Appendix B 



Iceberg / Ship Target Discrimination 

 with Side-Looking Airborne Radar 



LTJG N. B. Thayer, USCGR 

 CDR N. C. Edwards, USCG 



Introduction 



Since 1983, the 

 International Ice Patrol (IIP) has 

 been using a Motorola AN/APS- 

 135 Side-Looking Airtwrne MuHi- 

 Mission Radar (SLAMMR) as its 

 primary method of iceberg 

 reconnaissance in the North 

 Atlantic. The ability to detect 

 icebergs with a side-looking 

 airtxime radar (SI_AR) in poorer 

 zero visibility, plus the ability to 

 search larger areas, has resulted 

 in a significant increase in the 

 number of icebergs tracked by 

 IIP. 



Because SLAR can be 

 used with the sea surface 

 obscured by clouds, IIP 

 frequently conducts 

 reconnaissance flights when 

 visual confirmation of SLAR 

 targets is not possible. Without 

 visual confirmation, distinguishing 

 between icebergs and vessels is 

 sometimes difficult. 



Without visible cues on the 

 SLAR film (target movement, 

 wakes, brash, radar shadows, 

 strength of return) which improve 

 target identification, it is difficult to 

 distinguish between targets with 

 similar radar return, e.g., small 

 teebergs and vessels. IIP has 

 planned its search legs and the 

 track spacing equal to one-half 



the total SLAR sweep width (i.e., 

 25 nm). This type of search plan 

 gives 200% coverage between 

 parallel legs and provides two 

 views of each target within the 

 search area. Despite these 

 efforts to maximize cues, it is still 

 sometimes difficult to distinguish 

 vessels from small and medium 

 icebergs . For example , fishing 

 vessels often drift or move stowly, 

 producing no wake and showing 

 little or no nrxjvement between 

 looks. In addition, the search legs 

 going to and from the search area 

 as well as the outlying legs of the 

 search itself do not afford double 

 SLAR coverage. As a result, 

 approximately 35% of the search 

 area is seen only once on SLAR, 

 eliminating the chance to detect 

 movement and decreasing the 

 probability of picking up other 

 cues from SLAR images. 



This study measures the 

 en'or rate in SLAR target 

 identifk»tk>n, using single looks 

 at individual k^eberg and ship 

 targets without visual cues. 



Methods 



To conduct this study, it 

 was necessary to find a source of 

 SLAR targets with visual 

 confirmation. The best source of 

 targets with positive identif Nation 

 of both target size and type was 

 the BERGSEARCH '84 (Rossiter, 

 et al. , 1 984) data and the 1 985 

 SLAR experiment conducted by 

 IIP and the Coast Guard Research 

 and Devetopoment Center 

 (Robe, era/., 1985). These two 

 sources provided SLAR film from 

 7 days of IIP operattons with 

 shipboard ground truth data, 1 60 

 ship and k^berg targets in all. All 

 of the film used in this study was 

 collected at an altitude of 8,000 

 feet on the 50 km SLAR range 

 scale, standard conditions during 

 IIP iceberg reconnaissance. 



The films were duplicated 

 and the duplicate films were 

 examined for suitable targets for 

 the study. All targets without 

 obvbus cues were used. 

 Although targets were not 

 selected for ambiguity, all of 

 those used were quite 

 ambiguous, since they were all 

 single targets without 

 accompanying visual cues. With 

 the limited number of vessels and 

 k;et)ergs involved in the two 

 source experiments, some 

 targets were used more than 

 once, but separate SLAR passes 



53 



