Improving quantitative structure models with an Huxley protocol based filter

Abstract

Abstract Quantitative structure models (\texttt{QSMs}) are topological orderedcylinder models of trees whichdescribe the branching structure up to the tips. \\Methods:We present unpublished tree describing parameters which can bederived from a single Quantitative Structure Model (\texttt{QSM}). The parameters areused to build two \texttt{Radius} correctionfilters.\\Results:For validation we use QSMs produced from an open point clouddata set of tree clouds with the SimpleForest software. We comparethe QSM volume against the harvested reference data for 65 felledtrees. We also found QSM data of TreeQSM, a competitive andbroadly accepted QSM modeling tool. Our RMSE was less than 40% of the TreeQSM RMSE.\\For other error measures, the\texttt{r\textsuperscript{2}\textsubscript{adj.}}and the \texttt{CCC}, the relative improvement looked even better withreaching only 27 % and 21 % of the TreeQSM errors respectively.\\Conclusions:In forest ecology we should use the here presented pipelineto build accurate \texttt{QSMs} for reasons of: \\\texttt{Quality - }With the invention of the \texttt{QSM} \texttt{Radius} filter techniquesweimprove tree volume prediction capabilities utilizing \texttt{QSMs}.\\\texttt{Quantity - }More data can be collected with \texttt{QSMs} than with traditionalmethods. Here we use modelsbuild on more than ten thousand measurements.\\