Remote Sensing Image Based On GF-1 Mining Area Of Information Extraction And Modeling

The study is based on GF-1 remote sensing data to establish a more mature and reasonable extraction methods and processes for mine information, which try to be more efficiently and more quickly to extract of the mine information for a thorough analysis of the key technical problems in the process of information extraction. By comparing the fusion methods and effects to analyze the GF-1 data, the results show that the remote sensing image is more saturation and clear and the surface features' boundary is more specific getting by GS fusion method; By calculating the correlation between bands and the statistical information inside single band, combined with band index OIF, to further determine the best imaging band of remote sensing image is band4\3\2; Based on IDL platform to achieve the Mean Shift algorithm which enhance GF-1 data excellent, the image data has a better visual interpretation effect after the enhancement processing which is useful for the classification of the follow-up work, the purpose of previous studies is provide image enhancement technical support for GF-1 or the remote sensing image of a similar resolution. Through a large number of experimental to determine the most suitable segmentation scales and parameters about the tailings in the study area which is the target object. By analyze the segmentation results to choose the best texture band to participate in multi-scale segmentation process, experiments show that add the two texture features which are Contrast and Correlation at the same time, the segmentation result of the mining area will be best; By statistical analysis of the mining area and the surroundings' spectrum characteristic,which are in the research area to established the tailings' extract model IOT. Along with texture and geometric features using object-oriented method to extract the mining area and the surroundings' information, the precision of extraction is 83.377%. Compared the maximum likelihood classification based on pixels which extraction precision is 68.024%, we can concluded that the object-oriented classification method of GF-1 tailings information extraction has the absolute advantage.