机器视觉及激光轮廓扫描的电流互感器外观信息自动识别算法

为了提高电流互感器外观信息的识别精度,对基于机器视觉及激光轮廓扫描的电流互感器外观信息自动识别算法进行了研究.利用线激光轮廓传感器获取了电流互感器外观轮廓的三维点云数据,依据三维点云数据重构了电流互感器外观轮廓图像,丰富了外观细节信息.按照外观图像形状信息特征参数,采用二叉决策树分类器识别了污渍与裂纹等外观形状信息.利用形态学字符定位方法与垂直投影法,自动识别外观图像空间信息即字符信息.实验结果表明,该算法识别电流互感器外观信息的准确率高于９９％ ,识别速率为 １．０１ 个/ s ,具备较好的自动识别性能.

Automatic Recognition Algorithm of Current Transformer Appearance Information Based on Machine Vision and Laser Contour Scan

In order to recognize the accuracy of the appearance information of the current transformer , an automatic recognition algorithm for the appearance information of the current transformer based on machine vision and laser profile scanning is studied．The line laser profile sensor is used to obtain the three-dimensional point cloud data of the appearance contour of the current transformer , and the appearance contour image of the current transformer is reconstructed according to the three-dimensional point cloud data , which enriches the appearance detail information．According to the characteristic parameters of the appearance image shape information , a binary decision tree classifier was used to identify the appearance shape information such as stains and cracks．Using the morphological character positioning method and the vertical projection method , the spatial information of the appearance image is automatically recognized , that is , the character information．Experiments show that the accuracy of the algorithm for identifying the appearance information of current transformers is higher than ９９％ , and the recognition rate is １.０１ piece / s , which has good automatic recognition performance．