Reading and understanding house numbers for delivery robots using the ”SVHN Dataset”

Abstract

Detecting street house numbers in complex environments is a challenging robotics and computer vision task that could be valuable in enhancing the accuracy of delivery robots' localisation. The development of this technology also has positive implications for address parsing and postal services. This project focuses on building a robust and efficient system that deals with the complexities associated with detecting house numbers in street scenes. The models in this system are trained on Stanford University's SVHN (Street View House Numbers) dataset. By fine-tuning the YOLO's (You Only Look Once) nano model results with an effective detection range from 1.02 meters to 4.5. The optimum allowance for angle of tilt was ±15°. The inference resolution was obtained to be 2160 * 1620 with inference delay of 35 milliseconds.