Remote Sensing Image Fusion Model For Creating Soil Erosion Image

Soil erosion is the main reason for environment degradation globally. Remote sensing is the only method that can monitor land exhaustively and directly by providing good coverage of the earth at a great range of scales. Meanwhile, in this paper, an overview of the history and theory for multiresolution image fusion approaches is presented. First, the principle and development of the multiresolution image fusion approaches is given. Then, remote sensing image fusion techniques, including the theory of fusion, main algorithms, applications of remote sensing image fusion, and criteria for evaluating fused images are included. Finally, problems of remote sensing image fusion are also discussed.In this paper, the main study focus on basic concepts, soil erosion models, procedure of data processing, image fusion models, image fusion architecture etc. involved in producing soil erosion image of Hubei province at large scale and soil erosion image of Wuhan region at small scale. The following aspects are mainly discussed:(1) Basis on reviewing the history and present situation of the study on soil degradation process in the world and China, this paper summarizes the recent development in the forming mechanism, the models of soil degradation, and the assessment indexes for soil degradation. Finally, the main problems and futures of the research work in the field are also put forward.(2) This paper tries to gather and review the currently remote sensing approaches in soil erosion monitoring, whose dominances and limitations are also discussed.(3) This paper proposes a quick method of producing soil erosion based on moderate resolution imaging spectroradiometer (MODIS) data. Firstly, the land use image, and vegetation coverage image are obtained using interpreted TM image and MODIS vegetation index model, respectively. Secondly, slope image is created from Digital Elevation Model (DEM) . Thirdly, based on the soil erosion intensity classification, soil erosion intensity image is obtained by overlaying the images of land use, vegetation coverage and slope.(4) The efficient integration of remote sensing images with multiple spatial, temporal and spectrum resolution is a focus of image processing study. A full review of the origination and development of image fusion is given here. The concepts, and basic theories in this field are expatiated. Current image fusion techniques are classified into three types. The strong points and shortcomings of each type are analyzed.(5) For local area, we obtain high spatial resolution MODIS image by fusing TM and MODIS images. Then, vegetation coverage image is obtained using the fused MODIS based on normalized difference vegetation index model. Finally, soil erosion intensity image is obtained by overlaying the images of land use, vegetation coverage and slope of the same region.In this paper, there are three innovative tasks:(1) The land use image is obtained by interpreting manually the TM/ETM images, and the MODIS images are used to monitor the vegetation change. Therefore, the TM/ETM images and the MODIS images are fused at macroscale.(2) The spatial resolution of the MODIS images is improved using image fusion methods by means of the TM image, and the soil erosion image at small scale is obtained based on the fused MODIS images. Therefore, the TM/ETM images and the MODIS images are fused at microscale.(3) The available MODIS data, land use data, and elevation data are used to monitor water and soil preservation dynamically. Therefore, the TM/ETM images and the MODIS images are fused at time scale.