最大熵阈值的红外图像人体目标分割方法

为了有效地分割红外图像中的人体,提出了一种基于最大熵阈值的红外图像人体目标分割方法.对红外图像进行滤波处理消除噪声干扰,分别计算图像的目标熵和背景熵,最大化目标与背景熵的和,在目标和背景的分布中获得最大信息,以此为准则选择分割阈值.利用形态学方法进行后处理进一步消除噪声干扰.实验结果表明:与经典的阈值分割方法相比,文中方法效果更好,且运算速度快.

Method for Pedestrian Segmentation in Infrared Images Based on Maximum Entropy Threshold

The method for the pedestrian segmentation based on the maximum entropy threshold is presented in order to effectively segment the pedestrian in infrared images .The infrared image is filtered first to eliminate the noise effect .Then the object entropy and the background entropy are calculated respectively ,and the sum of entropies of object and background is maximized to obtain the maximum information from the object and background distributions in the image .According to the information the threshold is egmented .The final treatment is made by the morphological method to further eliminate the noise effect .Experimental results show that the proposed method has advantages over the existing classic thresholding methods of good segmentation quality and high calculation speed .