›› 2012, Vol. 32 ›› Issue (2): 35-42.doi: 10.3780/j.issn.1000-758X.2012.02.006

Previous Articles     Next Articles

Space Target Recognition Based on Improved Kernel FCM

 WANG  Xiao-Xue, YANG  Yong-Sheng, JING  Zhong-Liang   

  1. (SchoolofAeronauticsandAstronautics,ShanghaiJiaoTongUniversity,Shanghai200240)
  • Received:2011-10-12 Revised:2012-04-25 Online:2012-04-25 Published:2012-04-25

Abstract: A recognitionsystemwaspresentedtoaccomplishspacetargetclassification.Firstly,Imagefeaturevectorswereextractedaccordingtoinvariantregionmoments,shapeandimagedescriptorsofspaceobjects.Andthen,clusteringalgorithmisappliedtotheclassifyspacetarget.AnimprovedkernelclusteringalgorithmbasedonVoronoidistancewasproposedwhichhadacrispermembershipfunctionandwasrobustfornoiseandoutliers.Experimentsshowthattheimprovedkernelfuzzyclusteringalgorithmismoreaccurateandvalidthanthatoftheconventionalmethods.

Key words: Image recognition, Kernel cluster method, Feature extraction, Space target