Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Automated Defect Detection for Mass-Produced Electronic Components Based on YOLO Object Detection Models

Wei-Lung Mao, Chun-Chi Wang, Po-Heng Chou, Yen-Ting Liu

TL;DR

This work tackles automated defect detection in mass-produced DIP components, addressing limited defective data with ConSinGAN augmentation and evaluating multiple YOLO detectors. The study finds that YOLOv7 augmented with ConSinGAN delivers the best accuracy (mAP0.5 ≈ 95.5%) and a fast detection time (~285 ms), outperforming threshold-based baselines by a wide margin. The proposed system is embedded in a SCADA-enabled production line, enabling real-time, six-sided DIP inspection with minimal manual intervention. This approach demonstrates a practical path toward scalable, data-efficient AI-powered quality control in electronics manufacturing.

Abstract

Since the defect detection of conventional industry components is time-consuming and labor-intensive, it leads to a significant burden on quality inspection personnel and makes it difficult to manage product quality. In this paper, we propose an automated defect detection system for the dual in-line package (DIP) that is widely used in industry, using digital camera optics and a deep learning (DL)-based model. The two most common defect categories of DIP are examined: (1) surface defects, and (2) pin-leg defects. However, the lack of defective component images leads to a challenge for detection tasks. To solve this problem, the ConSinGAN is used to generate a suitable-sized dataset for training and testing. Four varieties of the YOLO model are investigated (v3, v4, v7, and v9), both in isolation and with the ConSinGAN augmentation. The proposed YOLOv7 with ConSinGAN is superior to the other YOLO versions in accuracy of 95.50\%, detection time of 285 ms, and is far superior to threshold-based approaches. In addition, the supervisory control and data acquisition (SCADA) system is developed, and the associated sensor architecture is described. The proposed automated defect detection can be easily established with numerous types of defects or insufficient defect data.

Automated Defect Detection for Mass-Produced Electronic Components Based on YOLO Object Detection Models

TL;DR

This work tackles automated defect detection in mass-produced DIP components, addressing limited defective data with ConSinGAN augmentation and evaluating multiple YOLO detectors. The study finds that YOLOv7 augmented with ConSinGAN delivers the best accuracy (mAP0.5 ≈ 95.5%) and a fast detection time (~285 ms), outperforming threshold-based baselines by a wide margin. The proposed system is embedded in a SCADA-enabled production line, enabling real-time, six-sided DIP inspection with minimal manual intervention. This approach demonstrates a practical path toward scalable, data-efficient AI-powered quality control in electronics manufacturing.

Abstract

Since the defect detection of conventional industry components is time-consuming and labor-intensive, it leads to a significant burden on quality inspection personnel and makes it difficult to manage product quality. In this paper, we propose an automated defect detection system for the dual in-line package (DIP) that is widely used in industry, using digital camera optics and a deep learning (DL)-based model. The two most common defect categories of DIP are examined: (1) surface defects, and (2) pin-leg defects. However, the lack of defective component images leads to a challenge for detection tasks. To solve this problem, the ConSinGAN is used to generate a suitable-sized dataset for training and testing. Four varieties of the YOLO model are investigated (v3, v4, v7, and v9), both in isolation and with the ConSinGAN augmentation. The proposed YOLOv7 with ConSinGAN is superior to the other YOLO versions in accuracy of 95.50\%, detection time of 285 ms, and is far superior to threshold-based approaches. In addition, the supervisory control and data acquisition (SCADA) system is developed, and the associated sensor architecture is described. The proposed automated defect detection can be easily established with numerous types of defects or insufficient defect data.

Paper Structure

This paper contains 22 sections, 12 equations, 16 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Automated Defect Detection System
  3. Automated Mechanism
  4. Defective Types of DIPs
  5. Threshold-Based Image Detection
  6. DL-based YOLO Model and ConSinGAN
  7. Data Pre-Processing
  8. Data Augmentation
  9. ConSinGAN
  10. YOLO Object Detection Model Series
  11. YOLOv3 re25_YOLOv3
  12. YOLOv4 re26_YOLOv4
  13. YOLOv7 re27_YOLOv7
  14. YOLOv9 re_add_8_YOLOv9
  15. Performance Evaluation
  16. ...and 7 more sections

Figures (16)

  • Figure 1: System architecture diagram.
  • Figure 2: 3D simulation schematic.
  • Figure 3: Different sides of Normal DIP, (a) front, (b) back, (c) bottom, (d) top, (e) right, and (f) left sides.
  • Figure 4: Four types of defective DIP, (a) surface glue overflow, (b) surface scratches, (c) surface dirt, and (d) bent pins.
  • Figure 5: The defective features from threshold-based detection.
  • ...and 11 more figures