摘要
基于多分辨率分析的图像分割技术是当前图像处理的重要内容,提出了小波变换多分辨率分析方法与改进分水岭分割算法相结合的综合分割方法。此方法利用逐层影射和小波反变换可以得到高分辨率图像,与在原始图像上直接进行传统分水岭分割算法相比较,该方法的实验结果能有效地减少分水岭算法图像过分割现象,经实验证明了该方法的有效性和实用性。
Image segmentation based on multi-resolution analysis is an important component in current image processing. In this paper we propose a comprehensive image segmentation method, which combines the wavelet transforming multi-resolution analysis with the improved watersheds segmentation. The method uses stepwise mapping and wavelet inverse transformation together to get high-resolution images. Comparing with the algorithm which segment directly on primitive image with conventional watershed, the experimental result of this method can effectively reduce over-segmenting phenomenon caused by watershed segmentation algorithm and the method has been demonstrated to have validity and practicability by the experiment.
出处
《计算机应用与软件》
CSCD
2009年第4期62-64,共3页
Computer Applications and Software
基金
河南省教育厅自然科学研究资助项目(200410465201
200510465002)
关键词
图像分割
小波变换
高分辨率图像
分水岭算法
过分割
Image segmentation Wavelet transform High-resolution images Watershed algorithm Over-segmentation
