摘要
针对现有的阈值选取方法应用于目标与背景面积相差悬殊的红外图像时常导致严重的误分割现象,本文提出了一种基于对称交叉熵及背景与目标面积差的红外目标图像阈值选取方法。对称交叉熵能确保分割后类的内聚性好,而背景与目标面积差可抑制均等分割的趋势,将两者综合构成了更为合理的阈值选取准则函数。首先导出了一维阈值选取公式;然后给出了二维直方图斜分阈值及二维直方图斜分的简化阈值选取方法,抗噪性能明显改善;最后与二维斜分的最大熵阈值、Otsu阈值及非对称交叉熵阈值选取方法进行了比较,实验结果表明,本文方法在分割效果上具有明显的优势。
When existing methods are applied to segment infrared images with significant area difference between background and target,they often lead to severe segmentation errors.Thus,a threshold segmentation method for infrared target images is proposed based on the symmetric cross-entropy and the area difference between background and target.The symmetric cross-entropy can make the cohesion performance better,and the area difference between background and target is used to inhibit the tendency of an equal division.Therefore,a more reasonable threshold selection rule is formed comprehensively.First of all,the 1-D threshold selection formula is derived.Then it is extended to the 2-D histogram,and the threshold method and the simplified thresholding method based on oblique segmentation of 2-D histogram are proposed,so that the anti-noise performance is improved obviously.Finally,the suggested methods are compared with the threshold methods of the maximum entropy,the Otsu and the symmetric cross-entropy based on oblique segmentation of 2-D histogram.The experimental results show that the suggested methods are more effective in segmenting infrared images.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2010年第12期1871-1876,共6页
Journal of Optoelectronics·Laser
基金
国家自然科学基金资助项目(60872065)
关键词
红外图像分割
对称交叉熵
背景与目标面积差
二维直方图斜分
infrared image segmentation
symmetric cross-entropy
area difference between background and target
oblique segmentation of 2-D histogram
