基于遗传算法的视频交通量检测图像分割方法

Image Segmentation Method for Genetic Algorithm Based Traffic Volume Video Detection

将遗传算法应用于交通量视频检测图像分割中 ,在Otsu法的基础上 ,采用多次迭代以获得最优分割阈值 ,能够在搜索空间内找到全局的最优分割阈值 ,从而能更有效地把背景和目标分割开来 ;由于在寻求最优解的过程中采用并行计算 ,其运算速度优于Otsu法 ,满足了交通量检测的实时性要求。

This paper applies the Genetic Algorithm to the image segmentation for the video traffic volume detection. Based on the Otsu method, the optimum segmentation threshold can be obtained by repeated iterations and the global optimum segmentation threshold can be found in the searching space, so the object and the background are segmented more efficienty. The method proposed by this paper has a faster operation speed than the Otsu method and can meet the real time requirement for traffic detection because of using parallel operation in the progress for searching optimum solutions.