最佳小波包基的矿区植被及红边位置提取

Wavelet Packet Based Extraction of Vegetation and Red Edge Position in Mine Area

小波包变换能同时对植被光谱信息的低频和高频分量进行分解,并能克服小波变换时间分辨率高而频率分辨率低的缺陷从而具有能够探测植被细微变化的优势。实验利用Hyperion高光谱影像对云南省普朗铜矿区植被像元的光谱进行最佳小波包基参量获取与植被信息识别,并在此基础上提出一种提取重金属污染下植被红边位置的最佳小波包基系数应用模型。研究结果表明基于最佳小波包基参量的植被信息识别及基于最佳小波包基系数的重金属污染探测具有可行性与一定优越性。

Wavelet packet transform can decompose simultaneously the low frequency and high frequency compo-nents of vegetation spectral information, and owns the advantages of detecting the small change of vegetation be-cause it can overcome the defects of high time resolution and low frequency resolution of wavelet transform. Experi-ment studied the extraction of the best wavelet packet bases and vegetation information about vegetation pixel spec-tral based on Hyperion hyperspectral images of Pulang in yunnan copper mining area , then a diagnostic model is built based on the best-base wavelet packet coefficients on detecting the spectral ＆quot;red edge＆quot; signal singularity of vegetation polluted by heavy metal. The study shows that it is feasible and superior to identify the vegetation infor-mation based on wavelet packet best-base parameters and detect its heavy metal pollution based on the wavelet packet coefficients of best bases.