首页|基于激光超声的玉米虫蚀粒检测研究

基于激光超声的玉米虫蚀粒检测研究

Detection of Corn Insect Damage Based on Ultrasonic Laser

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
虫蚀影响玉米的外观质量和营养价值,研究采用了一种基于激光超声的玉米虫蚀粒检测方法对玉米虫蚀粒进行检测.首先,利用脉冲激光照射完善粒和虫蚀粒玉米的表面产生激光超声信号.然后,提取了超声信号时域峰值因子和脉冲因子,频域重心频率和均方频率,Hilbert域高频能量作为特征参数.最后,将这5个特征分别作为BP神经网络和粒子群优化支持向量机(PSO-SVM)算法的输入对虫蚀粒和完善粒进行了分类识别.实验结果表明,PSO-SVM算法建立的分类模型对玉米完善粒和虫蚀粒的分类识别更准确,训练集和测试集准确率分别为99.72％和98.33％,所采用的方法是可行的.
In order to detect insect-damaged corn kernels affecting the appearance quality and nutritional value of corn,a laser ultrasound-based method was proposed in this paper.Firstly,pulsed laser was used to irradiate the surfaces of intact and insect-damaged corn kernels,generating laser ultrasound signals.Then,time-domain peak factor and pulse factor,frequency-domain centroid frequency and mean frequency,and high-frequency energy in Hilbert domain were extracted as feature parameters from the ultrasound signals.Finally,these five features were used as inputs for particle swarm optimization support vector machine(PSO-SVM)to classify and identify insect-damaged kernels and intact kernels.Experimental results indicated that the classification model established by PSO-SVM algorithm was more accurate for the classification and recognition of maize perfect grains and insect etched grains,and the accuracy of the training set and test set was 99.72％and 98.33％,respectively,and the method a-dopted was feasible.

ultrasoniccorn insect damageultrasound signalfeature extractionparticle swarm optimization sup-port vector machines

卢涛、赵中义、吴才章、赵志科

展开 >

河南工业大学电气工程学院,郑州 450001

超声 玉米虫蚀 超声信号 特征提取 粒子群优化支持向量机

河南省科技研发计划联合基金河南省自然科学基金河南省自然科学基金中原学者工作站项目

222103810084162300410054222102220100224400510030

2024

10.20048/j.cnki.issn.1003-0174.000726

中国粮油学报
中国粮油学会

中国粮油学报

CSTPCD北大核心
影响因子：1.056
ISSN：1003-0174
年,卷(期)：2024.39(6)