When the Telangana IT Department approached IIIT Hyderabad looking for a way to curb illegal sand mining, the challenge was not just regulatory — it was technical. India's trucks, unlike their counterparts in more standardized markets, often carry hand-painted registration plates: idiosyncratic in font, spacing, and style, and largely invisible to commercial automatic number plate recognition (ANPR) systems built for uniform plates.
The result of that collaboration is Vahan Eye, a field-deployed ANPR solution developed by iHub-Data, the technology translation hub at IIIT Hyderabad. The system has been running continuously since September at Chityal on the Vijayawada–Hyderabad highway, cross-checking trucks entering Telangana against a whitelist of nearly 40,000 vehicles approved by the Telangana Mineral Development Corporation (TGMDC).
"Typical license plates are actually easy to detect," said Dr. Veera Ganesh Yalla, CEO of iHub-Data. "But in India, especially with trucks, plates are often hand-painted, inconsistent, and highly variable." Commercial solutions, he noted, can cost tens of lakhs per camera in licensing and maintenance alone — a prohibitive expense for large-scale government deployment.
Rather than building from scratch, the iHub-Data team adapted an existing research prototype developed by Prof. Ravikiran Sarvadevabhatla's team at IIIT Hyderabad's Centre for Visual Information Technology. The core handwritten character recognition component was rebuilt and strengthened, then integrated as a plug-in into an open-source platform, allowing government agencies to adopt the technology without overhauling existing infrastructure.
The system has faced real-world complications, including poor lighting at night and festival garlands obscuring number plates, but has continued to improve as it processes live data. It was built by a team of fewer than five engineers using modern deep learning models including YOLO and RF-Detr.
The team is now working on adapting the technology for traffic violation detection involving two-wheelers, in partnership with the state police department.