Research On Key Technologies Of The Real-time Detection Of Dynamic Ship Draft And Data Processing

Inland navigation,which has low cost,large volume,less pollution and other advantages,plays an important role in China's economic development.However ultra-draft problems in inland shipping are more serious with the increasing shipping traffic recent years,which seriously affects the safety and efficiency of navigation. The law-executors from maritime check ship's drft mainly relay on visual inspection of ship's waterline at present,this mothod wastes time,consumes energy and affects the normal navigation.There is no effective technical means to detect ship's draft automatically under the premise of avoiding boarding ships currently.According to the requirements of detection on navigation ship draft, the real-time detection and warning system of dynamic ship draft based on ultrasonic sensors'array is researched and developed.Without affecting the normal navigation of the ship,it realizes fast and real-time detection on ship draft under the situation of disembarka-tion.The system currently has been put into trial operation in the Three Gorges Project of Yangtze river.The automatic synchronous rasing and lowering control subsystem is responsible for closed-loop control of speed and position of the door,where ultrasonic sensor array is installed, to ensure the detection door safely and smoothly work in the right underwater place.The data acquisition and processing subsystem collects ship draft's data with the help of ultrasonic sensor array fixed on the door, then the subsystem process the cllected data to exclude abnormal data and split draft data of navigation ships in parallel.Finally it sends the data of ship draft to the integrated information management subsystem,which compare the measurement with the maximum draft the current channel allows and advance warning of ultra-draft ships.The point cloud data of ship draft is obtained using the system.The abnormal data in the point cloud is grouped into two categories on the basis of distribution features and field interference source.The median filtering algorithm about the point cloud data is proposed to eliminate the interference of the first abnormal data when ultrasonic sensors are used to detect the bottom of ship. Because the second exception data only appears in the Ship Stern, the Least-Square linear regression and prediction algorithm is put forward to eliminate interference caused by the propeller wake.The two algorithms get verification by the field data. Field tests showed that the two algorithms were practical and effective.After preprocessing on collecting data to eliminate all kinds of noise, this paper makes a detailed analysis of the draft data division about the parallel navigation vessel,proposes a segmentation algorithm on point cloud data of ship draft based on the width of ship,and analysis key technologies of identification on ship into or out of the detection door and ship's width. The algorithm was validated and simulated using field data.Finally this paper gives field measurement results,determine systematic precision,and analysis measurement uncertainty.The results showed that the system me-asurement accuracy was better than 0.1m,which met design requirements.The algorithms used in the system were effective and practical.