Title : 
A ship detection system for RADARSAT-2 dual-pol multi-look imagery implemented in the ADSS
         
        
            Author : 
Crisp, David James
         
        
            Author_Institution : 
Intell., Surveillance & Reconnaissance Div., Defence Sci. & Technol. Organ., Edinburgh, SA, Australia
         
        
        
        
        
        
            Abstract : 
A ship detection system for use with RADARSAT-2 dual-polarisation multi-look imagery is described. The system is built using the Defence Science & Technology Organisation´s (DSTO) automated target detection software system known as the Analysts´ Detection Support System (ADSS). The ADSS is flexible and allows the combination of many different modules in parallel as well as sequential pipelines. This paper discusses a particular system implemented in the ADSS to detect ships in RADARSAT-2 ScanSAR Narrow dual-polarisation imagery. The imagery has been well ground-truthed and detection performance is presented in the form of ROC curves. A particular aspect of the problem that is considered in detail is how best to combine the HH and HV polarisation channels in the pre-screening stage.
         
        
            Keywords : 
marine radar; radar detection; radar imaging; ships; ADSS; DSTO; Defence Science & Technology Organisation; HH polarisation channels; HV polarisation channels; RADARSAT-2 ScanSAR narrow dual-polarisation imagery; RADARSAT-2 dual-pol multilook imagery; RADARSAT-2 dual-polarisation multilook imagery; ROC curves; analysts detection support system; automated target detection software system; detection performance; pre-screening stage; sequential pipelines; ship detection system; Detection algorithms; Detectors; Marine vehicles; Noise; Oceans; Synthetic aperture radar; Thermal noise;
         
        
        
        
            Conference_Titel : 
Radar (Radar), 2013 International Conference on
         
        
            Conference_Location : 
Adelaide, SA
         
        
            Print_ISBN : 
978-1-4673-5177-5
         
        
        
            DOI : 
10.1109/RADAR.2013.6652006