Pune: Junnar Forest Division Explores AI-Based System to Monitor Leopard Sightings

Pune, 27th August 2024: The Junnar division of the forest department is considering the implementation of an artificial intelligence (AI) system designed to alert both officials and villagers of leopard sightings by analyzing the animals’ behavior. This innovative approach involves installing camera traps at strategic locations to monitor wildlife activity.

The proposed system would utilize camera traps capable of capturing images of leopards or other animals within a range of 500 meters. These images would be transmitted via an Internet of Things (IoT) platform to a cloud server, where AI technology would process the data. The AI system would compare the captured images with an existing database of leopard photos and locations, helping to identify the behavior, timing, and movement patterns of the leopards.

Once the system identifies a potential threat based on the leopards’ behavior and movement, it would send immediate alerts through emails and text messages to the registered cellphone numbers of forest officials and local villagers near the sighting.

While this system shows promise, there are significant challenges to its implementation. During a field visit, engineers discovered that the proximity of homes to sugarcane fields—where leopards often roam—poses a timing issue for the alerts. “The houses are just 10 meters from the fields, and we need at least three seconds to send the alerts via satellite. Unfortunately, this isn’t enough time, as the leopard could already be at their doorstep,” explained Piyush Dhulia, Programme Director of Valiance Solutions.

Valiance Solutions, the company proposing the system, has previously implemented similar technology in the Tadoba Andhari Tiger Reserve, where it successfully reduced human-tiger conflicts. The system has also been replicated in 13 additional villages near Tadoba, as well as in Manas Tiger Reserve and Corbett National Park.

However, the conditions in Junnar present unique challenges. Unlike Tadoba, where settlements are at least 200 meters away from wildlife habitats, the proximity of human homes to the fields in Junnar requires rapid alert responses. “In Tadoba, we had the luxury of more time to process and send alerts, but in Junnar, the landscape demands quicker reaction times,” Dhulia noted.

Thermal cameras, which have been effective in other areas, might not be suitable for Junnar due to the dense sugarcane fields and high moisture content. Instead, the team may need to employ specialized vision cameras to achieve the desired results. Additionally, the scattered nature of homes in Junnar, as opposed to the clustered settlements near Tadoba, further complicates the system’s deployment. “Given the layout, nearly every house would need its own system,” Dhulia added.

Forest officials in Junnar are currently visiting villages identified as hotspots to gather detailed information, including GPS coordinates and records of past leopard attacks. “We aim to map these households and create a detailed map of the hotspots, which will help us better understand leopard activity and movements,” said Amit Bhise, Assistant Conservator of Forests in Junnar.

The first trial of this AI-based alert system is set to take place in Narayangaon, Umraj, and Bori—areas where human settlements are somewhat distant from leopard habitats, allowing for better observation of leopard movements. Deputy Conservator of Forests Amol Satpute confirmed that the initial system has been set up in Bori, a location known for its high leopard density.

“The success of this project in Junnar will hinge on our ability to send timely alerts to forest officials and villagers, adapting the system to the unique landscape here,” Dhulia concluded.