Photometric Stereo-Based 3D Reconstruction Method for the Objective Evaluation of Fabric Pilling

Fabric pilling evaluation has been considered as an essential element for textile quality inspection. Traditional manual method is still based on human eyes and brain, which subjective with low efficiency. This paper proposes objective semi-calibrated near-light Photometric Stereo (PS). images are digitalized by self-developed image acquisition system. The 3D depth information of each point could be obtained PS algorithm then mapped to 2D grayscale image. After that, the non-textured filtered using Gaussian low-pass filter. segmentation conducted global iterative threshold method, K-Nearest Neighbor (KNN) finally selected a tool grade classification fabric pilling. Our experimental results show that proposed system achieve excellent judging performance evaluation.