International Journal on Advanced Science, Engineering and Information Technology, Vol. 8 (2018) No. 4, pages: 1366-1371, DOI:10.18517/ijaseit.8.4.6513

Counting and Separating Damaged Seeds of Soybean Seeds using Image Processing

Monlica Wattana, Buris Siriluk, Suwanan Khotwit


This research proposed an image processing technique for counting soybean seeds and separating damaged seeds. The technique used was to adjust the image to black and white so that the soybean seeds differ from the background color. The second part, of the research, was the separation of soybean seeds using the distance transform method and the region growing method, to count soybean seeds. In the third part, it focused on the separation of damaged seeds by the size, the circle shape and HSV of soybean seeds. From 30 soybean seed images, the percentage of accuracy of counting and separating damaged seeds by naked eyes was 100 percent, and the average time spent was 13.70 seconds. The percentage of accuracy of counting the seeds by the developed program was 100 percent and the accuracy of the separation of damaged seeds was 99.80 percent. The average time spent was 6.49 seconds. The experimental results showed that the developed program took 2 times less to count the soybean seeds than the naked eyes. Therefore, the proposed algorithm of the program can help save the time for counting the soybean seeds and separate the damaged seeds.


image processing; seed count; damaged seeds; soybean seed quality inspection

Viewed: 606 times (since abstract online)

cite this paper     download