Power equipment image enhancement processing based on YOLO-v8 target detection model under MSRCR algorithm

Abstract

Abstract With the rapid development of the power industry, higher requirements have been put forward for real-time monitoring and fault identification of power equipment. However, images of power equipment in actual scenes are often affected by problems such as uneven illumination and color distortion, leading to a decrease in the performance of the target detection model. Hence, this paper suggests merging the Multi-Scale Retinex with Color Restoration (MSRCR) algorithm with the YOLO-v8 target detection model to enhance the visual quality of power equipment images and boost the accuracy and efficiency of target detection. Initially, the MSRCR algorithm enhances image brightness, contrast, and color restoration and preserves edge and detail features. Subsequently, the paper explores the architecture of YOLO-v8, incorporating the SE (Squeeze-and-Excitation) attention mechanism. This mechanism dynamically adjusts channel weights to optimize feature processing in input data. The final experimental results show that using the MSRCR algorithm to enhance the data and combining it with the SE attention mechanism have improved by about 3.2% compared to the original YOLO-v8 model. In comparative experiments with other algorithms, the method proposed in this article achieved an accuracy of 94.3% and a recall rate of 92.6%, which are both higher than other models. By enhancing power equipment images with the MSRCR algorithm, the YOLO-v8 model has significantly improved both target detection accuracy and recall rate. In summary, the MSRCR power equipment image enhancement processing method proposed in this article based on the YOLO-v8 target detection model can effectively improve the visual quality of power equipment images and improve the accuracy and efficiency of target detection.