【摘要】 研究安防监控中人员遮挡区域的有效图像分割,对后期特征识别的准确性有重要的实际意义。在安防视频监控过程中,行走的人员对监控对象进行遮挡,但是受到距离、走向、人员随机波动等因素的干扰,上述遮挡过程存在很强的动态性,遮挡图像之间的前后帧无法建立遮挡区域关联性。传统的图像分割方法在建立分割规则与关联规则时,两者之间无法建立准确的联系,无法对人员波动造成的遮挡不定区域进行准确分割。提出基于人工鱼群算法的安防监控中人员遮挡区域的有效图像分割方法。详细阐述人工鱼群觅食过程,建立人工鱼群模型,将人工鱼群理论应用到安防视频监控中人员遮挡区域分割过程中,确定人工鱼群数量,针对人工鱼群进行初始定位,设置适应度函数,根据分割阈值最优值,实现安防视频监控中人员遮挡区域的有效分割。实验结果表明,利用人工鱼群算法进行安防监控中人员遮挡区域的有效图像分割,能够快速收敛,获取理想的分割结果,提高分割的准确性。更多还原

【Abstract】 An effective image segmentation method is put forward based on artificial fish algorithm in the security monitoring of people keep out area. The artificial fish feeding process is analysed,the artificial fish model is established,and the artificial fish theory is applied to image segmentation in the security video monitoring process in personnel keep out region. The number of artificial fish is determined,initial positioning of artificial fish is obtained,fitness function is set up,and effective image segmentation is achieved according to the optimal segmentation threshold value. Experimental results show that the proposed method can obtain fast convergence and improve the segmentation accuracy.更多还原