论文部分内容阅读
根据每种地物都有自己独特而唯一的“光谱指纹”特性,对土地覆被变化情况做出判断时便可以对照“指纹”配对进行快速检测,各种乱砍滥伐、违法乱占林地、毁林开垦等破坏森林资源的行为将无所遁形。通过分析典型地物的光谱指纹特性,提出一种基于光谱知识和植被指数的高光谱图像自动识别方法,最终提取林地范围内地表发生变化的地块,从而实现对检查时段内林地变化情况的准确定位与快速定性。
According to each feature has its own unique and unique “spectral fingerprint ” characteristics of land cover changes to make judgments can be controlled “fingerprint ” pairing for rapid testing, a variety of deforestation, Illegal occupation of forest land, deforestation and other destruction of forest resources will be found invisible. By analyzing the characteristics of spectral fingerprint of typical objects, an automatic hyperspectral image recognition method based on spectral knowledge and vegetation index is proposed. Finally, the patch of land surface changes within the scope of forest land is finally extracted, Positioning and rapid qualitative.