Research on Electric Equipment Identification System based on Computer Vision

In order to enhance the intelligence of the power system, the functionality of the equipment is assessed in real time through the transmission of images and videos from a remote location. Although the integration of image and video data into the power system marks a notable technological advance, the existing monitoring system is only capable of transmitting a considerable volume of data, which cannot be analysed in an intelligent manner and still necessitates manual processing. This approach is inefficient and inaccurate. Prolonged manual monitoring is detrimental to the physical and mental wellbeing of personnel, which constrains the potential for further enhancement of the intelligence level of the power system. The initial stage of the process is image preprocessing, which is employed to eliminate extraneous noise and background interference, thus ensuring the integrity of the input data. Subsequently, a convolutional neural network (CNN) is utilised to extract the underlying, intricate features from the image. By means of multi-layer convolution and pooling operations, the model is able to automatically identify the key features of the device, including shape, texture and structure. These features are then passed on to the classifier for precise identification. This method is highly automated, capable of efficiently processing complex image data, reducing the potential for human error, improving the accuracy and efficiency of electrical device identification, and providing the flexibility to adapt to different application needs. The results of the experimental analysis demonstrate that the system exhibits high efficiency and accuracy in practical applications, which is beneficial for enhancing the automation level of the power system.