Real-time Detection System For Underwater Disaster-causing Organisms In Coastal Nuclear Power Plant

In recent years,there have been phenomena of marine-induced biological outbreaks in coastal nuclear power plants in various countries.After the outbreak of marine disaster-causing organisms,a large number of water inflow into the nuclear power station will threaten the safety of the nuclear power plant's cooling system,and even cause the nuclear power plant to be forced to stop,causing economic losses to the nuclear power plant,threatening the surrounding marine ecological environment and the safety of human society.In order to pre-warn the outbreak of disaster-causing organisms such as prawn,jellyfish and sea cucumber,this paper designed a real-time detection and identification system for underwater disaster-causing organisms in coastal nuclear power plants.The system takes real-time detection and identification of the prawn as an example,including hardware and software implementation.When the system is running,the underwater camera acquires underwater video information and transmits it to the control computer through the network data line in real time.The video data can be displayed in real time and utilize the pre-trained convolutional neural network-Faster R-CNN for the marine life(shrimp)for detection and identification,real-time detection of underwater disaster-causing organisms in coastal nuclear power plants,and the ability to issue early warnings to nuclear power plant personnel.The underwater camera used in the system is self-developed by this paper.The software part analyzes and compares various deep learning target detection algorithms according to the characteristics of shrimps,and selects the target of the Faster R-CNN based on region selection.Identification algorithm.