基于线阵扫描方式下的机车车顶状态检测方法

针对机车车顶异物存在和瓷瓶受损影响机车运行安全的问题，本文提出了基于线阵扫描方式下的机车车顶状态检测方法。论文用两台黑白线阵相机乒乓采集车顶图像，基于车速将图像校正后拼接成全景图，并利用区域划分法来划分大场景车顶图片，最后用SIFT特征点提取、改进后的RANSAC图像配准等算法与模板图匹配来定位瓷瓶、查找异物。实验结果表明该方法快速、有效，能够准确地找出检测图中的异物和定位瓷瓶。

Detection Method of the State of Locomotive Roof Based on Liner Array Scanning

Foreign objects and damaged porcelain insulators will cause accidents during operation of the locomotive. In order to solve the problem, a detection method of the state of locomotive roof based on liner array scanning is proposed. In this method, two monochromic line scan cameras are used to take images on ping-pong mode and roof image mosaic is completed after all images are corrected based on the speed. Then, a characteristic points based region division algorithm is proposed to divide the big complete roof image. Lastly, template match algo- rithm is used to detect foreign objects as well as to locate the porcelain insulators with SIFT characteristic point extraction algorithm and the improved RANSAC image registration algorithm. The result shows the proposed strategy is effective and efficient and can detect foreign ob- jects and locate porcelain insulators accurately.