Peanut Shape Recognition Based on Fourier Descriptor

A sort of software for peanut shape identification based on artificial neural network was developed. Images of peanut which is advantageous for carries on the characteristic extraction were acquired by means of red component extraction, filter, image division, edge examination, and so on. The method to describe the shape of irregular peanut was studied, in which the Fourier transform and Fourier inverse transform were applied. It was concluded that the first thirteen harmonics of the Fourier descriptor were enough to represent the primary shape of peanut, The method achieved an accuracy of 90% for oblong peanuts, 93.3% for simple peanuts, 96.7% for trilateral peanuts, 100% for elliptic peanuts and 93.3% for circular peanuts.