边缘检测和区域生长相结合的图像ROI提取方法

摘要

自然彩色图像感兴趣区域的提取与分析,对图像处理有着重要的意义.针对自然彩色图像,提出一种边缘检测和区域生长相结合的感兴趣区域提取方法.彩色图像经过预处理后,首先采用基于相位信息的方法来检测边缘.得到图像的大致区域分布.然后再结合边缘信息自动获得种子点,实现阈值的自适应确定进行区域生长.生长结果再经过后续处理最终得到感兴趣区域.实验结果表明,该方法能够更加精确、快速地对自然彩色图像进行处理,得到较好的结果.%The extraction technique of a Region of Interest (ROI) from a natural color image is an important step towards highly efficient image communication. An ROI search algorithm based on edge detection and region growing is proposed. After the pretreatment, the phase congruency was applied to detect edge and obtained the major geometric structures in a color image for seeded region growing. The threshold value is adaptively adjusted with the image automatically and initial seeds can also be computed automatically from the obtained color-edge regions. Then ROI can be extracted by region growing and the following treatment. Experimental results show that this combined method can detect an ROI from a color image accurately and quickly.