Review of seed quality and safety tests using optical sensing technologies

Seeds are of great importance to agricultural and industrial production. As such, rapid and non-destructive detection methods must be developed for the industry and consumers to ensure seed quality and safety. Among currently available non-destructive detection methods, optical sensing is one of the most promising technologies and has been used in seed evaluation for several years. This review primarily focuses on three widely used optical sensing technologies, namely, machine vision, near-infrared spectroscopy and hyperspectral imaging. Fundamental and essential instruments of the three technologies are introduced, followed by techniques in image and spectral processing, multivariate data analysis and commonly used software. Classification and prediction methods for attribute assessment of external or internal quality of various seeds are also compared. Result shows that optical sensing technologies are accurate non-destructive detection methods with great potential for quality control and safety assessment of seeds.