Enhanced Threshold Segmentation for Accurate Digital Meter Reading Recognition using an Improved Northern Goshawk Optimization Algorithm

Abstract In the realm of industrial automation, digital meters are increasingly prevalent, yet manual meter reading remains inefficient and error-prone. This paper presents a novel computational method aimed at enhancing the precision and speed of digital meter reading recognition. The proposed approach commences with image preprocessing, involving grayscale conversion, histogram equalization to amplify meter digital information, and median filtering to eliminate noise. To tackle the challenge of image binarization in complex scenarios, we introduce an Improved Northern Goshawk Optimization (INGO) algorithm, which leverages Tent chaotic mapping and adaptive Cauchy variation to refine the OTSU threshold segmentation process. By expanding the search range and mitigating local optimal traps, INGO ensures the precise determination of the optimal binarization threshold. Here, we show that the proposed method achieves a substantial improvement in Peak Signal-to-Noise Ratio (PSNR), with an average increase of 6.35% compared to traditional OTSU and NGO-optimized methods, resulting in clearer and more accurate digital meter readings. This research underscores the potential of intelligent algorithms in automating industrial meter reading, contributing to enhanced efficiency and reliability in energy management systems.