Towards Image-Based Automatic Meter Reading in Unconstrained Scenarios: A Robust and Efficient Approach

Existing approaches for image-based Automatic Meter Reading (AMR) have been evaluated on images captured in well-controlled scenarios. However, real-world meter reading presents unconstrained scenarios that are way more challenging due to dirt, various lighting conditions, scale variations, in-plane and out-of-plane rotations, among other factors. In this work, we present an end-to-end approach AMR focusing Our main contribution is the insertion of a new stage pipeline, called corner detection counter classification, which enables region be rectified -- as well rejection illegible/faulty meters prior recognition stage. We also introduce publicly available dataset, Copel-AMR, contains 12,500 acquired field by service company's employees themselves, including 2,500 faulty or cases where illegible occlusions. Experimental evaluation demonstrates proposed system, has three networks operating cascaded mode, outperforms all baselines terms rate while still being quite efficient. Moreover, very few errors tolerated applications, show our system achieves impressive rates (i.e., > 99%) when rejecting readings made with lower confidence values.