[The effect of adjusting the ANN's parameters on the identifying of biology spectrum].

ABSTRACT

A method to identify the visible spectrum of micro areas on the biological surface with artificial neural net (BP-ANN) was introduced in this paper. The visible spectra (from 500 nm to 730 nm) of the micro areas with some rotten or scars on the surface of the apples were measured with fiber sensor spectrometer. A kind of ANN with a single hidden layer was created to identify the characters on the surface automatically. The effects of different ranges of output, different training functions, different number of single hidden layers, and different noise levels on the ANN were also studied.