Designand Realizationof Integrated Ship Identification System Basedon Remote Sensing Imageand AIS Data

Since the 21st century, marine resources have been received more and more attention. Shipping, biological and mineral resources and other resources to a country's economy is playing an increasingly important role. In order to protect China's maritime interests from infringement, and to achieve the goal of the search and rescue of the sea,climate and environment monitoring, remote sensing satellite automatic detection identification must be the general trend. Nowadays, remote sensing satellites have been able to realize the monitoring of the whole day,all-weather and large area of the ocean, and most of the ships have installed the ship automatic identification system (AIS). This paper discusses the design and implementation of a comprehensive ship identification system based on remote sensing and AIS data, to realize the full automation from remote sensing image uploading, remote sensing image management, remote sensing image detection and AIS data matching.In this paper, we first study the domestic and foreign research results in this field, then formulate requirements and formulate documents according to the actual needs and the mature technologies that have been realized, after that, we discuss two key algorithms — CFAR algorithm and AIS matching algorithm, and the key technologies such as Canny edge detection, land mask based on texture segmentation, ship wake type and feature calculation ship speed are applied to the system. The main work flow of the system is image management, image preprocessing,automatic detection, image enhancement, feature extraction, AIS assistant recognition, adding to the topic and exporting .shp file. Finally, the remote sensing image of the Feiyun River near the mouth of the East China Sea is actually applied to the system. The number of ships that can be detected automatically is 765, most of which are 100 meters in length and 60 meters in width. After matching with the AIS data, there are 509 matching specific ships, matching rate of about 70%. The remaining unmatched shipscan be selected by automatic or manual identification of the ship types after feature extraction. After analyzing the matching rate and accuracy, we can see that our system realizes the integrated ship detection and recognition based on remote sensing image and AIS data,and can be applied practically.