Debrigarh Wildlife Sanctuary in Odisha has deployed AI-enabled, on-demand camera traps and an AI surveillance tower to identify wildlife, transmit live feeds and improve monitoring across remote forest areas. The system is the first of its kind reported in Odisha and is intended to support, rather than replace, conventional foot patrolling.
The camera traps use a Convolutional Neural Network-based artificial intelligence engine that can currently identify eight species: gaur, leopard, sambar, spotted deer, wild pig, sloth bear and nilgai. The system processes images on the device and filters records considered relevant to wildlife monitoring, reducing the need for officials to manually review large volumes of footage.
Forest authorities have also established a central control room to monitor live feeds from the surveillance network in real time. Officials said the system is expected to improve the tracking of animal movement, help identify unusual activity and support faster responses to potential threats, subject to connectivity and operating conditions.
“Foot patrolling remains our topmost priority. However, AI-based wildlife monitoring through AI-integrated camera traps has significantly strengthened wildlife monitoring in Debrigarh,” Anshu Pragyan Das, divisional forest officer of Hirakud wildlife division, said. Das added that the technology had improved surveillance efficiency and supported more informed decisions on wildlife protection and habitat management.
An AI surveillance tower has been installed in the Chourasimal area. According to officials, it has a reported data-capture radius of around 5-6km and covers approximately 70 sq km of the sanctuary. The tower-based system is being complemented by 18 AI camera traps deployed across a 2-3km forest stretch, with each unit placed roughly 200 metres apart.
The network is designed to extend monitoring into forest interiors and locations that are difficult to cover through regular patrols. The on-demand feature allows field personnel to remotely activate individual camera units and retrieve information when required. During daylight hours, the cameras capture photographs and videos, while thermal imaging is used at night to support surveillance in changing light conditions.
The cameras communicate through a wireless mesh network, allowing individual units to relay data to one another without relying on cellular SIM cards. This is intended to address weak or unavailable mobile connectivity in remote forest areas. The arrangement also means that the effectiveness of data transmission will depend on the operation of the mesh network and the availability of the surveillance units in the field.
Forest officials said the technology would help document species presence, understand habitat use and identify areas requiring focused conservation attention. They also made clear that AI monitoring is being integrated with existing protection practices, with foot patrolling continuing as the primary method of forest protection.
