红外图像活动轮廓分割算法

Active contours based algorithm for infrared image segmentation

红外图像通常存在边缘模糊、对比度低、背景复杂等特点,传统的活动轮廓模型难以达到理想的分割效果。针对这种情况,提出了一种能够用于红外图像分割的边缘活动轮廓模型。该模型采用了结合图像局部熵信息定义的边缘停止函数,同时提出了一种针对红外图像的自动选取初始轮廓的方法,进一步提高了算法的效率。实验结果表明,采用该模型能够有效分割红外图像。

Infrared images always have blurry edges,low contrast and complicated background,therefore the segmentation of infrared images by traditional active contours is very difficult.This paper proposed an edge-based active contour model for the segmentation of infrared images.This model used a new edge stopping function which used the local entropy of image.To further improve the efficiency,this model used an automatic initialization scheme which was based on the feature of infrared image.The experiments show that the proposed model can efficiently segment infrared images.