›› 2015, Vol. 35 ›› Issue (6): 57-.doi: 10.3780/j.issn.1000-758X.2015.06.007

Previous Articles     Next Articles

DetectingandTrackingMethodforSpaceTargetGraspingBasedonContourFeaturesDetection

KANG  Ru-Fei, WANG  Shi-Cheng, LIU  Zhi-Guo, YANG  Dong-Fang   

  1. (DepartmentofControlEngineering,TheSecondArtilleryEngineeringUniversity,Xi′an710025)
  • Received:2015-06-02 Revised:2015-08-04 Online:2015-12-25 Published:2015-12-25

Abstract: Duetothecomplexbackground,changesoftheilluminationandrelativemotions,betterperformancesineffectivenessandrobustnessareneededformovingtargetgraspingtasks.Anewdetectingandtrackingmethodbasedondynamiccontourfeatureswasproposed.Thegraygradientmethodandtheactivecontourmodelwereintegratedinthismethod.Atfirst,onthebasisofthetraditionalgraygradientmethod,animprovedgraygradientmethodwasproposed,whichincludedshapematching,areascreeningandadaptiveadjustmentofgraythreshold.Accordingtothedetectingresultsoftargetcontour,thetimedomainconstraintsbetweenimagesequenceswerealsoinvolved.ThedetectingresultswereaddedintheGVFsnakemethodtoimprovethemeasuringaccuracyandeffectiveness.Intheend,consideringtherealtimedemandofthegraspingtasks,thetrackingmethodswasparallelwithCUDAcalculationarchitectureandanexperimentwasconductedwithabundleofsequentialimages.Theresultshowsthatthetrackingratecanreach10framespersecond,improvingthetrackingspeedwithoutsacrificingtheaccuracy.

Key words: Contourdetection, Spacetargetgrasping, Timedomainconstraints, Spacemission