Defect detection method for complex surface based on human visual characteristics and feature extracting

Du, Yubin; Cao, Pin; Yang, Yongying; Wang, Fanyi; Liu, Rongzhi; Wu, Fan; Zhang, Pengfei; Chai, Huiting; Jiang, Jianjun; Zhang, Yihui; Feng, Guo-Hua; Xi, Xiang; Li, Yanwei

doi:10.1117/12.2511490

Cited by 1 publication

(1 citation statement)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared with the defect detection of motor armatures of finished products [9], the production line detection based on machine vision is more economical. For the defect detection of non-standard micro workpieces, it is difficult to quantitatively distinguish the defect level and category by classic machine vision technology [10,11,12]. Meanwhile, with the development of visual detection based on artificial neural networks [13,14], in which image features are fed into the region of interest (ROI) for classification and segmentation [15,16,17], the accuracy of detection is constantly improved.…”

Section: Introductionmentioning

confidence: 99%

An Industrial Micro-Defect Diagnosis System via Intelligent Segmentation Region

Xia

Wang

Feng

2019

Sensors

View full text Add to dashboard Cite

In the field of machine vision defect detection for a micro workpiece, it is very important to make the neural network realize the integrity of the mask in analyte segmentation regions. In the process of the recognition of small workpieces, fatal defects are always contained in borderline areas that are difficult to demarcate. The non-maximum suppression (NMS) of intersection over union (IOU) will lose crucial texture information especially in the clutter and occlusion detection areas. In this paper, simple linear iterative clustering (SLIC) is used to augment the mask as well as calibrate the score of the mask. We propose an SLIC head of object instance segmentation in proposal regions (Mask R-CNN) containing a network block to learn the quality of the predict masks. It is found that parallel K-means in the limited region mechanism in the SLIC head improved the confidence of the mask score, in the context of our workpiece. A continuous fine-tune mechanism was utilized to continuously improve the model robustness in a large-scale production line. We established a detection system, which included an optical fiber locator, telecentric lens system, matrix stereoscopic light, a rotating platform, and a neural network with an SLIC head. The accuracy of defect detection is effectively improved for micro workpieces with clutter and borderline areas.

show abstract