Moisture content assessment of dried Hami jujube using image colour analysis

To investigate the feasibility of image colour information in predicting moisture content dried Hami jujube, images were obtained under different space models, and model component mean chromaticity frequency sequences R, G, B, H, S, V, L*, a* b* extracted through analysis. After optimising sequence, was established compared. The results showed that GA-ELM (genetic algorithm - extreme learning machine) by CARS (competitive adaptive reweighted sampling) method to optimise 12 features S sequence had best prediction effect, with Rc 0.917, Rp 0.934 residual predictive deviation (RPD) 2.507. Therefore, can accurately predict jujube.