Shape Judgment of Agricultural Products via Neural Network

Judgment of Shape and Size of Pimento

Abstract

Judgment of the complicated shape of agricultural products such as pimentos was tried with the neural network and the results of judgment were compared with those by human observation. The applicability of the neural network to the sizing and grading system was examined. In this research, several kinds, of neural network with three layer were constructed, and the shape and size of pimentos were judged with this network. The influence of the kinds of input signal to the learning time and the judging results were discussed.
The input signal which could reproduce the original shape decreased the learning time and judging time including the time for computing the input signal. However, the judging results were not so influenced by the kind of input signals and the degree of agreement with the human judgment was not so good. The reason of this result might be partly the quantitative and qualitative difference between the inputs into the neural network and the human.