Although medicinal plants play an important role in the drug industry and health care and attract much attention, there are few studies on medicinal plant seed identification.” This paper presents an automatic system for medicinal plant seed identification using machine vision based on a study of 75 species. Results of this study have led to 23 different classifications for seeds based on their aspect ratio, eccentricity and six colour features (red, green, blue, hue, intensity and saturation). The use of different combinations of morphological and colour features led to a range of training and test accuracy figures during seed identification. In combinations of just colour features, the highest average accuracy values belonged to six colour features with 99 and 88% training and test accuracy, respectively. In different combinations of one morphological feature and various colour features, the highest average accuracy values were seen in the combination of one morphological and two colour features with 92 and 80% training and test accuracy, respectively. In different combinations of two morphological and various colour features, the highest average values occurred in a combination of two morphological and three colour features with 65 and 58% training and test accuracy, respectively. The accuracy of purity testing is one of the most important concerns of ISTA in seed testing studies, and this study shows that machine vision could support effectively seed purity analysis

Keywords