基于机器视觉的葡萄颗粒计数分级方法研究
Method of Counting the Grains of Grape Based on Machine Vision
本文提出了一种基于机器视觉的葡萄整穗颗粒计数方法。首先选择42串葡萄作为研究样本，分别用5种模型对葡萄果面面积和整穗颗粒数的数据进行最小二乘法拟合，选取拟合度最大，残差平方和最小的对数方程作为拟合方程。然后选择15串葡萄作为测试样本验证上述模型，将样本容量扩充至57串再次进行数据拟合，并且将所得到的拟合方程作为最优拟合方程。最后将该最优拟合方程运用于葡萄分级试验平台，计数精度达到88.83%。
he method of counting the grains of grape based on machine vision was proposed. Choosing the 42 bunch of grapes as the research sample, the data which include area of grape’s surface and counting the grains of grape were fitted by the least square method through five model, and select the logarithmic equation with maximum degree of fitting and minimum residual sum of squares as the fitted equation. The model mentioned above was validated by choosing 15 bunch of grapes, and select the equation which were fitted by 57 bunch of grapes as the best fitted equation. The best fitted equation was used in grape grading system and the experiment results showed that the counting accuracy was 88.83%.
李伟、林初靠、陈英
农业科学技术发展农艺学园艺
机器视觉，葡萄分级，颗粒数，最小二乘法
machine vision，grape grading，counting of grape，least square method
李伟,林初靠,陈英.基于机器视觉的葡萄颗粒计数分级方法研究[EB/OL].(2008-06-03)[2025-08-16].http://www.paper.edu.cn/releasepaper/content/200806-46.点此复制
评论