Comparative Study of Distance Discriminant Analysis and Bp Neural Network for Identification of Rapeseed Cultivars Using Visible/Near Infrared Spectra

摘要: The potential of visible/near infrared spectra as a method nondestructive discrimination various rapeseed cultivars was evaluated, ability distance discriminant analysis (DDA) and BP neural network (BPNN) for identification shown in this article. spectral curves ranging from 350 to 2500 nm were obtained by VIS/NIR spectroscopy, the principal component (PCA) applied perform clustering analysis. first 6 principle components (PCs) extracted PCA employed inputs DDA BPNN, respectively, then two different models built. Forty-five samples each species total 225 5 categories extracted. One hundred fifty elected randomly training sets set up model which validated prediction formed remaining 75 samples. result error BPNN be ±0.15, indicated that no exceeded threshold value, therefore distinguishing rate 100%. displayed recognition 100% achieved. Although methods mentioned presented paper good approaches cultivars, with functions more intuitive than convenient machine recognition.