Quantitative Typical Land Cover Remote Sensing and its Application in Earthquake Evaluation

Land cover classification and change detection plays an important role and significance in geographical conditions and environmental monitoring, which is a hot and difficult remote sensing topic. The difficulties in land cover remote sensing are how to realize the detecting by a quantitative and automatic way. At present, there is no general classification algorithm which is proper for different multi-remote sensing data's quantitative processing. In this paper, a multi-source remote sensing image classification method based on spectral reflectance characteristics is studied. Using the geometric and radiometric calibration parameters, sensor spectral response function and satellite-earth-sun's orbit parameters, multi remote sensing data's ground reflectance and land cover monitoring classification could be realized automatically. In earthquake evaluation, two GF2 multispectral remote sensing images are processed. The typical land cover elements, including vegetation, water and bare land, are detected and classified. The evaluation of land cover change proportion is calculated quantitatively and automatically.