Heavy water reactor unit refueling identification method based on target detection algorithm

The invention belongs to the technical field of heavy water reactor nuclear power station refueling, and particularly relates to a heavy water reactor unit refueling identification method based on a target detection algorithm. Comprising the following steps: step 1, training a refueling target identification model; and step 2, a refueling target identification process. The method has the beneficial effects that the deep learning algorithm is utilized to perform feature extraction on the refueling form trend, the refueling power reduction interval of the nuclear power plant is accurately identified, the refueling working condition result generated in the operation process of the power plant is recorded, convenience is provided for engineers to count the occurrence frequency of the refueling working condition, and the working efficiency is improved. And meanwhile, the power plant can conveniently and effectively control and track the refueling working condition, and the operation state of the power plant is eff