摘要
总结了雷达遥感图像分类技术的发展过程 ,指出新的分类技术正朝着采用新特征 (如雷达极化信息与干涉信息、多参数极化干涉信息、多时相信息、DEM与地理信息等 ) ,应用新理论 (如小波理论、分形理论、模糊理论 ) ,设计新算法 (如改进的最大似然法、上下文分类法、改进的神经网络分类算法等 )
With more and more information acquired by radar remote sensing the classification technology for radar imagery is heading towards high precision, high correctness and rapidness due to the application of new algorithms, theories, ancillary information and characteristics. In the paper, the development of new classification technology for radar imagery is overviewed. To improve the precision and stability, the authors consider that new characteristics (polarimetric and interferometric information, multi_temporal and geographic information etc.), new theories (like wavelet, fractal and fuzzy theory etc.), and new designed algorithms (such as improved max_likelihood, context Classifier and neural network classifier etc.) should be applied to the classification process of radar imagery.
出处
《国土资源遥感》
CSCD
2001年第3期1-7,共7页
Remote Sensing for Land & Resources
基金
国家自然科学基金项目 (4 99890 0 1)
