ageometricsimilarit ymeasurementmethodanda pp licationstolinearfeature

geometricsimilarit ymeasurementoflinearfeaturesistheke ytomatchin gma pdata,fu-sion,andclusterin g.thisp a p erp resentsanewmethodforg eometricsimilarit ymeasurementofdi g it-alma plinearfeaturesbasedonthetraditionaldiscretefréchetdistance.wederivedaformulaforcomp utin gdiscretecurvesfréchetdistancebasedonrecurrenceandp resentscurvessimilarit ymeas-urementmodelbasedonavera g efréchetdistance.theavera g efréchetdistanceisobtainedb yreco g-nizin gandcomp utin gminimalp athbetweenp ointsintwocurves，whichcanavoidbi gg isherroroftra-ditionalfréchetdistance.meanwhile,thisp a p erdemonstratesthattheavera g efréchetdistancedeliv-ershi g heraccurac y,theoreticall y.inordertomeasurethep artialandoverallsimilarit ybetweentwocurves，wealsop resentap artial-overalldiscretefréchetdistancebasedontheavera g efréchetdis-tance.finall y,thisfréchetdistancewasa pp liedtomatchin gma pdata,fusionandcontourinter p ola-tion.ex p erimentswerep erformedtoshowthefeasibilit yandsu p eriorit yofthemethod.