Design and Performance of Soybean Seeds Sorting Machine Based on Image Processing
Mas Wisnu Aninditya (a*), Wawan Hermawan (b), Slamet Widodo (b)

a) Indonesian Agricultural Engineering Polytechnic, Indonesian Ministry of Agriculture
Jalan Sinarmas Boulevard No.1, Kabupaten Tangerang, Banten 15338, Indonesia
b) Agricultural Engineering Study Program, Department of Mechanical and Biosystem Engineering, Faculty of Agricultural Technology, IPB University
Jalan Raya Dramaga, Babakan, Kec. Dramaga, Kabupaten Bogor, Jawa Barat 16680
*maswisnua[at]gmail.com

Abstract

One of the obstacles in increasing soybean production in Indonesia is the ability to supply good quality soybean seeds. Indonesia is still constrained by the manual sorting system, which has low capacity and relatively low quality of sorting results. To overcome this problem, it is necessary to design a machine that is able to sort soybean seeds with a higher capacity, good results quality, and at an affordable price for farmers. The designed machine has a sieving system for sorting standard-sized seeds, a seed image capture and image processing system for selecting mature seeds, and an ejector system for removing rejected seeds. The prototype then was tested to evaluate the sorting performance, using soybean seeds of Anjasmoro var.
The test results showed that the accuracy of the image processing was 79% and the accuracy of the ejector was 89%. The sorting capacity is directly proportional to the seed feeding rate, in which the sorting capacity was 3.5 kg/h at 3 seeds/s feeding rate.

Keywords: Image processing, Performance, Sorting machine, Soybean seed

Topic: Agricultural Engineering