METHOD FOR INCREASING AN ACCURACY OF DETECTING DECAY BY THE ARBOTOM® 3-D TREE TOMOGRAPH ON PICEA ABIES (L.) H. KARST TREES DAMAGED BY HETEROBASIDION ANNOSUM (FR.) BREF

Abstract

ARBOTOM® 3-D tree tomograph is a non-destructive decay-sensing instrument based on measurements in standing trees. The accuracy of ARBOTOM® 3-D tree tomograph in detecting decay was evaluated for 703 standing trees in 27 Norway spruce (Picea abies) stands. Trees were measured at three different heights, 0 m, 1 m, and 2 m above the ground. Sections were later inspected for the presence of decay by increment borings and wood croscuts. A system of stochastic differential equations and multivariate normal copula density function were used to develop decay detecting rule. In the present study, the Vasicek stochastic process was employed to analyze propagation velocity data of the stress wave that diffuses through the wood of damaged and healthy tree stems in standing trees. Ornstein–Uhlenbeck family stochastic processes offer a possibility of capturing important distributional differences between damaged and healthy trees stress wave velocity datasets and make the model flexible to internal decay of a tree stem. The study reveals that the Vasicek type stochastic process is a possible candidate for statistical classification of waveforms. After the conditional probability density functions to ratio of difference maximal and minimal propagation velocities to minimal and maximal propagation velocity were established, a bivariate normal copula distribution model was constructed. The results are implemented in the symbolic computational language MAPLE.

Keywords: Internal decay, Norway spruce, non-destructive sensing, stochastic differential equation.

Article DOI: http://doi.org/10.15544/RD.2015.057