Deep Learning for Avian Monitoring: Bird Detection at Chilika Lake

Chilika Lake serves as an important habitat for migrating birds. Located on the east coast of India, in Odisha, it is the largest wintering ground for migratory waterfowl in the Indian subcontinent, covering over 1100 km \(^{2}\) . Manual counting and tracking can be a difficult task due to the vast area and the large number of birds and species visiting the lagoon throughout the year. This study focuses on a method to detect and classify Cormorant birds, specifically three species: the Little Cormorant (Phalacrocorax niger), the Great Cormorant (Phalacrocorax carbo), and the Indian Cormorant (Phalacrocorax fuscicollis). The approach involves analyzing video files containing these birds and processing each frame using a custom parallel processing system. The method achieved an accuracy of 99.27% across the three species. This study proposes an automated process for bird detection as an alternative to traditional manual counting methods.