Researchers at the Indian Space Research Organization (ISRO)’s Space Applications Center (SAC) have developed an artificial intelligence model that offers a comprehensive set of airport control and aircraft tracking functions, as well as than surveillance.
The Times of India reports that the model was developed in collaboration with an institute and tested at airports in Ahmedabad, Mumbai and Pune. The model uses CosmiQ Works’ RarePlanes dataset, an open-source machine learning dataset from CosmiQ and AI Reverie, which incorporates real-world and synthetically generated satellite imagery.
In addition to the RarePlanes dataset, the model was trained on the Airbus aircraft detection dataset and Indian remote sensing satellite imagery.
“The deep learning data models were called YOLOv5 and YOLOv7, which belong to the You Only Look Once (YOLO) family, widely used in various applications, including aircraft detection in satellite imagery,” said a senior official at the Times of India.
The model would have an accuracy level of 94% for large planes and 88% for small ones. YOLOv7 performed better in these tests. The research team is currently working on extending the model’s capabilities and integrating synthetic aperture radar (SAR) satellite data, which can contribute to aircraft detection in extreme weather conditions.
Related News :