Abstract <p>Accurate detection and measurement of holes in plate images are critical for quality control in manufacturing. This work aims to develop a non-contact, computer vision-based system capable of detecting and measuring holes in metal plates with high accuracy. This paper presents a comprehensive system combining YOLO (You Only Look Once) object detection with image processing to address the accurate measurement of holes in a metal plate. The proposed method utilizes YOLOv8 to detect plates within the images, followed by preprocessing to enhance image quality. Perspective transformation is applied to correct plate orientation, ensuring precise measurements. Measurements of the hole’s diameter and inter-hole distances in pixels are performed and are then converted to real-world measurement (in mm) using fixed and reference values. Experimental evaluation demonstrates high detection accuracy and measurement precision, with the system successfully identifying holes and visualizing results using bounding boxes and annotations. The outcomes confirm the effectiveness of the proposed approach as a robust, automated solution for industrial inspection and quality assurance, with potential for real-time integration and scalability in future applications.</p> Graphical Abstract <p></p>