Rice fatty acid detection method combining hyperspectral technology with improved LSSVM

Abstract

Objective: To solve the problems of low accuracy and poor efficiency in the existing rice quality testing methods of food enterprises. Methods: Based on a hyperspectral data acquisition system, a fast and non-destructive detection method for stored rice quality was proposed, which combined an improved bacterial foraging algorithm and least squares support vector machine. By applying the improved bacterial foraging algorithm, the hyperparameters (regularization parameter and kernel parameter) of the least squares support vector machine were optimized to achieve rapid and non-destructive detection of rice quality. its performance was analyzed through experiments. Results: The proposed method can achieve rapid and non-destructive detection of fatty acid content in stored rice, with a determination coefficient of 0.940 5, root mean square error of 0.543 5, and an average detection time of 1.12 seconds. Conclusion: The proposed detection method has high detection performance, which can be used for the identification and detection of rice quality.

Keywords

• rice
• fatty acid
• hyperspectral data
• bacterial foraging optimization algorithm
• least squares support vector machine
• rapid non-destructive testing