Study on the Biological Surface Using the Reflected Spectrum and Artificial Neural Network Technology

Abstract

This paper gave a method to identify the visible spectrum of micro areas on the biological surface with the back propagation artificial neural network (BP-ANN).The visible spectrum (from 500nm to 730nm) of the micro areas with some specks on the surface of the apples was measured with the self- made fiber sensor spectrometer. A kind of BP-ANN with single hidden layer was devised to identify the characteristics on the biological surface automatically. It was also studied that the performance of the ANN with the different ranges of the output, the different numbers of the single hidden layer's units, the influence to the ANN if the noise was added to the input signals. Finally a three-stage ANN was founded to identify the four sorts of apples, the fleckless, the pushed, the scared, and the rotten�» 25,10,10 and 10 respectively were selected as training samples, and 15,10,10 and 10 respectively were selected as testing samples. This BP-ANN can achieve 85% accuracy if 20% noise was added.