Intelli entsystemforthedetection ofburied obectsbased on hu-Stutures usinmetaldetector
HadjHamma Tadjine, M. Reuter, S. Hameit · 2006
Hybridclassification structures ofburied object in different soils usingtimedomainmetaldetectors are presented inthis paper. Thisapproach consists ofself organising memoriesand Backpropagation nets combinedwiththe inter neuralWVHUStructures, whereby special pre Processing methods like D Sand fitting methods are usedtoextract thefeature specific items. Inthis work, an enduser supporting systemfor humanitarian deminingtasksto semiautomatically classify signals isalsoproposed. Inaddition these structures enable tointegrate several aspects ofa scene intoa mega conceptcorresponding totheChunking processofthebrain. Thecategorization ofconcepts will hereby bedonebyactive whustructures whichbecome dominant since they sum up theactivities oftheinvolved neurons instead ofconsidering one andonly one high active winnerneuron. Usingtheproposed method, results confirm, thatthepresence ofburied objects in landminecouldbe localized and some ofthem classified.