形态学增强的连接相对熵在自动目标检测中的应用

提出了一种基于多尺度形态学和连接相对熵的红外图像自动目标检测方法;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具有一定的鲁棒性和自适应性,能够实现目标的自动、非参数化的有效检测,且效果良好,便于下一步的目标识别和跟踪.

Mathematical Morphology Enhancement of Joint Relative Entropy for Automatic Target Detection

For the automatic target detection in infrared image sequences, a new infrared target detection algorithm based on morphological filtering and joint relative entropy is proposed. This algorithm detects target in three steps. Firstly, it removes background and clutter with multiscale morphological operations; then applies joint relative entropy based thresholding method; finaly the thresholded binary image is filtered with morphological operation again. Unlike existing methods, this method is fully automatic, parameter free, and independent of local statistics. The experimental results verify the effectiveness and robustness of this algorithm which can facilitate the target recognition and tracking in the next step.