Research of Optimizing Goggles Defect Detection

• Main Authors: HSIEH, CHUNG-NAN, 謝忠男
• Other Authors: WANG, JING-WEIN
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/47hcv6

碩士 === 國立高雄科技大學 === 光電與通訊工程研究所 === 107 === Goggles are used in various fields included industry, medical science, sports, and normal using, but the products of protecting eyes cannot have any defects in skiing. Therefore, it is important and necessary to develop the goggles defect detection system. From hardware, we improve the surroundings of image capturing and the lighting design so that the original images will be much clearer than before. From software, in order to accelerate images processing, we improve the available system in our laboratory, which can detect the smallest defect down to size 200 um * 200 um. In order to prove whether the system is working or not, we test the new system with 120 goggles samples. After the experiment, we have three conclusions. First, the recognition rate of the new system is 100% which means we can find every defect on the goggles. Second, the time of inspecting every defect is 0.7 seconds on average, which decreases by 85% running time per piece. In conclusion, the new system certainly accelerates the goggles defect detection system.