Wood chip crack detection based on linear scale-space differential. (April 2021)

• Record Type: Journal Article
• Authors: Li, Shaoli; Li, Dejian; Yuan, Weiqi
• Is Part Of: Measurement. Volume 175(2021)
• Publication Date: 2021-04
• Subjects: Machine vision -- Defect detection -- Crack -- Multi-layer post-feedback LSD model; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2021.109095
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: A MP-LSD model for the extraction of wood chip crack defects is proposed. The feature of the model is suppressing the noise of wood fiber furrow. The principle of the model is the multiple iteration of LSD algorithm. The crack is regarded as the combination of points with different width and depth. Abstract: In this research, the wood chip refers to the product such as tongue spatula, which is rectangular and thin. And as the common production defect on the surface of the wood chip, crack threatens the safety of users' oral cavity seriously. However, some thin and shallow cracks are very similar to the wood fiber furrow on the wood surface, which makes the extraction effect of the current detection algorithm unsatisfactory. Aiming at this problem, a method based on multi-layer post-feedback LSD (linear scale-space differential) model is proposed in this paper. The aim of the model is to extract the crack defects and suppress the noise of wood fiber furrow to the greatest extent. We conduct the experiments with the images captured by a self-established automatic optical inspection system. Then the effect of the proposed algorithm is compared with the state-of-the-art methods. And we also conduct the contrast experiment with the current manual method via the established system and the result shows higher effectiveness.