This paper proposes a magnetic tile surface defect detection method based on improved ho-momorphic filtering and the MT(Magnetic tile)saliency fusion model to address the hard deeetection of defect types of the magnetic tile image due to uneven illumination and complex background texture.The study involves using an improved homomorphic filtering function to separate the high-frequency and low-frequency components of the image,linearly overlaying the high-frequency and low-frequency images en-hanced by the limited contrast histogram equalization method,using a universal model of MT significance fusion to detect different types of the magnetic tile defects.The result shows that the final defect detection is obtained by using morphological closure and noise removal algorithms,the proposed method can effec-tively and conrrectly detect the different types of magnetic tile surface defects,with a correct detection rate of 92.95%.
关键词
表面缺陷检测/显著性检测/同态滤波/AC模型
Key words
surface defect detection/significance detection/homomorphic filtering/AC model
维普
万方数据