摘要
为解决苹果分级准确率低和速度慢等缺陷,提出一种基于粒子群优化(PSO)改进算法的最小二乘支持向量机(LS-SVM)苹果分级检测方法,通过对苹果特征的优化选择,从而大规模缩减分类前LS-SVM训练样本数据,提高分类器训练效率。苹果分级实验表明,此方法能从红富士苹果的16个形状特征中提取出5个最优特征,用最优特征分级的正确率达96%以上,效果显著,该方法具有可行性。
To solve the low accuracy and velocity of apple grading,a grading method of LS-SVM based on improved PSO algorithm is presented,which is a new improved form by synthesizing the existing model of PSO.This method can optimize features and reduce the number of training samples greatly before classification,so as to enhance the training efficiency.Experimental results on the Fuji apple database show that the method can extract 5 optimal features from 16 shape features and achieve a high correct identification ratio which is up to 96%,the application effect is very notable so the proposed method is effective.
出处
《计算机与现代化》
2012年第5期161-163,166,共4页
Computer and Modernization
基金
盐城工学院重点建设学科开放基金资助项目(XKY2010021)
关键词
苹果分级
粒子群优化算法
最小二乘支持向量机
apple grading
particle swarm optimization（PSO）
least squares support vector machines（LS-SVM）
