基于掘进机位姿的特征检测技术研究

Research on feature detection technology based on heading attitude

在掘进机定位的过程中,为实现掘进机位姿自动化无接触检测的功能,针对在运动目标检测过程中采用传统背景差分法或帧间差分法存在的准确率不高、特征缺失的问题,提出一种基于改进的视觉特征检测算法。该算法利用机器视觉、图像处理等,通过改进的三帧差分法结合边缘检测、均值法背景建模对运动目标进行检测,最后通过形态学处理得到运动目标。仿真结果表明,采用该算法可弥补传统背景帧差法和帧间差分法的缺点,有效提高运动目标检测的实时性、准确性和检测效率。

in the process of locating the roadheader, in order to realize the automatic non-contact detection function of the tunneling attitude, aiming at the problem that the traditional background difference method or the frame difference method exists in the process of moving object detection, a new algorithm based on improved visual feature detection is proposed. The algorithm uses machine vision, image processing and so on, through the improved three- frame difference method combined with edge detection and mean-value background modeling to detect moving objects, and finally obtains moving targets by morphological processing. Simulation results show that the algorithm can make up the disadvantage of traditional background frame difference method and frame difference method, and improve the real -time, accuracy and detection efficiency of moving target detection effectively.