【摘要】 应用在高架配电线路上的绝缘漆喷涂机器人在自然因素影响下，无法保持匀速作业，导致其位置跟踪难度较大。于是提出一种新的机器人作业位置跟踪方法。对采集的目标图像完成高斯滤波处理，最大程度强化目标图像细节信息。检测目标图像显著域，获取可能存在目标的感兴趣区域，同时在原图内提取对应特征，完成作业位置识别。引入Bouguet算法，校正目标图像。利用双目立体匹配的视差获取机器人作业位置，最终实现配电线路绝缘漆喷涂机器人作业位置的跟踪。仿真结果表明，研究方法能够准确跟踪机器人作业位置，目标位置的识别误差低于0.5cm,平均耗时为43.6s,有效提高了绝缘漆喷涂机器人作业效率。更多还原

【Abstract】 The insulation paint spraying robot applied on elevated distribution lines cannot maintain a uniform speed of operation under the influence of natural factors, making it difficult to track its position. Therefore, a new robot job position tracking method is proposed. Gaussian filtering on the collected target image was completed to maximize the enhancement of target image details. And then, the salient region of the image was detected to find the region of interest where the target may exist. Meanwhile, the features in original image were extracted correspondingly to complete the location recognition. Moreover, Bouguet algorithm was introduced to correct the target image. Furthermore, the parallax with binocular stereo matching was employed to obtain the working position of robot. Finally, the tracking for the working position of insulating paint spraying robot was realized. Simulation results show that the proposed method can accurately track the working position of robot, the recognition error is less than 0.5cm; In addition, the average time consumption is 43.6s. This method improves the efficiency of the robot.更多还原