基于累积图像特征点匹配的智能皮带测速方法

在煤炭工业生产运输过程中可能出现皮带打滑问题影响设备运行,现有的基于传感器的测速方法可能因为贴合问题出现误测.提出一种基于视觉的智能皮带测速方法,针对皮带低纹理特征点不足的问题,通过累积多帧特征点对估计运动模型.同时,实际场景可能出现的视频抖动问题可以通过背景特征点估计出偏差,并进行补偿.另外,图像坐标系下的检测结果可以根据标定得到的摄像机内外参数转换到世界坐标系下皮带运动速度.实验结果验证了方法的有效性和准确性.该方法实现了煤炭运输皮带的智能测速,可用于皮带打滑故障的及时报告.

A Method of Intelligent Speed Measurement of Coal Conveyer Belt Based on Accumulated Feature Point Matching

During the production and transportation in coal industry, belt slip may happen, which influences equipment operation. Existing speed detection methods are based on sensors and may cause error due to joint problem. Therefore, in this paper, a vision-based intelligent speed detection approach is proposed. To handle the problem of insufficient feature points in low-textured belt, motion model is estimated through accumulating point correspondences from multiple frames. Meanwhile, the shake problem of video in practice can be eliminated relying on background feature points. In addition, the detection result in image coordinate system can be transformed to the actual speed of belt in world coordinate system through calibrating the internal and external parameters of the camera. Experiment demonstrates the efficiency and accuracy of the proposed method. It implements the intelligent speed measurement of conveyer belt and applies to report the slip problem timely.