Document Type : Original Article
Author
Assistant Professor, Agricultural Engineering Research Department, Fars Agricultural and Natural Resources Research and Education Center, Agricultural Research Education And Extention Organization, Shiraz, Iran
Abstract
Considering the problems of visual and human evaluation, machine vision, can be used as a suitable substitute for human vision. The purpose of this research was to build and evaluate a machine vision system to determine the breakage percentage and the number of wheat seeds. The system consists of three parts: suction box, sampling box and imaging box and evaluated for two cultivars Torabi and Azar wheat. In each type of wheat, the performance of the suction device was evaluated with two seed plates and four suction values. In each amount of suction, the number of singled seeds and the seeds of sticked together on each hole were counted and their percentage was calculated. The image prepared was transferred to MATLAB software and breakage determination algorithm and wheat seed number was coded and presented. The results showed that the most suitable treatment for wheat of Torabi cultivar, was seed plate with 1 mm hole and suction of -100 mm Hg with 95.31 percent singled seeds and 4.69 percent of sticked together seeds. For Azar cultivar, seed plate with 1 mm hole and suction of -120 mm Hg with a percentage of singled seeds of 91.6 and a percentage of sticked together seeds of 8.4 was the most appropriate treatment. The validation results of the algorithm showed that its accuracy for determining the percentage of breakage and the number of wheat seeds were 85.33 and 98.76%, respectively. It is suggested that this system be evaluated for seeds of different sizes.
Keywords