ViT-Siamese Cascade Network for Transmission Image Deduplication.

With the large-scale use of various inspection methods such as drones, helicopters, and robots, the generated power inspection images have increased significantly, which has brought huge pressure on data storage and transmission. At the same time, with the rapid development of artificial intelligence technology in the electric power field, the cost of manual labeling required for model training has become a major pain point. This paper studies a transmission image deduplication technology based on ViT-Siamese cascade network, which reduces the amount of data and the cost of data annotation. This paper first investigates the research status of image similarity at home and abroad, and then studies the transmission image deduplication technology based on the ViT-Siamese cascade network, which greatly reduces the complexity of similarity calculation, and finally trains the model on the transmission image data set. Firstly, this paper investigates the research status of image similarity at domestic and international. Next, transmission image deduplication technology based on ViT-Siamese cascade network is studied, which greatly reduces the complexity of similarity calculation. And finally, the AI model is trained on transmission image dataset, and effectiveness and feasibility of the technology in transmission scene processing are verified by experiments.