Artificial Intelligence Based Bird Census System

Birds play a vital role in connecting nature and life, serving as key indicators of Earth’s biodiversity. Recent research indicates that global warming, climate change, and pollution significantly impact living creatures. Understanding migratory bird behavior is essential for studying ecology. Among the bird sanctuaries of Karnataka, the population of birds in Rangana Thittu, Kokkare Bellur, and Kaggaladu, each of which is officially known as a wildlife refuge, is recorded for migratory reasons with regular bird censuses. Producing exhaustive headcounts using photographs is a skill that is difficult but absolutely vital for ecologists. Through true-to-life shots taken at various bird havens, this project is working on Artificial intelligence (AI) enabled bird counting. We hope that by using Faster Region-based Convolutional Neural Network (Faster RCNN) and YOLOv5 to count birds, we can get to a stage where the two systems are at least as accurate and efficient as traditional counting methods. The experimental results demonstrate YOLOv5 performs well compared to Faster RCNN. This novel method provides an invaluable insight into our avian ecology, helping to unravel their relationships within ecosystems and contributing towards the conservation of biodiversity.