Document Type : Original Article

Author

• Mohammad Shaker

Assistant Professor, Agricultural Engineering Research Department, Fars Agricultural and Natural Resources Research and Education Center, Agricultural Research Education And Extention Organization, Shiraz, Iran

Abstract

In this research, a machine vision system was used, which included three parts: suction box, sampling box, and imaging box. The above machine vision system was evaluated for seeds of corn and pinto bean. For each type of seed, the performance of the suction device, with three seed plates and four suction values was evaluated. In each amount of suction, the total number of seeds of sticked to the seed plate, the number of singled seeds and the number of seeds of sticked together on each hole were counted and their percentage was calculated. After that, for the two types of seeds tested, the algorithm for determining the percentage of breakage and the number of seed coding and validation of the algorithm was evaluated in 30 repetitions. The results showed that for corn seed, treatment 10 (seed plate with holes 2 mm and suction value -100 mm Hg) and for pinto bean seed, treatment 11 (seed plate with 2 mm holes and suction value -120 mm Hg) was the most suitable treatment. The validation results of the algorithm for determining the percentage of breakage and the number of seeds for corn seeds showed that the average accuracy of the algorithm was equal to 100%. For seed of pinto beans, the algorithm validation results showed that the average accuracy of the algorithm in determining the percentage of breakage and the number of seeds was 95.27 and 99.47%, respectively.

Keywords

• Seed purity test
• machine vision
• image processing
• breakage