基于机器视觉几何特征参数的大蒜根部尺寸预测

Prediction of Garlic Root Size Based on Geometric Feature Parameters of Machine Vision

在出口大蒜的深加工工序中,如何精确快速地获取大蒜根部位置形状尺寸信息是实现自动化精准切须作业的关键步骤.以大蒜根部蒜胡的形状尺寸预测为研究目标,利用机器视觉技术提取大蒜的12个绝对形状特征建立大蒜特征训练集,通过建立回归预测模型,实现大蒜蒜胡周长的预测,并完成实验测试.测试结果表明,所建立的BP神经网络模型的预测值与实际值之间的平均绝对百分比误差为3.59%,SVR模型得到的平均绝对百分比误差为4.62%,两类预测模型均可实现对大蒜根部蒜胡周长的预测.

In the deep processing of export garlic,how to accurately and quickly obtain the position,shape and size information of the root of the garlic is a key step to achieve automatic and precise whisker shaving operations.This paper took the prediction of the shape and size of garlic roots as the research target,used machine vision technology to extract the 12 absolute shape features of garlic to build a garlic feature training set,and established a regression prediction model to achieve the prediction of the circumference of garlic whisker and completed experimental test.The test results show that the average absolute percentage error between the predicted value and the actual value of the established BP neural network model based on geometric feature parameters is 3.59%,and the average absolute percentage error obtained by the SVR model is 4.62%.The two types of prediction models can achieve the prediction of the circumference of garlic root.