Rapid urbanization prompts the need for smart city innovations to ensure predictive maintenance of infrastructure in place. We propose a real-time automated inspection system for urban asset monitoring powered by state-of-the-art deep learning algorithms, on-the-ground video data acquisition (from a car-mounted camera and GPS module), and data visualization. Our system swiftly identifies and locates critical urban assets like traffic signs, garbage bins, and trees along city routes and automatically geo-tags them, overlaying the real-time geolocation data onto a digital map. Experiments within a study area of 10 major routes in Guwahati recorded the detection model to achieve a precision of 0.845, a recall of 0.803, and a mAP of 0.808, demonstrating positive performance. Our proposed system is designed to assist in efficient infrastructure mapping and monitoring by local government authorities over a period. Our prototype and workflow are cost-effective based on open-source approach and can be adapted to other cities.