Research on Water Leakage Fault Detection of Air Conditioning and Refrigeration Plant Room Based on Computer Vision Technology

Air conditioning and refrigeration plant rooms have always been at risk of equipment damage and energy wastage due to water leakage faults. At present, only manual methods of water leakage fault detection, especially in the dark light environment, are obviously unable to avoid the potential risks. In this paper, we propose an improved algorithm YOLO-SEE (SENet+EIOU) based on computer vision technology to identify water leakage faults in potential leakage areas of air conditioning and refrigeration plant rooms under different lighting environments to reduce the risk of subsequent damage caused by faults. The self-made water leakage dataset of refrigeration plant rooms has 2485 images, the attention mechanism SENet is inserted into the backbone part of the model to improve the feature extraction ability of the model, the loss function EIOU is improved to increase the convergence speed of the model and provide better location results, and the performance of the algorithm is validated by comparing the algorithms such as SSD, YOLOv5s and etc. on the test set. The experimental results show that the improved YOLO-SEE algorithm achieves an average precision of 88.9% for water leakage detection in air-conditioning and refrigeration plant rooms, and still ensures the accuracy and real-time performance under different lighting environments, which can meet the needs of real-time monitoring of water leakage faults in actual refrigeration plant rooms.