基于视觉的无人机着舰中红外合作目标的分割方法研究

针对基于视觉的无人机着舰中红外合作目标分割的需求背景,提出一种组合分割方法。相关研究发现随着待分割目标所占图像中比例变化,Otsu算法在分割时效果亦随之改变。当目标比例较小时可能会造成不同程度的误分割,而最大熵分割算法在实际处理中又含有较多噪点。本文首先根据图像的能量信息找出包含合作目标的感兴趣区域,同时在原图运用双Otsu算法进行分割,最后对两者的结果进行逻辑＂与＂运算得出最终分割结果。大量实验结果表明,本文的组合分割方法与Otsu算法和最大熵算法相比分割效果有明显改善。

The Study of IR Segmentation of Cooperative Target in Vision-Based UAV Landing on Ship

A new combinatorial method is proposed in this paper for the background of the IR segmentation of cooperative target of vision-based UAV landing on ship.Relative studies have found that with the change of percentage of cooperative object in image,there will be different degrees of error in the segmentation using Otsu algorithm for IR cooperative target when the percent is small.Meanwhile,the result of the maximum entropy segmentation algorithm used to deal with in practice is also proved to contain some noi...