Advancements and Challenges in Low-Light Object Detection

The study of computer vision has been drawing tremendous attention to low light object detection as the need for reliable vision systems that can function in difficult lighting circumstances is growing. Low levels of light pose a significant challenge to computer vision systems, particularly in the domain of object detection. While some efforts have been made to improve lighting, others have been concentrated on lowering noise. This research offers a thorough examination of both conventional and deep learning techniques used in low light object recognition, a crucial area of computer vision with important applications in everyday life. First, the basic ideas and techniques of classical algorithms, is highlighted. The discussion then turns to deep learning techniques, emphasizing current developments and their increasing prominence in the industry. A significant portion of the study examines current related works and carefully weighs the advantages and disadvantages of each. This comparison highlights the advancements in algorithmic sophistication while also illuminating the inherent trade-offs associated with various strategies. Typical obstacles including fluctuating illumination levels, noise interference, and dynamic range constraints are carefully investigated to comprehend how they affect algorithms. The study also describes different performance metrics that are used to assess object detection algorithms in low light. Analysis results are summarized and future research avenues are suggested in an effort to close the gaps in the state-of-the-art approaches and improve the performance of low-light object detection algorithms.