Changzhou Zhonglou Uses an AI Smart Eye to Strengthen Fuel Station Tax Supervision

A China Tax News article republished by Jiangsu tax authorities on July 22 says Zhonglou district in Changzhou is using a real-time supervision model built on lightweight video transmission and AI recognition to address hidden fuel sales and off-book circulation in the refined-oil sector. Working with police, tax officials connected cameras to the existing city surveillance network to capture vehicle type, license plate, and entry and exit times at stations, then used backend analysis to estimate fueling volume and compare it with declared revenue. The district now covers 32 fuel stations and says six private operators have voluntarily paid 7.1 million yuan in back taxes. The report describes one case where captured traffic data showed around 450 vehicles a day at a station whose declared revenue was far below peers, allowing officials to confirm underreporting with image evidence and model-based estimates. The article says the system now forms a full loop of data capture, intelligent estimation, warning, risk response, and correction, while also sharing data with commerce and emergency-management departments, showing AI and visual recognition moving into practical tax administration and cross-agency governance.