Elephant–railway conflict minimisation using real-time video data and machine learning

Elephant–train collision has been a major issue for both the railway as well as the forest departments. In this study real-time video data is analysed for detecting elephant to alert the train driver in case of elephants crossing the railway track in which the train is approaching. The HAAR feature extraction and adaptive boosting-based machine learning algorithm are used for detecting elephants from real-time video data. The experimental result shows the average precision of the proposed technique in detecting elephants using real-time video data is more than 96%.