A New Method for Feature Enhancement of Grid Equipment Images

With the continuous development of the state grid, power transmission equipment is also increasing, and how to efficiently inspect and maintain power transmission equipment has become a key concern in the industry. In order to reduce the consumption of manpower in the inspection process, the current method is mainly to send drones to take patrol photos and then use deep learning to identify faulty equipment. However, in the actual process, power transmission equipment is generally built in areas with more natural vegetation, so the pictures obtained from drones have problems such as unclear targets and too many irrelevant areas, which become the main factors affecting the effect of abnormality identification. Based on this, this paper proposes a method of image feature enhancement from the perspective of image processing, which can effectively improve the feature representation in the original image. Experiments are conducted on three datasets constructed by ourselves, and it can be seen that the image processing method proposed in this paper achieves good experimental results, which can effectively improve the detection effect of deep learning models on faulty equipment photos.