一种改进的超宽条带噪声消除算法

针对高光谱遥感图像中的超宽条带噪声干扰现象，在深入研究高光谱图像特点和条带噪声产生机理的基础上，提出了一种新的基于最小序列值、小波变换和矩匹配相结合的滤波算法（OWM算法）。该算法主要包括灰度对比度处理、最小序列值处理、小波变换系数归零处理和矩匹配处理等四个步骤。用实际的高光谱图像进行了一系列的验证比较实验，获得了好的实验效果。实验结果表明OWM算法不仅能够有效滤除高光谱图像中的超宽条带噪声，而且还具有较好的普适性。

Improved filtering algorithm on ultra-wide stripe noise

Aiming at the interference phenomenon of ultra-wide stripe noise in hyperspectral remote sensing images, on the basis of study on the characteristics of hyperspectral images and the generation mechanism of the strip noise, a new filtering algorithm based on the combination of the minimum order value, wavelet transform and moment matching was proposed, called OWM (order, wavelet and moment) algorithm. The algorithm mainly consists of four steps: the gray contrast processing, the minimum order value processing, the zeroing processing of the wavelet transform coefficients and the moment matching processing. A series of experiments were carried out with real hyperspectral images, and good results were obtained. These experimental results show that the OWM algorithm not only can effectively filter the stripe noise in hyperspectral images, but also has a good range of applications.