CAJ | 학술논문

针对生产线喷涂件表面喷涂质量检测问题，设计了基于多特征的图像检索系统，用来检测喷涂件表面喷涂效果。基于图像的SIFT特征、纹理、颜色直方图构建图像检索系统；通过CCD图像传感器获取样例图像作为查询样本，由在线获取的图像流构成图像库，定义相似度函数及阈值，在图像流中查找是否存在满足阈值的图像。通过仿真实验表明：该方法能较准确地检测出生产线中的残次产品，为喷涂产品表面质量检测提供一个新的思路。
침대생산선분도건표면분도질량검측문제，설계료기우다특정적도상검색계통，용래검측분도건표면분도효과。기우도상적SIFT특정、문리、안색직방도구건도상검색계통；통과CCD도상전감기획취양례도상작위사순양본，유재선획취적도상류구성도상고，정의상사도함수급역치，재도상류중사조시부존재만족역치적도상。통과방진실험표명：해방법능교준학지검측출생산선중적잔차산품，위분도산품표면질량검측제공일개신적사로。
For quality evaluation on spraying production,we proposed a monitoring system based on image re?trieval. Firstly,we designed a contend-based image retrieval system based on SIFT feature,texture feature and color histogram. In the system we choose image from CCD imaging sensors as query sample and image database is made up of online images. Similar function and threshold value is defined and we look for images which fit threshold value. Simulation experiments show that the system can accurately detect the defect prod?ucts and provide a good idea for the product quality testing.