货车转向架交叉杆弯曲故障辨识算法研究

Study on Recognition Algorithm of Curving Fault of Cross Rod of Goods Wagon Bogie

以货车转向架交叉杆弯曲故障的智能检测为背景,提出了一种针对图像信息中刚体结构弯曲形态进行模式识别的算法;从分析故障特征入手,通过梯度投影分析线性分割特征提取出交叉杆区域并将交叉杆旋转至与水平向垂直的位置;基于Sobel算子提取出交叉杆边缘;针对交叉杆边缘的分布,提出了一种应用Hough变换分段求取边缘线段极角均值估计的分类器设计算法;通过实验,验证了算法的有效性,并分析了算法的实时性。

This paper proposed an algorithm on intelligent recognition on the curved fault on the basis of the cross rod of goods wagon bogie in the image.After analyzing the character of curved fault of cross rod,the region including the cross rod is extracted by using the gradient projection,then turn the cross rod in the vertical position in the image.The edge of cross bar is extracted by using the Sobel Operator.According to the distribution feature of the edge maps,the classification based on the average value of polar angle of the segmented lines of edge obtained by the Hough Transformation is designed.Demonstrated by the experiment,the algorithm is precise and efficient.