Research on distribution infrared image defect system based on improved Mask-RCNN

With the acceleration of power grid construction, the scale of electricity continues to expand, and the mileage of distribution lines is also constantly increasing. The power lines are exposed to the natural environment for a long time and require regular inspections to eliminate hidden dangers, defect detection of distribution equipment has become a key link in ensuring the stable operation of the power system. With the deep application of drones in the field of power grid and the development of artificial intelligence technology, using drones to inspect and collect line images has gradually become mainstream. Infrared imaging technology is playing an increasingly important role in identifying defects in distribution networks. However, existing image detection mainly focuses on visible light images, making it difficult to penetrate the surface of objects and detect internal defects, merely detecting through visible light images may have adverse effects on the safe and stable operation of the power system. This article proposes an improved Mask rcnn defect detection algorithm, which is combined with CBAM attention mechanism to further enhance the model's attention to insulators, fuses, and isolation switches, achieving more accurate detection and segmentation. The experimental results show that compared with traditional segmentation algorithms, the improved algorithm has higher recognition accuracy and better segmentation effect, which can better achieve infrared image temperature difference and defect detection.