尘雾环境下工作面护帮板视频目标定位

Video Target Positioning of Working Face Guard Plate in Hazy Environment

工作面护帮板的精确定位是液压支架动作决策的基础和关键。在采煤机截煤过程中出现的大量煤尘以及除尘设备喷出的水雾严重干扰了工作面视频中护帮板的识别准确性,从而严重影响护帮板定位的精度。针对此问题,面向工作面尘雾环境提出一种护帮板视频目标精确定位算法。采用常值暗通道先验知识,实现护帮板视频目标的清晰化,从而克服煤尘和水雾对护帮板视频目标定位的影响。同时采用摄像头转角动态基准校正机制,通过补偿校正摄像头装配误差和转角漂移,对视频中的护帮板目标实现精确定位。实验结果表明,该算法在尘雾环境下对不同摄像头转角下的护帮板视频目标实现了精确的定位,取得了良好的效果。

The precise positioning of the working face guard plate is the basis and key to the decisionmaking of hydraulic support. The large amount of coal dust and the water mist sprayed by the dust removal equipment during the coal cutting process of the shearer seriously interfer with the recognition accuracy of the guard plate in the video of the working face, thereby seriously affect the positioning accuracy of the guard plate. Aiming at this problem, proposed an algorithm for precise positioning of the video target of the guard plate for the hazy environment of the working face. The prior knowledge of the constant dark channel was used to realize the clarity of the video target of the guard plate,so as to overcome the influence of coal dust and water mist on the video target positioning of the guard plate. At the same time, the camera angle dynamic reference correction mechanism was adopted, accurate positioning of the guard plate target in the video was achieved by compensating and correcting the camera assembly error and rotation angle drift. The experimental results show that this algorithm can accurately locate the guard plate video target under different camera rotation angles in hazy environment and has achieved good results.