改进的OTSU法在遥感水体信息提取中的应用

The application of improved OTSU algorithm in water romote sensing information extraction

在遥感信息解译中,利用指数法提取水体信息一直存在阈值选取困难的问题。为了进一步解决该问题,将改进的OTSU法用于阈值选取。综合考虑目标像元和背景像元类间方差、类内方差两个指标,并结合粒子群算法,对研究区水体信息进行提取。实验结果表明,提出的算法能更快地确定阈值,有效地提取水体信息。

In the interpretation of remote sensing information,the decision of threshold value has always been difficult when the method of index water information extraction is used. In order to further solve the problem,the improved OTSU method for threshold selection is used. Considering the maximum variance between target and background pixels,and the variance whin categories,combined with particle swarm optimization（PSO） algorithm,water information was extracted in the studied area. The experimental results show that the proposed algorithm can quickly determine the threshold value,effectively extract water information.