Computer vision techniques for automatic determination of yield effective bad condition storage effects on various agricultural seed types

Plant seeds can be preserved for a long time. Long-term storage of seeds is very important for the continuity of life. Seed storage time heavily depends on plant species, seed maturation degree, preliminary processing, seed germination value and the moisture content. Environmental conditions such as temperature of the storage place, humidity, and light, or infestation of parasitic living beings; insects and fungus are some other first order parameters responsible for the seed storage time, too. Germinated and bug infested seeds are neither suitable for food consumption nor for cultivation. Automated discrimination of these situations with respect to healthy seeds requires an algorithm to be applied in the sense of seed image classification. This paper draws some image processing related steps to be applied for seed images which discriminate the germinated or bug infested ones. The problem itself stems from the seed type, the reason why some seed variations like beans show close texture to germinated sprouts. The experiments are handled by using some sort of seed types readily available in the seedsman such as lentil, chickpea, bean, corn, barley, cowpea, and wheat seeds. Each of the training and test pictures is obtained in an uncontrolled environment without the care of light and image acquisition details that makes the problem tougher. A brand-new database namely ITUSEED-II has already been constructed. The analysis of database related to healthy and badly conditioned seeds using a new computer vision method namely Relational Bit Operator (RBO) is presented in this paper.