论文部分内容阅读
基于机器视觉技术、光电测试技术和图像处理技术研制轮轨接触几何状态可视化检测装置。检测装置的硬件系统包括高速摄像机、光源系统、测速单元等;软件系统包括图像采集软件和数据处理软件。检测装置的工作流程分为图像采集和数据处理2步。图像采集主要完成检测区域的标定图像、钢轨轨头廓型图像和轮轨接触图像的采集。数据处理主要完成标定参数的提取,矫正、拼接图像,识别接触曲线,获取检测结果。关键技术是:通过畸变矫正函数将拍摄的有畸变的图像还原为无畸变的图像;通过标定参数中标记点的信息,将矫正后的内外两侧图像拼接,得到轮轨接触的完整图像。现场实际应用表明:该装置能够准确检测轮轨接触状态,并给出接触参数和相关报表,达到了实用化的要求。
Based on machine vision technology, optoelectronic test technology and image processing technology, a wheel / rail contact geometry visual inspection device was developed. The hardware of the testing device includes high-speed camera, light source system, speed measuring unit and so on. The software system includes image acquisition software and data processing software. Detection device workflow is divided into image acquisition and data processing 2 steps. The image acquisition mainly completes the calibration image of the detection area, the head rail profile image of the rail head and the image acquisition of the wheel and rail contact image. Data processing to complete the calibration of the main parameters of extraction, correction, stitching images, identify contact curves, access to test results. The key technologies are: the distortion-corrected image is restored to the undistorted image through the distortion correction function; the corrected inner and outer images are stitched together by the information of the marked points in the calibration parameter to obtain the complete image of the wheel-rail contact. The practical application in the field shows that the device can accurately detect the wheel / rail contact status, provide the contact parameters and related reports, and meet the practical requirements.