Establishment of near infrared prediction model for nutritional indexes of rice processing by-products
In order to achieve rapid detection of nutritional substances in by-products of rice processing and promote on-demand feed processing and grading of processing by-products,a study was conducted using a total of 302 samples of rice processing by-products(107 broken rice samples,77 belly white rice samples,and 118 discolored grain samples).Through standard methods of physicochemical data detection,sample pretreatment,near-infrared spectroscopy acquisition,spectral data preprocessing,feature wavelength selection,and model establishment,quantitative prediction models for moisture,crude protein,crude fat,and crude fiber nutritional parameters for the overall categories of broken rice,belly white rice,and discolored grain in rice processing by-products were ultimately established.The results showed that the coefficient of determination(R2)for the established models was greater than 0.85,and the relative prediction error(RPD)was greater than 2.5,indicating good accuracy and stability.The study indicates that the method is rapid,efficient and accurate,and greatly improves the comprehensive utilization rate of by-products of rice processing.
broken ricebelly white ricediscolored grainmoisturecrude proteincrude fatcrude fiber
国家科技期刊平台
NSTL
万方数据
