基于改进Camshift算法的钻杆计数方法

针对采用传统的Camshift算法跟踪钻杆机器时相似颜色物体间易产生干扰的问题,提出了一种基于改进Camshift算法的钻杆计数方法:使用改进的Camshift算法实时捕捉钻杆目标,并且通过设置锚点以及判断时间间隔来进行钻杆计数。改进的Camshift算法采用模版匹配的方法定位跟踪目标,智能地选择合适的方法提取颜色直方图,并引入粒子滤波器,通过一定机制与纹理特征相结合,得到跟踪目标在下一帧的准确位置。实验结果表明,改进后的算法解决了相似颜色干扰问题,并且在部分遮挡条件下仍能够准确跟踪目标,从而实现对钻杆的准确计数。

Drill counting method based on improved Camshift algorithm

In view of problem that it is easy to produce interference between objects with the similar colors while using traditional Camshift algorithm for drill tracking,a drill counting method based on improved Camshift algorithm was proposed.Improved Camshift algorithm was used to capture drill targets real timely,and then drill counting was realized by setting anchor and determining time interval.The improved Camshift algorithm uses template matching method to locate tracking target,intelligently selects appropriate method to extract color histogram,meanwhile introduces particle filters to get exact location of the tracking target in the next frame by combining texture features through certain mechanisms.Experimental results show that the improved algorithm solves the interference problem of similar colors,and is still able to accurately track the target in conditions of partial cover,consequently achieves accurate drill counting.