针对多源目标图像区域分割问题,采用了多分辨率阈值选取方法,依据图像小波分解后的最小距离求得最优多分辨率阈值,对多源图像进行分割,标定重点目标区域。仿真研究表明,利用小波变换在较小尺度下对区域边缘点定位比较准确,可得到图像区域边缘增强图,并与全局阈值分割结果比较发现,此方法提取出的有效区域,为图像信息后续研究提供了一种可行的技术支撑。

In view of multi-source target image region segmentation, a multi-resolution threshold selection method was used to get the optimal resolution threshold based on minimum distance of the image wavelet decomposition, multi-source image was segmented, and key target area was calibrated. Simulation studies have shown that the use of wavelet transform at smaller scales could locate the edge of the region accurately. The enhanced images of the edge image area can be obtained, and global threshold segmentation results could be found. By this method, the effective area was extracted, and a viable technical support was provided for follow-up of the image information.