Hyperspectral detection method for starch content of potato

Wei Jiang, Shuwen Wang, Yongcun Fan, Junlong Fang
2015 unpublished
The purpose of this research is to explore to detect and predict the starch content of potato by the hyperspectral imaging technique. The original images of 118 samples were acquired by the hyperspectral imaging system, and then the suitable pretreatment was selected. The prediction model for starch content of potato was constructed by the 100 hyperspectral images randomly selected which was dealt with smoothing, PCA and PLS. The determination coefficient of this model was 0.8234, and the root
more » ... 8234, and the root mean square error of 0.5633. The other 18 sample images were used to validate the accuracy of model, and the determination coefficient R2 was 0.9031 and the root mean square error (RMSEP) was 0.5025. The results shown that it is feasible to detect and predict the starch content of potato by hyperspectral image technology.
doi:10.14257/astl.2015.111.09 fatcat:6docijilybdh7igmjcj67pyut4