The pointer-type instrument recognition method based on OpenCV

Pointer-type instruments remain widely used in industrial environments due to their cost-efficiency and reliability. However, manual reading of such instruments is labor-intensive, error-prone, and unsuitable for real-time monitoring. This paper presents a lightweight and interpretable method for automatic pointer-type instrument recognition using classical image processing techniques provided by OpenCV. The proposed approach includes image preprocessing, circular dial detection via the Hough Circle Transform, scale calibration, pointer extraction using probabilistic Hough Line detection, and value computation through geometric analysis. Experimental results on standard analog gauge images demonstrate the methods accuracy in identifying gauge boundaries, calibrating tick marks, and determining the pointer's angular position. The approach enables flexible value mapping and delivers reliable readings under standard lighting conditions. Compared to deep learning methods, this solution offers better computational efficiency and easier integration into resource-constrained systems. It is particularly suitable for industrial automation and retrofitting of legacy analog devices. Future work will focus on improving robustness under complex backgrounds and dynamic conditions.