In textile industry production, fabric defect inspection is a vital step to ensure the quality of fabric before spreading, cutting and so on. Recently, image characteristic of nonlocal self-similarity (NSS) is widely applied to image denoising due to its effectiveness. Actually, fabric defect detection can be considered as a problem that finds noises in an image. Based on the reason, we propose a simple yet effective method, namely nonlocal low rank approximation (NLRA), for fabric defect detection. In NLRA, an image to be processed is divided into many patches. For a given patch, we search its several similar patches and group them as a matrix. Then, the clean image patch can be reconstructed through solving the low rank approximation of the matrix. Finally, a new image will be synthesized from these estimated patches, the defects can be located by finding the difference between the original fabric image and the reconstructed image. Experimental results prove the validity and feasibility of the proposed NLRA algorithm.
展开▼
