AUTOMATIC VIDEO DETECTION FOR GOLDEN MONKEYS IN SHENNONGJIA NATIONAL PARK, CHINA

Abstract

Infrared-triggered cameras serve as essential tools for wildlife resource surveys, allowing camera traps to capture wildlife activity while minimizing the impact on the ecosystem. Traditional monitoring methods heavily rely on human resources for visual discrimination, which is inefficient and susceptible to environmental influences. Therefore, this paper focuses on Shennongjia National Park as a case study to address the limitations of traditional approaches. We constructed a Shennongjia wildlife object detection dataset using video data from the fixed-erected infrared cameras and proposed a supervised learning-based automatic monitoring method for golden monkeys, aiming to achieve intelligent target detection. With these processing results, we were able to capture golden monkeys' tracks and conduct statistical analysis of their life range and distribution characteristics. Through the integration of sensors and deep learning techniques, we developed a golden monkeys detection monitoring system to visualize the monitoring results and assess the spatial and temporal distribution characteristics of golden monkeys’ activities.