A Comprehensive Framework for Automated Quality Control in the Automotive Industry

2025-12-05

AI Summary

To address low detection accuracy, high false-positive rates, and poor production-line adaptability in surface and thread defect inspection of aluminum high-pressure die-cast (HPDC) automotive components, this paper proposes a vision-based quality inspection framework tailored for fully automated production lines. The system employs a dual collaborative robotic platform equipped with high-resolution cameras, custom lenses, and optimized ring lighting, integrated with image tiling preprocessing, an enhanced YOLOv11n model (incorporating ensemble learning and bounding-box fusion), and quantized image processing algorithms to achieve precise defect localization and severity assessment. Compared to baseline models, the framework reduces false positives by 32.7% and achieves an mAP@0.5 of 98.4%, while enabling rapid cross-part-type and cross-line deployment. Key contributions include: (i) the first application of a dual-robot cooperative vision system to HPDC thread defect inspection; and (ii) a lightweight, efficient object detection optimization strategy that significantly improves robustness for small-object detection without compromising real-time performance (≥25 FPS).

Technology Category

• Computer Vision: Vision for Robotics & Autonomous Driving
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Hardware-aware ML

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

This paper presents a cutting-edge robotic inspection solution designed to automate quality control in automotive manufacturing. The system integrates a pair of collaborative robots, each equipped with a high-resolution camera-based vision system to accurately detect and localize surface and thread defects in aluminum high-pressure die casting (HPDC) automotive components. In addition, specialized lenses and optimized lighting configurations are employed to ensure consistent and high-quality image acquisition. The YOLO11n deep learning model is utilized, incorporating additional enhancements such as image slicing, ensemble learning, and bounding-box merging to significantly improve performance and minimize false detections. Furthermore, image processing techniques are applied to estimate the extent of the detected defects. Experimental results demonstrate real-time performance with high accuracy across a wide variety of defects, while minimizing false detections. The proposed solution is promising and highly scalable, providing the flexibility to adapt to various production environments and meet the evolving demands of the automotive industry.

Problem

• Automates quality control in automotive manufacturing
• Detects surface and thread defects in aluminum castings
• Uses deep learning for real-time, high-accuracy inspection

Innovation

• Robotic vision system with high-resolution cameras for defect detection
• Enhanced YOLO11n model with slicing and ensemble learning
• Specialized lenses and lighting for consistent image acquisition