机构地区: 湖南大学电气与信息工程学院
出 处: 《仪器仪表学报》 2011年第11期2564-2571,共8页
摘 要: 障碍物检测识别是实现输电线路除冰机器人自主除冰作业的关键技术之一。针对220 kV输电线路的结构特点,提出了一种障碍物视觉检测识别算法。算法首先对机器人采集的原始图像进行中值滤波以减少图像噪声并减小障碍物内部的灰度差异;然后利用Otsu算法来确定Canny算子的参数,较好地提取出图像的边缘;最后利用改进的随机Hough变换(RHT)提取出几何图形基元,并施加一定的结构约束,实现了对障碍物的检测识别。在模拟线路上的实验结果表明,该方法能有效地对高压输电线路的导线以及防震锤、绝缘子等障碍物进行检测识别。 Obstacle detection and recognition is one of the key techniques in transmission line deicing robot.A vision based obstacle detection and recognition algorithm is designed based on the structure of 220 kV transmission line.Firstly,the original image captured by the robot is median-filtered to reduce noise and gray-scale difference within the obstacles.Then the Otsu algorithm is used to determine the parameters of Canny operator in order to extract edges from the image.Finally,the improved random Hough transform(RHT) is used to extract the geometry elements.Given certain structure restraint,the obstacles are detected and recognized.Experiment results for simulation transmission line show that the algorithm is effective for wire and obstacles such as vibration damper and insulator.
关 键 词: 除冰机器人 障碍识别 图像处理 结构约束 变换
领 域: [自动化与计算机技术] [自动化与计算机技术]