Research And Implementation Of Ship Target Recognition System In The RS Image

Now the data of visible light remote sensing image used by ship target surveillance department is rapid increased, while the ship target recognition is also mainly depend on manual interpretation, resulting in the extremely low efficiency.It can not meet the requirements of monitoring the of marine vessels.The paper analyses the business requirements of ship target recognition system of large-format remote sensing image, and designs the framework and functions of the system. It designs four-story structure including data layer, data access layer, application service layer and application layer. It realizes the scheduling of data management, ship target recognition capabilities.This system uses multi-layer structure, and each functional module is a server software, making the system can be easily configured, on-demand cutting and easily upgraded. It uses rapid Auto Image registration and multi-level multi-feature Auto recognition algorithm. It realizes the service software to make the image registration and recognition can be parallel processing, easy to operate using the interface simplifies designed.The ship target recognition system of large-format remote sensing image of designed gives a powerful means to detection、confirm、track and monitor sea ship. However the image segmentation algorithm of the paper used is an accomplishment realized before, do not have good Effective, needing improved to the system. At the some time the system need to accumulate standardized data and reduce the time of image registration, giving a strong data support to the allocation of system resources.