engleski

Neural Network for Classification based on Nondestructive Testing

Proposal for classification of wood members using neural network is presented here. Using nondestructive testing dynamic parameters are obtained for more then 100 wood members of oak timber species. Dynamic parameters (resonant frequency, decrement of vibrational decay and propagation speed of ultrasound waves), moisture, density, boundary conditions and geometry of samples were input while strength values and MOE were output variables. As a result run-time option of NN is obtained as main part of classification system.

neural network; wood grading; nondestructive testing

nije evidentirano