基于遗传优化神经网络在高压瓷瓶裂缝识别中的应用

Application of Neural Network Based on Genetic Algorithm in Recognition of Porcelain Bottles Crack

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。本文通过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五个特征值,得出图像信息。最后利用遗传算法和BP网络、RBF网络相结合的算法,分别实现对绝缘瓷瓶裂缝五种状态的分类识别。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它领域。

In order to ensure the security of power transmission lines, the power transmission lines can be inspected by vision system of inspection robot. In this paper, the images of porcelain bottles crack are collected by some hardware such as CCD, which is used to simulate the vision system of the robot. It is described how to finish the orientation of crack in the picture by the image preprocessing, such as smoothing and segmenting the object from the background by the threshold and eigenvector of figure. Five features are extracted from the image, which can reflect the information of the whole image. Finally, the algorithm is designed by combining genetic algorithm with BP and RBF network. It is used to realize the classifications, which divide porcelain bottles cracks into five classifications. These algorithms can be used effectively and reliably in recognition of porcelain bottles crack types by analyzing and comparing the results of the simulation in the experimentation. These algorithms can also be used effectively in other fields.