Abstract: For the segmentation of moving targets in the IR imaging guidance system, a new automatic segmentation method is presented. First, the IR image is filtered by multiscale morphological filters, suppressing the background and eliminating the clusters to enhance the contrast between the object and background. Secondly, the joint relative entropy threshold is applied to get the threshoded image; Finally, the thresholded image is filtered by morphological filter again. The experimental results verify the effectives and robustness of this segmentation method which can facilitate the target recognition and tracking in the next step.

Key words: Image segmentation, Morphological filtering, Joint relative entropy

摘要： 针对红外成像制导过程中运动目标的分割，提出了一种红外目标图像的自动分割方法，它首先对目标红外图像进行多尺度形态滤波，消除背景和杂波的影响同时增强目标的对比度，然后采用连接相对熵准则确定区分目标和背景的最佳门限，最后对得到的二值图再进行形态滤波，得到精确的目标分割图像。实验结果表明该方法效果良好，便于下一步的目标识别和跟踪。

关键词: 图像分割, 形态滤波, 连接相对熵

CLC Number: 

• plwu@eyou.com