How to distinguish Medicago seed species? 

Seed discrimination is crucial to ensure the overall fairness of the economical market and appropriate further seed processing. Distinguishing between hybrids and other species aids with maintaining seed purity across the industry. 

Currently, the most common method used for seed identification is molecular marking, however, it is a destructive method and cannot be used in real-time. A recent study elaborates on new innovative, rapid, and non-destructive methods for seed discrimination. Research conducted by the China Agricultural University explores how to effectively distinguish between sickle alfalfa (Medicago falcata), hybrid alfalfa (Medicago varia), and alfalfa seed (Medicago sativa). 

Seed identification by Stacking Ensemble Learning combined with spectral imaging

The researchers developed a three-layer model – Stacking Ensemble Learning (SEL) – to discriminate the appropriate seeds. The VideometerLab and its software were used for the imaging processing of the samples and their analysis. The model focus on both spectral and morphological features that allow for the distinguishing of Medicago seeds varieties. 

At the beginning of the study, the researchers used 33 features, however, by establishing a filtering process, they were able to determine a subset of features. It was shown that the contents of chlorophyll, anthocyanin, fat, and moisture of seeds allow for powerful discrimination of alfalfa seeds and its varieties. Furthermore, the analysis illustrated that intra-species variations in morphological features are greater than inter-species variations in the Medicago seed study.  

The newly developed SEL model appeared to be successful in terms of accuracy, reliability, precision, and sensitivity.  

The VideometerLab for seed discrimination  

These results provided a powerful model for discriminating Medicago species seeds and a theoretical basis for the development of low-cost and high-efficiency sensors. 

(Jia et al., 2022) 

Spectral imaging has been shown to be effective for various seed analyses, such as seed treatment, purity, germination, and more. The Videometer technology with its software has been tested in numerous studies in this industry. We are excited to see the positive results of our accurate, rapid measurements; it motivates us to enhance our robust solutions. Learn more about the recent study – seed identification on Medicago varieties with SEL and spectral imaging.