The Study On Detection And Analysis Of Marine Traffic System Based On Remote Sensing Image

In recent years,accompanied by the rapid development of China's shipping, navigable environment of coastal waters has changed.The current investigation methods such as VTS,AIS have been restricted by distance.Not all ships can be detected,surveyed and analyzed.It is necessary to develop a new ship traffic detection system to understand the traffic situation,improve the scientific decisions on safety planning and management,and ensure the safety of marine traffic by effective ship detection.The development of satellite technology has provided us a method which can detect in long-range,all weather and economically.This thesis discusses applying this technology in marine traffic survey to detect and analyze multi-element in large scale.At first,researching the relevant information at home and abroad,analyzing the possibility of creating marine traffic system based on remote sensing satellite, discussing the key technologies to detect ship which includes gravitational field,CFAR detection,PNN and adaptive algorithm.The process is:import the satellite image to system;image land masking;detect ship and extract ship from the image by ship detection algorithm;manual intervention,removing false alarms and adding missed targets,then save the result to the database.Based on the database,Statistic and analyze the data to achieve the statistical analysis of ship traffic,ship density and ship sizeAt last,taking the ERS-2 image as an example,processing and analyzing the images of waters near Yangtze estuary,extracting the ship information from the image. Analyzing the accuracy with AIS information.Selecting nine images,analyze the detection result of each image,knowing the distribution of ship density,traffic,ship size by analyzing the database.The results showed that detecting ship at sea using SAR imagery is entirely feasible,the analysis result is reliable.