基于轻量化语义分割网络的织物复杂纹理 疵点分割方法.

朱琳; 马爽

Aiming at the problem of textile and garment complex texture surface defect segmentation, the existing automatic detection methods still need to be improved in terms of computational complexity, segmentation accuracy and real-time performance. To solve this problem a method based on lightweight semantic segmentation network was proposed* The image enhancement technology was used to improve the clarity of the complex texture image of textile and garment and the salient region extraction method was used to locate the candidate defect area. A lightweight semantic segmentation network was constructed to accurately segment defects by using semantic information* The experimental results show that the proposed, method has obvious advantages in terms of average recall area which proves the efficiency and. accuracy of the proposed method, in defect segmentation task, and. helps to improve product quality and production efficiency.

SURFACE texture; IMAGE intensifiers; SURFACE defects; COMPUTATIONAL complexity; TEXTURED woven textiles

Wool Textile Journal, 2024, Vol 52, Issue 12, p102

1003-1456

Academic Journal

10.19333/j.mfkj.20240707806