船用多色宽幅钢板表面缺陷在线检测系统研究-机械工程专业论文.docxVIP

哈尔滨工业大学工程硕士学位论文 哈尔滨工业大学工程硕士学位论文 - - II - Abstract In ship manufacture process, large numbers of color steel strips are needed. Surface quality not only affects ship appearance, but also affects corrosion resistance, abrasion resistance and fatigue strength etc. To detect steel strip surface defects, the method of manual eyeballing is adopted generally, which costs much time and power, can not ensure detection quality, and suppresses the competition of ship building factory. Therefore, the development of surface defects detection system for wide color steel strips is an important project. Accordding to the features of that the gray acale of surface defects is different from the others, a set of surface defects online detection system is developed. The system can classify surface defects by image recognition technique with color and gray scale, and control the marking system to print defects area when defects are founded. On the basis of analyzing the marine steel strip surface defects and considering the basic performance requirements of dection system, the whole scheme of the surface defects detection system is designed，then hardware structure and control chart of the detection system are deseribed in detail. As marine steel strips are wide and colorful, at the same time their reflectivity is very low, the plan using many high-frequency fluorescent lamps is proposed to overcome the defects of uneven intensity and fuzzy image under common fluorescent light, and to ensure that brightness is even on steel strips surface with 4 meters width, and image is clear. The marking alarm module is designed and implemented. When defects are identified, the industy computer controls the module to work. The pictures from the imaging system are preprocessed. According to the image features of marine steel strip surface defects, the object image spectrum is decomposed to eliminate effects by color background. Based on the differences of color and gray scale between defcts with backgr