红外目标分割方法研究

红外目标分割算法对红外目标检测、跟踪具有非常重要的价值。本文利用背景和目标灰度特征,提出一种实现红外目标有效分割的方法,克服红外目标内部温度不稳定造成的误分割问题。本文方法首先采用基于灰度-显著度最大相关准则的二维直方图分割算法进行图像分割;然后,在分割后二值图上进行基于随机种子点选取的区域增长,提取背景;最后,采用形态学方法优化分割结果。相对传统的红外目标检测算法,这种算法具有更好的抗干扰能力,更强的鲁棒性。不仅可以应用于红外图像的目标分割,而且可以应用于其他类似的目标分割问题。

Research on infrared target segmentation

Infrared target segmentation is very valuable to infrared target detection and tracking. The research proposes an effective infrared target segmentation based on background and target gray character. The method avoids gray instability in the interior of the targets caused by temperature instability. Firstly, two-dimensional MCC based on gray-salience was proposed to segment targets. Then, improved region-growing algorithm based on random reed points was presented to extract background. Finally, morphological operation improved the segmentation performance. Compared with the traditional algorithms of target segmentation, this algorithm has better anti-jamming and robust. The method can be used not only for target segmentation but also for other similar segmentation.