遥感图像分割中的小波域改进信息割算法

Remote Sensing Image Segmentation Based on Modified Information Cut in Wavelet Domain

针对传统信息割(IC)算法对存在灰度变化的图像易引起误分割的问题,提出了一种小波域改进信息割(W-MIC)算法。使用一个新的结合像素点灰度关联和空间关联的Parzen窗对IC进行改进,以降低灰度变化对分割的影响;将MIC引入到小波域,利用小波的平滑作用降低参数选取复杂度,改善图像分割效果。遥感图像分割实验结果表明,W-MIC可有效降低参数选择效应,不仅避免了灰度变化引起的误分割,而且还较好地保持了图像边缘。

To solve the mis-clusters caused by the traditional information cut （IC） algorithm when it is applied to segment images with gray changes, a modified information cut in wave- let domain （W-MIC） algorithm is proposed. First, it modifies IC using a new Parzen windo- wing which combines the gray relevance and space relevance between image pixels in order to reduce the effect of gray changes to image segmentation. Further, considering the difficulties of selecting the optimal parameter in the above modified information cut （MIC）, MIC is in- troduced to wavelet domain to reduce the complexity of parameter selection via the smoothing role of wavelet. Remote sensing image segmentation results demonstrate that it effectively decreases the influence of parameters selection, and it can not only avoid the mis-clusters caused by gray changes, but also keep image edges.