基于机器视觉的君迁子种子尖端特征识别方法研究

摘要

针对君迁子种子的外形特点,基于Ramer轮廓拟合算法,结合Halcon算法工具包,提出了一种基于Ramer算法的君迁子种子尖端识别方法.结果表明, 该方法可以对种子轮廓进行拟合分割,确定种子的尖端位置.通过对5种不同品种君迁子的250粒种子进行识别,准确率达83.6%.%By analyzing the feature of the seeds, a method of identifying the seeds based on the Ramer algorithm was developed, which was based on the Ramer algorithm and the tool kit of Halcon.The experimental results showed that the method could fit and segment the seed profile, and finally determine the tip position of the seed.Through the identification of 250 seeds of 5 different species, the proposed method got accuracy of 83.6% for the tip position of the seeds.