摘要
对绝缘瓷瓶定位及其破损检测在高压输电线路巡检机器人视觉系统中的应用提出了一套方案。通过CCD摄像头等实现对机器人视觉的模拟,对采集到的图像进行去噪、分割等预处理,以绝缘子中部的特定区域作为识别的依据,基于其形状特征及不变矩特征提取特征值,提出了采用概率神经网络对绝缘子的特殊区域及背景区域进行识别,完成对绝缘子的定位。利用Harris角点检测法,结合纵向切线灰度值变化率判断瓷瓶是否破损,通过仿真和实验,验证了其可行性。
The methods of orientation and damage inspection of porcelain bottles applied in the vision system of inspection robot on the power transmission lines is proposed. Simulation of vision system of the robot is implemented by hardware such as CCD. The images collected by the system is preprocessed for filtering and segmenting the object from the background by the threshold. According to the characteristic region in the center of porcelain bottles, features are extracted based on the shape and moment invariants. In order to accomplish the orientation, probabilistic neural network is used for recognition of the characteristic region and background region. Damage of porcelain bottles is inspected by Harris comer detector and the gray level rate of longitudinal tangent. The method above is simulated and experimented. The simulation result of indicates the reliability.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第1期136-140,共5页
Computer Engineering and Design
基金
国家863高技术研究发展计划基金项目(2005AA420064)
关键词
绝缘子
特征提取
PNN神经网络
HARRIS角点检测
porcelain bottles
feature extraction
probabilistic neural network
Harris corner detector
