基于OpenCV的动态检测

Dynamic Detection Based on OpenCV

实现了一个基于OpenCV与机器视觉技术相结合的动态检测方法。在Visual Studio 2010结合OpenCV2.4.3的开发环境下,进行了摄像机抓取指定的视频帧、对抓取的帧图像进行灰度化处理、异常时将灰度化后的前景与背景差分、采用Otsu法确定动态阈值对差分图像进行阈值分割得到二值图、腐蚀膨胀二值图后计算非零点像素数判断场景中是否有物体移动。通过实验证明上述方法不仅对监控区域的动态物体检测效果很好,而且克服红外探测器自身的诸多问题。既可以沿用原有的监控摄像机,也可根据需要采用针孔和防暴等特殊摄像机以更好地实现安防监控的目的。

Realize a dynamic detection method based on OpenCV Combined with machine vision technology. Under Visual Studio 2010 and OpenCV2.4.3 development environment, grab the appointed frame from video camera. Conduct gray processing to the grabed frame. When abnormal after gray processing, foreground and background difference. Otsu method was used to determine the dynamic threshold, then threshold segmentation to the difference image to get the binary image. Corrode and expand the binary image. Judge there are moving objects in the scene or not after the calculation of non-zero pixels. Experiments show that the above-mentioned is not only very good for the dynamic detection to the monitoring region, but also overcome many problems of infrared detector itself. Not only can continue to use the camera original, according to the need can use pinhole and anti or other special camera in order to achieve the purpose of security monitoring better.