面向水体的HY-1B/COCTS图像条带噪声去除

Removal of stripe noise in HY-1B / COCTS image based on waters

HY-1B/COCTS是我国自主研发的第2颗海洋卫星HY-1B星搭载的水色水温扫描仪(Chinese ocean color and temperature scanner,COCTS)。由于COCTS的不同探测器对地物光谱响应有差异,其图像的行间存在明暗的条带噪声。针对COCTS图像在水色遥感中的应用特点,在分析可见光-近红外波段地物光谱特征的基础上,使用其第4和第8波段数据构建的归一化差分水体指数(normalized difference water index,NDWI)提取图像的水体信息;采用矩匹配法对COCTS图像水体区域的条带噪声进行去除,并以均值、标准差、信噪比、偏度、峰度、信息熵和平均梯度7个指标对条带噪声去除后的COCTS图像进行质量评价。结果表明,该方法能有效消除图像的条带噪声,使图像质量得到改善。

COCTS is a Chinese ocean color and temperature scanner which is carried on the 2 nd oceanographic satellite （HY-1B） launched by China. HY-1B/COCTS images have the alternate dark and bright lines due to the difference of the detectors within COCTS responding to spectral signal reflected by ground objects. In this paper,the authors analyzed the spectrum of ground objects in visible and near infrared bands of COCTS, retrieved the water information by the normalized difference water index （ NDWI） constructed with the 4th and 8th band of COCTS, and then removed the stripe noise for waters in COCTS image by the moment matching. Finally, 7 indexes, namely, mean, standard deviation, signal - to - noise, skewness, kurtosis, information entropy and average gradient, were selected to evaluate the de - striped image quality. The results show that the moment matching based on waters performs well in removing the stripe noise in COCTS image, the contrast of digital number is decreased, and the image quality is improved.