Abstract:

Evaluation of maturity grading of fresh corn ears was realized through computer vision. Image characteristics of maturity grading were extracted after twodimensional discrete wavelet decomposition. In HIS color model wavelet decomposition of H component was carried out, and the color characteristics of H component were extracted from low frequency subband. The texture characteristics were extracted from high frequency subband in RGB color model. Dimension reduction of characteristic parameters was implemented using principal component analysis method. The first four principal component values were taken as the network inputs. The probabilistic neural network was developed for maturity grading of fresh corn ears, and the grading accuracy is about 95%.

Key words: food machinery, twodimension discrete wavelet decomposition, fresh corn ear, maturity grading, neural network