A Study on Object Detection Performance Through Data Augmentation Under Adverse Weather Conditions

This study compares and analyzes the YOLOv8, Faster R-CNN, and SSD models using transfer learning techniques to improve object recognition performance in autonomous vehicles. Road environments vary significantly due to external factors such as weather and road conditions, making object recognition challenging. To overcome these issues, data augmentation techniques were applied to artificially modify the training data. The evaluation of the three representative object detection algorithms revealed that YOLOv8 demonstrated balanced performance in terms of accuracy and speed under adverse conditions.