摘要
为了保证高压输电线路的正常运行,可以通过高压输电线路巡检机器人视觉系统完成高压输电线路的检测。通过CCD摄像头等硬件模拟机器人的视觉,完成对绝缘瓷瓶裂缝图像的采集,并通过滤除噪声、图像分割等预处理操作和形状特征,完成图像中裂缝的定位。对于聚焦放大后的裂缝图像提取不变矩等4个特征值,得出图像信息。最后利用遗传算法和RBF网络相结合的算法,实现对绝缘瓷瓶裂缝5种状态:横向、纵向、块状、网状、无裂缝的分类识别。通过仿真和实验比较表明,该算法可以有效、可靠地运用于绝缘瓷瓶裂缝类型识别研究中,并可方便地应用于其他领域。
To ensure the security of power transmission lines, an inspection method is presented by using the vision system of inspection robot. The images of porcelain bottle cracks are collected by CCD, which is used to simulate the vision system of the robot. The cracks in the picture are located by the image preprocessing, such as smoothing, segmenting and eigenvector of images. Four features, such as moment invariants ect. , are extracted from the image to get the information of the whole image. An algorithm is designed by combining genetic algorithm with RBF network. The porcelain bottles cracks are divided into five types： transverse, longitudinal, block, alligator and non-distress. The simulation and experiment results show that the algorithms could be used effectively and reliably in recognition of porcelain bottles crack types. The algorisms could also be used effectively in other fields.
出处
《控制工程》
CSCD
北大核心
2009年第5期561-565,共5页
Control Engineering of China
基金
国家863基金资助项目(2005AA420064)
