Analysis Methodology Research On Identification Characteristics Of Ship Target’s Remote Sensing Images

As an important maritime reconnaissance target, ship target has high movementspeed, and complex movement environment characteristics. In the processing of remotesensing images，the technical capabilities of ship target detection and identificationdirectly affect the extracting of intelligence from original satellite data. A reasonableand effective method need to be choose to realize the accurate interpretation of shiptarget recognition.In this paper, we took ship targets as our main study object. By analyzing kinds ofcharacteristics of various ship targets，we proposed some important characteristics inship identification by using remote sensing images，and constructed a simulation systemfor target identification through them.The main results are as follows:1. Analysis of typical target characteristic parameters from remote sensingimages.Based on analyzing typical ship target characteristics item by item, we illustratedsome kinds of important characteristics in three ways like space distribution,configuration and motion features, and their related parameters, such as geometryparameters interpreted from ship images, activity parameters of ship targets andgeographic information, etc.2. Study on the analysis methodology on identification ship target characteristicsWe analyzed the methodology on activity parameters of ship targets andgeographic information, especially to the typical target characteristic and geographicparameters.3. Study on simulation of ship target interpretation in remote sensing imagesBy analyzing and comparison the extracted characteristic parameters of ship targetsand ship formation, we established the simulation models of typical ship interpretingprocesses, to explain the implementation methods of ship target identification.4. The engineering practice of automatic ship identification from remote sensingimagesCombining studies in actual projects, we developed a knowledge-base for ship identification by using feature analyzing theories proposed above.