摘要
本文就在复杂背景下的运动目标检测,对传统的帧间差分法和背景帧差法进行改进,对Canny算子检测边缘检测运动目标和增加计数器来对背景模型实时更新的基本算法进行改进,对他们的缺点进行改进,同时引入了中值法提取的背景作为初始背景,大津法进行目标分割,再结合集合关系。将得到的两种二值化图像进行与运算,最后利用图像形态学滤波得到准确的运动目标。
This paper research the moving object detection In the complex background,the traditional frame,background frame difference method difference method,Canny edge and increasing the counter to updata the background are improved,combining the advantages of them,at the same time,the extracted background of median method as the initial background,Otsu method is used for object segmentation,Combining relations of set.The binary images will be operated by the intersection of them.The accurate moving targe will be obtained by morphological operation.
出处
《科技信息》
2013年第15期167-168,共2页
Science & Technology Information
