基于小波分析和数学形态学相融合的高光谱数据去噪

Hyper spectral data denoising based on wavelet analysis and mathematical morphology

为了更好的消除高光谱数据噪声,加快高光谱数据去噪去噪速度,设计了基于小波分析和小波分析的高光谱数据去噪方法。首先收集原始的高光谱数据,并采用小波分析对其进行变换,通过设置阈值去除高光谱数据中的噪声,然后采用数学形态学对高光谱数据进行处理,实现高光谱数据的进一步去噪。从主观和客观两个方面对高光谱数据去噪方法的性能进行评价,本文方法的高光谱数据去噪效果好,解决了当前方法存在的不足,加快了高光谱数据去噪速度,研究结果可以为高光谱数据去噪的研究提供有价值参考。

Abstract ： in order to eliminate the noise of hyperspectral data and accelerate the denoising speed of hy- perspectral data, a hyperspectral data denoising algorithm based on wavelet analysis and wavelet analysis is designed. The first collection of original hyperspectral data, using wavelet analysis to transform it, by setting the threshold is divided into sub blocks, using wavelet analysis to decompose each block, and then use mathematical morphology for hyperspectral data processing, to achieve high spectral data denoising. To evaluate the performance from the two aspects of subjective and objective denoising algorithm for hyperspectral data, hyperspectral data the algorithm denoising effect, solves the shortcomings of the current algorithms, accelerate the denoising speed of hyperspectral data, the research results can denoising of hyperspectral data provide valuable reference.