基于RGB颜色空间的矿井运动目标检测及跟踪方法

摘要

On the basis of studying the gray-level plane frame-difference method,a method of detecting video moving objects was put forward based on the color change frame-difference meth-od in RGB color space.The algorithm calculates the frame difference of two adjacent color video frames on RGB three planes separately,and determines whether the changes of the pixel values of RGB three planes for each pixel spot is are in the same direction and calculate by superimposing them.As a consequence,the image of moving objects in frame-difference image is clearer and the noise is more prominent,which makes the foreground image and the background image more easi-ly separated,and subsequent morphological processing such as binarization,filtering,filling and expansion corrosion easier.By using above algorithm and Camshift algorithm,the aim of detec-ting and tracking the moving objects in coal mine was achieved on Labview.The field test resultsshow that the combined algorithm can complete the real-time and accurate detection and tracking of moving objects in coal mine,can realize the linkage with existing production safety monitoring system,which can further ensure the safety of mine production.%在研究灰度平面帧差法的基础上,提出了一种基于RGB颜色空间色变帧差法的运动目标检测方法,算法对彩色视频帧在RGB 3个平面上分别进行相邻两帧帧差计算,比较每个像素点RGB 3个平面像素值的变化是否同向并分别进行叠加计算,所得帧差图像中的运动目标图像更清晰、噪点更突出,使前景图像与背景图像更容易分离,后续的二值化、滤波、填充及膨胀腐蚀等形态学处理更加容易.在Labview软件平台上,采用上述算法与Camshift算法联合实现了对矿井运动目标的检测和跟踪.现场测试结果表明:该联合算法可以完成对矿井运动目标的实时、准确检测与跟踪,还可与矿井现有安全生产监测监控系统实现联动,可最大限度地确保矿井生产安全.