Research On Ship Identification System Based On Maritime Monitoring Platform

Water traffic is an important part of the Great-Traffic in our country.China has a vast water area and a large number of ports,so the demand for a ship identification system that can be used for maritime surveillance is extremely strong.The ship identification system can effectively provide assistance for surface safety,provide supervision evidence for water traffic accidents,greatly enhance the efficiency of water security implementation,and protect the hydrological environment to a certain extent.Compared with the background of the airborne system,the standardization of the terrestrial system,the waterway is dynamically changed due to the horizontal plane,and the surface environment changes greatly due to the reflection of light in the water surface environment.Therefore,the target recognition method on the traditional water surface is not effective.In order to solve this problem,this paper combines the traditional image processing technology with the emerging deep neural network method to target and segment the water surface,and then lay a foundation for future applications such as water surface video image processing.This paper combs the application of traditional technology methods on the water surface,and studies the application of deep learning technology in ship target image recognition.The main work is as follows:1.Investigate the traditional video capture technology,and carry out experiments to compare and select the video capture technology suitable for the water surface environment,so that the system does not have to run the target detection program at all times,but captures the effective ship image before proceeding to the target.Detection effectively speeds up the operating efficiency and time of the system.2.A variety of different feature matching methods have been studied and compared,and the matching procedure is high and the matching speed is fast,which can effectively improve the database retrieval efficiency.3.Learning and studying the target detection depth neural network SSD,construct its own ship database,and train the model ship target detection model.4.A simple ship target identification system has been implemented,which is beneficial for the maritime supervision authorities to keep track of ships and effectively improves the monitoring of the maritime sector by the maritime sector.