71%, and the Kappa coefficient is the highest, reaching 0.875 3. The classification accuracy is higher than that of other transformation methods. Therefore, The LBV transformation can improve the quality of ZY-3 remote sensing image, and it can be applied to ZY-3 remote sensing image processing and information extraction.
The recognition and extraction of seismic targets in high resolution post-earthquake images pose great challenge to the traditional image classification method. This paper introduces an object-oriented classification method for high resolution post-earthquake images classification, which integrates fractal texture features into a gravitational self-organizing map (gSOM). The method can be summarized as follows. First of all, the mean shift (MS) segmentation algorithm is adopted for initial segmentation in order to obtain homogeneous geographical objects, and the objects are regarded as the basic classification units in the subsequent process. Secondly, the characteristics of objects are quantified by the adaptive combination of the spectral bands and the fractal second order statistics as the texture information extracted from the original seismic image. Finally, the objects as classification units are clustered under the gSOM. For the purpose of controlling the uncertainty in the classification results, these various clustered results are assembled by the consensus function with the least cost. The qualitative and quantitative experiments on the Wenchuan County seismic images demonstrate the effectiveness and accuracy of the proposed method, which not only maintains the integrity of large damage targets, but also reflects details of the small targets at the same time. Also, the method shows the potential in the new technology for high resolution post-event image classification.
The geo-positioning accuracy is an important parameter in evaluating the geometric quality of the satellite imagery. In this paper, integrating the stationary orbit and area-array image characteristics of GF-4 satellite and taking Google Earth images as geometric references, the authors analyzed the geo-positioning accuracies of multispectral camera images during and after the commissioning phase period. This study was concentrated on the relationships between the image geo-positioning accuracy and the sun azimuth angle, the sun altitude angle, the satellite azimuth angle, the satellite zenith angle, and the satellite attitude angle. Meanwhile, the influence of satellite attitude angle and imaging time on the compensation of camera’s attitude constant angle errors was discussed. The results can be used to improve the imaging model and hence are useful for the high accuracy geo-positioning of this type satellite imagery.
In view of the problems of scale effect, spectral diversity and classification feature optimization in the extraction of urban objects information from high spatial resolution remote sensing images,the authors, based on the object-based image analysis method and combined with data mining and machine learning,propose a multi-level segmentation and classification hierarchical model and its feature space optimization method for building extraction. First, according to the multi-scale characteristics of remote sensing information, a hierarchical relationship is set up for the difference of features of ground objects, and then a hierarchical structure based on information segmentation and classification is established based on the characteristics of spectral diversity to define the subtypes of ground objects. After that, the proposed Relief F-PSO combination feature selection method is used. Finally,on the basis of multiscale segmentation and feature optimization, the water surface distribution is obtained based on the random forest model, and finally the building information is extracted by the J48 decision tree algorithm. Experimental results show that the method can utilize a small number of image feature attributes to get high-precision building extraction results.
High spatial resolution surface albedo datasets are of critical importance for weather forecast and global climate change studies. The Chinese Huanjing-1 satellites (HJ-1 A/B) can provide wide swath, short revisit time, and high spatial resolution (30 m) remote sensing observations, and hence can be considered as a perfect input data source for generating high spatial resolution surface albedo datasets. In this study, the authors compared and evaluated two methods for estimating surface albedo from HJ-1 A/B CCD data: the direct estimation algorithm from surface reflectance (DEA-SUR) and the method based on MODIS kernel coefficients (MKC). The visual interpretation and clarity index methods were employed for evaluating the fineness of the images. The results show that the clarity and fineness of imagery were greatly improved by the DEA-SUR and MKC methods, compared with the MODIS surface albedo products. It has been demonstrated that the DEA-SUR method is much better than MKC method in avoiding the mosaic effects. Four sites (US-MMS, CN-Cng, Yingke, and Namco) were used for validating and comparing the DEA-SUR and MKC methods. The results show that the DEA-SUR method and the MKC method have similar estimation accuracies during the snow-free period (root mean squared error (RMSE) is 0.015~0.041). In contrast, the estimation error is much larger during the snow-covered period.
In-depth study of the spatial-temporal change process of the ecosystem service value in Loess Plateau reveals the evolution of the ecological environment in the region, which is of great significance to the improvement and protection of the ecological environment in Loess Plateau. In this paper, based on comprehensively considering the spatial heterogeneity and temporal evolution of ecosystem service, the authors selected the Loess Plateau ecosystem service value dynamic evaluation model to assess the ecosystem service value in Loess Plateau. The ecosystem service value variability index was used to assess the impact of land use change on ecosystem service value and the environmental economy coordination degree index(CDEE) was used to evaluate the coordinating relationship between eco-environmental quality and social and economic development in Loess Plateau. The results show that, from 2000 to 2010, the ecosystem service value of Loess Plateau increased by 48%. This is mainly due to the desertification control in the Loess Plateau which had led to an increase in regional vegetation coverage and a corresponding increase in ecosystem services. The ecosystem service value change rate in grassland was the highest (13.45). The CDEE was 0.115 3 in Loess Plateau from 2000 to 2010, which was at a low level of coordination. This study can provide a scientific basis for ecological environment governance and evaluation of protection effectiveness in the Loess Plateau.
The assessment of habitat quality has important implications for the biodiversity conservation. It can provide scientific basis for the restoration and protection of the ecological environment and the formulation of regional sustainable development policies. Taking the east of Jilin Province as the study area and based on Landsat TM/OLI images and digital elevation model (DEM) data in 2000 and 2015, the authors selected water situation, degree of disturbance, shelters, and food source as evaluation factors. In addition, the analytic hierarchy process (AHP) was used to determine the weight, thus establishing a habitat quality evaluation model and making a dynamic assessment of the habitat quality in the eastern part of Jilin Province. Moreover, the reasons for change were analyzed from the aspects of assessment factors, land cover changes, social economy, and climate change. The results show that the areas of the best and better habitat quality in 2000 accounted for 37.69% and 46.88% of the study area respectively. The areas of the best and better habitat quality in 2015 accounted for 41.20% and 44.66% of the study area respectively. From 2000 to 2015, the area of habitat quality that got better was 9 414.19 km 2, and the area of habitat quality got worse, being 5 627.5 km 2. Land cover change is one of the causes of habitat quality change, and the vigorous development of eco-tourism and returning farmland to forests in Jilin Province constitute the main driving forces for improving the quality of the habitats. The suitable climatic conditions provide a powerful guarantee for the long-term stability of the ecological environment in the eastern part of Jilin Province.
Impervious surface is an important land cover type. Extracting impervious surface from satellite images is crucial for land use and land cover change (LUCC) studies. Although several indexes have been proposed to detect impervious surface, there is a lack of systematic comparative analysis of these indexes. To address this problem, the authors estimated the performance of eight state-of-the-art impervious surface indexes using Landsat8 satellite images. The experimental results show that perpendicular impervious index (PII) performs best, yielding the highest detection accuracy of 89.6%. The accuracies of ratio resident-area index (RRI) and biophysical composition index (BCI) are slightly lower than the accuracy of PII, which are 87.5% and 87.4%, respectively. The accuracies of urban index (UI) and new built-up index (NBI) are 82.9% and 80.0%, respectively. Normalized difference impervious surface index (NDISI), normalized difference built-up index (NDBI), and index-based built-up index (IBI) fail to enhance the spectral characteristics of impervious surface from complex image background, thereby yielding the lowest accuracy (<75.0%). Importantly, the eight impervious surface indexes fail to distinguish the spectral characteristics of impervious surface from large bare land areas and the average detection accuracy is only 71.0%, hindering their applications in bare-land-rich areas.
Extracting surface water like lake water areas from satellite images quickly and accurately has been an important research topic, which is of great significance to the water disaster monitoring and water resource management. Sentinel-2 multi spectral imager (MSI) and Landsat8 operational land imager (OLI) data are two popular medium- to high- resolution data sources that are freely available. Using the Poyang Lake as the study area and employing four popular water indices, i.e., normalized difference water index (NDWI), modified normalized difference water index (MNDWI), automatic water extraction index (AWEIsh) and water index created with linear discriminant analysis (WI2015), the authors extracted water distribution from two types of images respectively. Water extraction results derived from different images and different water indices were analyzed. The accuracy of the water extraction results was evaluated by visual interpretation results of corresponding GF-1 images. The results reveal that, for these two remote sensing images, all water indices can detect most water body successfully. Among these indices, AWEIsh and WI2015 have relatively higher extraction accuracy, reaching 98% and 94% respectively on Sentinel-2 and Landsat8 images. Compared with Landsat8 images, Sentinel-2 images are capable of reflecting more detailed water body information, and the overall extraction accuracy is higher.
It is difficult to identify the relative ages of lithological units only relying on remote sensing images with few reference data. In case of lack of geological references and difficult fieldwork, it is limited to use remote sensing images for interpretation. To solve these problems, the authors firstly introduced the theory of structural hierarchy to geological interpretation and presented a new method to classify the regional structures and small-scale structures extracted from OLI and GF-2 data respectively. To some extent, this method can determine the relative ages of the lithological units, the events of tectonic evolution and the advantageous areas for metallogenesis in the areas with only a few geological reference data. And it provides a new way for exploiting the advantages of high ground resolution of GF-2 so as to promote the development of remote sensing geology in foreign areas.
Northeastern Jeddah area of Saudi Arabia is located in Africa-Arab metallogenic area, which is characterized by rich mineral resources, complex geological structure and magmatic hydrothermal activities, thus having well-developed polymetallic mineralization and tremendous metallogenic potential. For the purpose of further prospecting for polymetallic areas and delineating favorable targets, the authors used the Landsat8 images to interpret linear-ring structures in the whole study area, and employed the “principal component analysis-optimal density segmentation” method by using iron mineralization alteration anomaly to extract polymetallic deposits. Based on the comprehensive analysis, the authors studied the remote sensing image geological characteristics of 7 iron mineralization alteration zones in the study area and selected favorable metallogenic linear-ring structures to predict the favorable location of prospecting in the northeastern Jeddah area of Saudi Arabia. Three iron prospective areas were delineated. GF-2 and Google Earth images were used to verify the prediction of metallogenic favorable sites in northeastern Jeddah of Saudi Arabia. The results obtained further prove that the magmatic intrusion in the annular tectonic zone and the fracture junction are favorable places for metallogenesis in the study area. The results obtained by the authors could provide the reference for the further prospecting prediction in this region.
In order to discuss the regional hazard assessment method of geological disaster, the authors, based on the comprehensive study of hazard-formative environments of Yanchi County, extracted twelve conditioning factors, i.e., strata, lithology, soil, land use type, slope, aspect, topographic wetness index (TWI), stream power index (SPI), distance to river, distance to road, normalized difference vegetation index (NDVI) and precipitation in the evaluation and information of factors by employing GIS with remote sensing data and geological data. Then the judgment matrix of conditioning factors and factor classes were constructed by analytic hierarchy process (AHP) method, and geological disaster hazard index (GDHI) was built. Finally the geological disaster hazard of Yanchi County was assessed and the resulted hazard map was classified into five classes, including very low, low, moderate, high and very high hazard. Meanwhile validation of the hazard map was performed using success rate curve and receiver operating characteristics (ROC) technique. The results are as follows: ①The area percentage of very low and low class accounts for 34% and 28% respectively, mainly distributed among the middle and north of hill region; the area percentage of moderate class accounts for 25%, mainly distributed in the area of Mahuang Mountain, western Wanglejing and two sides of the main roads; the area percentage of high and very high class accounts for 12% and 1% respectively, mainly distributed in the area on two sides of rivers and Mahuang Mountain. ②The area under curve (AUC) value of success rate curve and ROC is 0.77 and 0.89 respectively, which shows a reasonable validation accuracy of hazard assessment. At the same time, disaster density shows the characteristics of a continuing increase in the density values from the very low class to the very high class, and the density of the very high hazard class has maximum value with 1.076/km 2. ③AHP method was successfully used to assess the geological disaster hazard of Yanchi County and AHP is suitable for hazard mapping in this region. The evaluation results could provide a reference for the prevention and control of geological disasters in Yanchi County.
In recent years, China has paid more and more attention to the mineral geology in the northwest frontier area. However, due to natural geography and other reasons, it is difficult to carry out large-scale manual investigation. By collecting and sorting the available data, the authors have found that there are gold, silver, copper, lead and other minerals in the vicinity of Shumu campsite of the northwest frontier area, and hence it is an important metallogenic prospective area of China’s mineral resources. In order to give full play to the advantages and leading role of remote sensing in prospecting in difficult and dangerous areas of Western China, the authors used ASTER remote sensing image data to extract alteration anomalies and controlling factors. On such a basis, various thematic factors which were used to evaluate the metallogenic favorable areas were obtained, and the correlation and usability between thematic factors were investigated. The establishment of remote sensing geological prospecting model and verification through known deposit point information have obtained good evaluation results, which can provide reference for similar study areas in the future.
Mapping the distribution pattern of mangrove species in regional scales with remote sensing technology is of great significance in the investigation, utilization and protection of mangrove resources. In this study, the authors mapped and analyzed the mangrove species distribution based on the spectrum characteristics of mangroves,vegetation index, texture information and shape parameters calculated from ZY-3 high-resolution multispectral images, in conjunction with the mangrove species sample points, which were collected by the unmanned aerial vehicle (UAV). The authors used object-oriented classification method,decision tree and support vector machine (SVM). The total area of mangrove forests in Leizhou Peninsula in 2014 was estimated at 5 949.3 hm 2, much less than the area reported in most previous studies. For each of the districts in Leizhou Peninsula, mangrove forests covered 1 556.0 hm 2 in Lianjiang City, 1 466.1 hm 2 in Leizhou City, 1 168.0 hm 2 in Zhanjiang Municipal City, 734.7 hm 2 in Suixi County, 479.8 hm 2 in Xuwen County and 544.7 hm 2 in Wuchuan City, respectively. Zonal distribution of native mangrove species is significant from sea to land, with Avicennia marina dominated in the low tide level, followed by Aegiceras corniculatum, Kandelia obovata, Rhizophora stylosa and Brugueria gymnorrhiza dominated from middle to high tide level; the exotic mangrove species Sonneratia apetala is mainly distributed on the land side of Avicennia marina in its introduction area. The proportions of each dominate species are Avicennia marina (41.9%),Sonneratia apetala (23.4%), Aegiceras corniculatum (20.9%),Kandelia obovata(5.4%),Rhizophora stylosa(4.8%) and Brugueria gymnorrhiza (3.6%), respectively. The results show that Sonneratia apetala planting in Leizhou City and Zhanjiang Municipal City has achieved remarkable success in the past several years; however, the risk of its invasive and distribution expansion should also be taken into consideration.
The analysis of the lake ice phenology in the high altitude and alpine area is of great significance for traffic capacity assessment on the lake ice in the cold season, disaster prevention and reduction of moraine lake burst and prediction on the flood disaster of the lower reaches in the warm season. On the basis of the OLI data from 2013 to 2017, four typical lakes(areas)in the Pangong Lake area were chosen for the analysis of the lake ice phenology in winter. The results show that the starting freeze time, the time of maximum ice amount, starting thaw time and totally thaw time of Zone1 and Zone2 in the Pangong Lake were almost simultaneous. Although Spanggur Lake and Moriri Lake both have higher altitude than Pangong Lake, and they shared the similar freeze processes. The starting thaw time of Spanggur Lake was later than Pangong Lake, while the totally thaw time was almost the same. The starting thaw time of Moriri Lake was about half to one month later than that of Pangong Lake, and the totally thaw time was one month later than other three lakes.
The Dimunalike iron ore belt is an important iron ore exploration area in western China, where the Dimunalike iron ore deposit, the Yuling iron ore deposit and the Hesu iron ore deposit were discovered. Its geological conditions are superior and the potential for prospecting is huge. Nevertheless, due to its special geological structure and harsh natural geographical conditions, there is a great difficulty in the mineral exploration. In this study, the authors made full use of WorldView-2 high spatial resolution remote sensing data and aeromagnetic data and conducted the decomposition of the Changshagou structure-ophiolite belt to analyze remote sensing and aeromagnetic characteristics of typical deposit of the Dimunalike iron deposit. The authors established a comprehensive prospecting model of sedimentary metamorphic iron ore deposit integrated with geology, aeromagnetic and remote sensing. By using this model, the magnetite ore-bearing target can be located precisely. The results show that the combination of remote sensing and aeromagnetic methods can accurately locate the information such as ore-bearing lithology, ore-controlling structure and mineralized zone. This will improve the accuracy and efficiency of geological exploration.
Image recognition based on low-altitude remote sensing imageries provides a new technological opportunity for forest survey and monitoring. In this study, the authors took the permanent gully in Benggang District, Anxi County, Fujian Province, as an instance and constructed the FC-DenseNet to identify tree species based on the low-altitude aerial optical image of UAV. First, the dense module in the FC-DenseNet model can extract the features of spectral images and enhance the information of the deep network, and the transition down block has an impact on reducing the image dimensions and highlighting the texture and spectral features; then, the transition up block can resize the scale of the predicted image to that of the original image, combined with information fusion of the shallow Dense module; finally, the Softmax classifier is used to achieve pixel-level classification so as to complete the tree species recognition. The results are as follows: ①The FC-DenseNet model based on the low-altitude aerial images not only could identify the difference between vegetation and non-vegetation but also could detect the their spatial distribution. The accuracy of the FC-DenseNet-103 model for vegetation and non-vegetation pixels is 92.1%, and the 103 layers’ network layer is the best network layer. ②Tree species are subdivided into 13 categories, and the accuracy of FC-DenseNet-103 model for dominant species reaches 79%.Some conclusions have been reached: The FC-DenseNet model based on low-altitude aerial optical images has a high tree classification accuracy. With the low cost of low-altitude aerial optical imagery, low data acquisition costs and short time cycles, forest resource surveys and forest species detection can be facilitated. The results obtained by the authors provide a new method in the field of tree recognition using deep learning.
Non-photosynthetic vegetation (NPV) is an important component of grassland ecosystem, which affects the flow and cycle of carbon, water and energy in the ecosystem. It is of great significance to quantitatively grasp the fractional cover of non-photosynthetic vegetation (fNPV) information for the scientific and effective utilization of grassland resources and the protection of the ecological environment. Taking the typical steppe of Xilingol in Inner Mongolia as the research area and using the regression analysis method, the authors used a variety of non-photosynthetic vegetation indices (NPVIs) based on MODIS (MCD43A4) data and field measured fNPV data to invert the fNPV model and evaluated the estimation effect of the model. The results show that the NPVIs based on MODIS data have a good correlation with fNPV. The correlations are as follows: DFI, SWIR32, NDTI, STI, NDI7, NDI5 and NDSVI. The DFI index inversion fNPV model has higher estimation accuracy. It can be applied to the rapid monitoring of large scale fNPV in typical steppe.
Muddy coastal areas have a unique and complex water environment. It is of great scientific significance to deeply analyze the water extraction efficiency of water index in this area. The authors took the Yellow River Delta coast as the study area and used the MODIS and Landsat remote sensing data of 2008, 2009 and 2015. The water extraction performance of 6 water index (NDWI, MNDWI, AWEInsh, AWEIsh, TCW, WI2015) were analyzed from spectral characteristics of land cover types. The best threshold of each water index was obtained through the ROC curve. The accuracy and extraction errors of water indexes in muddy coastal area were studied, and the influence of different land cover factors on water extraction was analyzed. The results show that the AWEInsh have the best performance in extraction of water, with an overall accuracy of 97.29%, mapping accuracy of 96.84%, and user accuracy of 97.69%. The accuracy of seawater extraction by different water indexes is higher than 90%. The extraction accuracy of land water is at general level and the map precision is less than 80%. The capability of NDWI for identifying tidal flat water is poor, and the accuracy of mapping is lower than that of other water indexes. The different water indexes have high omission error of land water, and the omission errors of seawater and tidal flat water are low. The MNDWI has the highest omission error of seawater. The influence of the tidal flat soil on the water extraction is the greatest, followed by the cultivated soil. The sparse vegetation, luxuriant vegetation, and built-up area have the least impact. This study provides a reference for the further development of water extraction methods suitable for muddy coastal areas.