Object Detection Using YOLOv5 and OpenCV

Abstract

Object detection is one of the main tasks in computer vision, aimed at recognizing and localizing objects in images or videos. In this study, we utilize the YOLOv5 model, which is well known for its efficiency in realtime object detection. We implement this method with the help of the OpenCV library for image processing. This research aims to evaluate the performance of YOLOv5 in detecting objects in various types of images, including landscape photos, cat photos, and traffic light images with vehicles. The model is trained using optimization methods with the Adam optimizer and assessed through metrics like accuracy, precision, recall, and IoU. The results indicate that YOLOv5 can detect objects with high accuracy and fast inference time, making it an ideal solution for various applications such as security monitoring, video analysis, and automatic recognition systems. The advantage of YOLOv5 over traditional methods such as histogram equalization and thresholding lies in its ability to perform realtime detection with optimal computational efficiency. Thus, this study demonstrates that YOLOv5 is a suitable choice for implementing deep learningbased object detection systems.