Performance Analysis of YOLO for Object Detection under Complex Illumination Conditions

Abstract

Object detection has become a crucial component in various computer vision applications, ranging from autonomous driving to surveillance systems. YOLO (You Only Look Once) has gained prominence due to its real-time performance and relatively high accuracy. However, its performance can be significantly affected by complex illumination conditions such as underexposure, overexposure, shadows, and varying light sources. This review paper provides a comprehensive analysis of YOLO's performance under these challenging illumination conditions. We begin with an overview of the YOLO architecture and its evolution, followed by a detailed exploration of how different illumination factors impact its detection accuracy and speed. We then delve into various techniques proposed to mitigate these issues, including image enhancement methods, adaptive thresholding approaches, and robust feature extraction strategies. Furthermore, we comparatively analyze the performance of different YOLO variants and other state-of-the-art object detectors under diverse illumination scenarios. The paper synthesizes the current research landscape, highlights the key challenges that remain, and discusses potential future directions for improving the robustness of object detection algorithms in adverse lighting conditions. This review aims to serve as a valuable resource for researchers and practitioners seeking to understand and address the limitations of YOLO in real-world applications with complex illumination.