Image-based Automatic Dial Meter Reading in Unconstrained Scenarios

The replacement of analog meters with smart is costly, laborious, and far from complete in developing countries. Energy Company Parana (Copel) (Brazil) performs more than 4 million meter readings (almost entirely non-smart devices) per month, we estimate that 850 thousand them are dial meters. Therefore, an image-based automatic reading system can reduce human errors, create a proof reading, enable the customers to perform themselves through mobile application. We propose novel approaches for Automatic Dial Meter Reading (ADMR) introduce new dataset ADMR unconstrained scenarios, called UFPR-ADMR-v2. Our best-performing method combines YOLOv4 regression approach (AngReg), explores several postprocessing techniques. Compared previous works, it decreased Mean Absolute Error (MAE) 1,343 129 achieved recognition rate (MRR) 98.90% -- error tolerance 1 Kilowatt-hour (kWh).