Animal Detection in Night Light Videos Using Deep Learning
Sudharshan Duth P, Maneesha Shibu, Nithyashree M · 2024
Nighttime surveillance and monitoring of wildlife are critical for various applications, including conservation efforts, ecological research, and security purposes. Understanding the behavior of animals makes it difficult to predict the presence of animals on roadsides, potentially resulting in human-animal conflicts. This research work provides a novel approach for detecting animals in nighttime videos using deep learning algorithms. The main aim is to effectively detect and recognize the nocturnal presence of the animal by which surveillance at night and animal wellbeing can be improved by contributing to SDG 3, 11 and 12. In this scenario, a set of 10 animals were taken as classes, and the experimental was carried out for animal detection and recognition utilizing the YOLO v8 model which enhances the efficiency and accuracy of animal detection at night when compared to other models. Overcoming the challenges of low-light images, blur images and occlusions. Experimental results showcase superior performance compared to existing models exhibit real-time processing capabilities, making it suitable for practical deployment in nocturnal wildlife monitoring systems. The animal elephant is identified more accurately in this study, with a 87.49% detection accuracy.