A High-Precision Automatic Pointer Meter Reading System in Low-Light Environment.

Xuang Wu,Xiaobo Shi,Yongchao Jiang,Jun Gong

doi:10.3390/s21144891

Abstract

At present, pointer meters are still widely used because of their mechanical stability and electromagnetic immunity, and it is the main trend to use a computer vision-based automatic reading system to replace inefficient manual inspection. Many correction and recognition algorithms have been proposed for the problems of skew, distortion, and uneven illumination in the field-collected meter images. However, the current algorithms generally suffer from poor robustness, enormous training cost, inadequate compensation correction, and poor reading accuracy. This paper first designs a meter image skew-correction algorithm based on binary mask and improved Mask-RCNN for different types of pointer meters, which achieves high accuracy ellipse fitting and reduces the training cost by transfer learning. Furthermore, the low-light enhancement fusion algorithm based on improved Retinex and Fast Adaptive Bilateral Filtering (RBF) is proposed. Finally, the improved ResNet101 is proposed to extract needle features and perform directional regression to achieve fast and high-accuracy readings. The experimental results show that the proposed system in this paper has higher efficiency and better robustness in the image correction process in a complex environment and higher accuracy in the meter reading process.

Highlights

Pointer meters are widely used in electric power systems, petrochemicals, military industries, and other fields
The performance comparison results of Mask-RCNN networks obtained by two training methods with and without transfer learning are shown in Table 2, in which the feature extraction networks were used with ResNet50-FPN and ResNet101-FPN respectively, and the target detection and instance segmentation of the models trained in the four cases were performed on real datasets under the same conditions as other experiments
This paper proposed an automatic reading system based on computer vision and deep learning for different types of pointer meters in the low-light environment

Summary

Introduction

Pointer meters are widely used in electric power systems, petrochemicals, military industries, and other fields. Ma [5] proposed a random sample consistency method to overcome background image interference. These methods run faster, but the traditional target detection and segmentation algorithms are less accurate and less robust. Salomon et al [9] detected the bounding box of the dial by YOLO-based methods, and Faster-RCNN extracted the needle and calculated the readings. These methods can be effective, but the detection success rate and reading accuracy are significantly degraded under disturbances such as tilt and uneven illumination, and preprocessing is needed for improvement

Methods

Results

Conclusion