基于机器视觉的重轨表面缺陷成像方法与系统设计-机械工程专业论文.docxVIP

万方数据 万方数据 Abstract In this dissertation, aiming at the complex geometrical characteristic and the low reflection and high radiation of the heavy rail surface, the imaging method of the heavy rail surface defects was researched on the base of the theory and technology of machine vision. The imaging system of heavy rail surface defects was achieved with the optimized design of imaging angle and the development of special lighting system. The system further enhance the imaging effect of heavy rail surface defects and provide the basis for the development of heavy rail surface defects on-line detection system. The main works of the research are as follows: Starting with the production processes of heavy rail, the geometric features and surface defects of heavy rail surface were analyzed in detail, and the reasons and regularity of defects were summarized. The light transmittance, reflectance and thermal radiation characteristics of thermal heavy rail surface were analyzed in detail combined with physical optics theory and radiation theory. The overall design of heavy rail surface defects imaging system was proposed on the base of the researches and machine vision theory system. The best segmentation scheme of the FOV (field of view) was chose for geometrical characteristic of heavy rail surface. The image acquisition system was designed combined with the selection of camera and lens. The imaging angles of cameras were computed with the optimization design on the base of the segmentation of cameras view field, and the best angles of cameras were selected through the imaging quality evaluation of the heavy rail surface. To improve the feature information of heavy rail surface, the effects of light source parameters and illumination mode were analyzed and the best image enhancement technologies were chose and applied on heavy rail surface images which were took under different illumination mode. The best illumination mode for different defects was chose by the effect compa