红外图像的轮廓提取在军事等领域具有广泛的应用.从视觉注意机制的角度入手,提出了一种基于视觉显著图的红外图像轮廓提取方法,利用红外图像的灰度局部标准差特征计算其显著图,再进行聚类和滤波消除背景信息对目标轮廓的干扰,得到了红外图像中显著目标的轮廓信息.与其他方法相比,提取的显著轮廓更准确,更符合人眼视觉注意的结果.%Infrared image contour extraction is widely used in the military field.The external contour and internal detail contour exist in current target contour extraction research,which does not fit the requirements of engineering practical application.So a method of infrared image contour extraction based on the visual saliency is proposed in the perspective of visual attention mechanism.The local standard deviation of the image is used to compute the characteristic saliency map and it is clustered and filtered to eliminate the interference of the background information,and then the accurate contour can be obtain.Compared with other methods,this contour obtained is more accurate,and fits the result of human visual attention.
展开▼