Design of an Optical Distortion Measurement System for Laminated Windshields Using Artificial Vision

Abstract

The evaluation of optical distortion in laminated windshields is a critical aspect of automotive safety and manufacturing quality control. Traditional inspection methods rely on manual visual assessment, which introduces subjectivity, variability, and human error. This research presents the design and implementation of an automated measurement system based on artificial vision and image processing to accurately detect and quantify optical distortion in windshields. The proposed system integrates OpenCV and Python to process high-resolution images captured under controlled conditions. A dataset of 100 windshield samples was analyzed, achieving an accuracy of 94.0%, precision of 94.3%, and recall of 94.0%, demonstrating superior performance compared to manual inspection. The methodology includes image acquisition, edge detection using the Canny algorithm, and circular shape recognition via Hough Transform. The experimental results confirm that artificial vision provides a faster, more reliable, and cost-effective alternative to conventional inspection methods. This work contributes to the automation of optical quality control, reducing production defects and improving windshield safety standards. Future work will explore deep learning techniques to enhance detection robustness under variable lighting conditions.