基于均值漂移的红外舰船图像分割算法

摘要

提出了一种有效的舰船红外目标图像分割算法，利用均值漂移分割算法的二维灰度和灰度邻域信息，滤除了海面的强杂波干扰，同时又不会损失舰船目标信息。利用不同类别的灰度权重，将红外舰船图像分割成为天空、海水和舰船3类，从而将舰船从图像中有效地分割。由于采用区域节点和灰度直方图来表征图像，与原始图像像素节点表征图像相比，区域节点的个数远远小于原始图像像素节点，从而提高了算法计算效率。计算结果也表明，该算法能够在海面强杂波的干扰下，有效地提取红外舰船目标。%A novel algorithm was presented to achieve an improved infrared obj ect segmentation performance,the pro-posed algorithm used discontinuity to preserve smoothing algorithm based on mean shift procedure to filter the powerful noise without the loss of the ship obj ect information.The regions produced can be represented by a planar weighted region adjacency graphs that incorporated topological information of the image structure into three parts,that is sky background, sea background and ship obj ect.Due to the less modes and histogram images produced by the regions of filtered image com-pared with that of the original image,the region merging based on mean shift algorithm has much less computational com-plexity compared with the traditional methods based on pixel gray value.A large number of examples were presented to show the superior performance of the proposed ship infrared obj ect segmentation algorithm.