一种新的基于背景的红外图像分割方法

针对红外图像中目标周围的背景往往比较均匀的特点,文中提出了一种基于背景区域的目标提取方法.首先利用局部熵信息检测到目标的位置,并在此基础上获得一个包含目标的感兴趣区域,最后通过基于直方图搜索的区域增长法来提取感兴趣区域中的背景部分,从而有效地将背景和目标进行了分割,实现了目标完整提取的目的.由于是在感兴趣区域进行提取而且对直方图进行搜索,克服了区域增长法计算量大的缺陷.试验结果证明,该方法能有效且完整地提取目标.

A Novel Background-based Infrared Image Segmentation Method

In terms of the feature of infrared images that the background around the object is always even,a novel segmentation method based on background information has been proposed in this paper.Firstly the position of the object was detected using local entropy information.And then a region of interest which contained the object was produced.At last the method using region growing based on histogram searching segmented the target efficiently by extracting the background of the region of interest(ROI) instead of the target itself.In order to reduce the computation cost,the method was used in the histogram of ROI.Experimental result shows that the method can effectively and integrally extract the target in the infrared image.