Indian Journal of Animal Nutrition

Guava is an important tropical and subtropical fruit which is grown widely in India. Guava stands fourth place among all fruits produced in India and are classified into different varieties based on colour, size, shape and firmness. Colour is one of the major maturity index for guava which varies from dark green during immature stage to full yellow when fully ripened. Manual classification is subjective, difficult and time consuming, Computer vision system has been proven to be an alternative method for objectively classifying fruits. This paper studies application of computer vision system to classify guava based on colour. Rightly matured guava (lucknow-49) has been harvested and images of these guavas were captured in RGB and converted to HSV colour model for classification. Colour features were extracted and histogram was plotted. The colour range was fixed by the manual expert's classification of guava fruits and analgorithm was developed for colour classification. A mechanical sorter device was also developed as an integral part of computer vision system which consists of pneumatic cylinder, solenoid and ejection board with sorting capacity of 20 fruits per minute.