Defect detection of 3D printing surface based on geometric local domain features

The 3D printing process has advantages in the processing of complex parts, and its application is more and more extensive. However, the 3D printing process is still immature compared to traditional machining, it is more prone to produce defect products. Detecting defects on the surface of 3D printed parts is very important to ensure product quality. However, existing methods are barely capable to detect small 3D printing defects, as small defects are partially submerged by noise. In order to solve the problem, we proposed a novel method based on 3D machine vision to detect small 3D printing defects in this paper. The proposed method can be divided into two steps, potential defect regions extraction and accurate defect detection. Specifically, in the first stage, MBH (moving least squares smoothing and rough boundary removal-based Harris keypoint) and INRoPS (Improved Normal Rotated Projection Statistics) feature descriptor is proposed to extract the potential defect region, which has the advantages of robustness and high computational efficiency. In the second stage, a defect detection method based on neighborhood point calculation is proposed, which can accurately extract the complete shape of the defect. Experiments prove that the proposed method is more accurate and robust than other methods for both potential defect region extraction and precise defect detection.

Automated summary

In this paper, a novel method based on 3D machine vision is proposed to detect small 3D printing defects. The method includes two steps, potential defect region extraction and accurate defect detection. Experiments prove that the proposed method is more accurate and robust than other methods for both potential defect region extraction and precise defect detection.