Real-Time Object Detection for Night Vision Scenarios Using Deep Learning

Abnormal activity detection is crucial for effective surveillance applications. While daytime surveillance has benefited from deep learning algorithms for object detection and tracking, achieving similar performance in night vision is challenging due to low illumination. Deep learning, a powerful machine learning technique, automatically learns image features for detection tasks. This project aims to develop a model capable of detecting objects under low illumination, specifically using thermal infrared images. The proposed approach focuses on utilizing thermal infrared images to detect external objects. The performance of the developed model is evaluated through experimentation and analysis, including comparative studies with existing methods. This assessment provides insights into the advantages and limitations of using deep learning for object detection in low-light conditions. With the continuous growth of deep learning, more efficient techniques have been implemented to address the challenges associated with object detection using neural networks and deep learning.