摘要
花椒是一种重要的烹饪调料和中药配料。本文将机器视觉技术引入到花椒品种的快速鉴别中,通过机器视觉硬件装置获取6类花椒样品图像共90幅,其中60幅作为训练样本,30幅作为测试样本。对所有样本分别提取了颜色和纹理的共10个特征参数,利用概率神经网络对特征数据进行鉴别,正确识别率为93.33%。本文研究的基于机器视觉的花椒品种鉴别方法可以快速准确地提取花椒样品的特征数据,为批量分选花椒奠定了技术基础。
Zanthoxylurn Bungeanum Maxim is a kind of important ingredients for cooking and traditional Chinesemedicine. In this paper,the computer vision technology was introduced to discriminate the varieties of Zanthoxylurn Bungeanum Maxim rapdily. The Zanthoxylurn Bungeanum Maxim image of six categories were obtained by the com-puter vision hardware device,and the total number of images was 90,of which 60 were used as training samples and30 as test samples. The 10 feature parameters of color and texture features were extracted from all the samples,whichwere identified by the probabilistic neural network. The correct recognition rate was 93.33%. The discriminationmethod of Zanthoxylurn Bungeanum Maxim varieties by computer vision can quickly and accurately extract the char-acteristic data of the Zanthoxylurn Bungeanum Maxim samples,which lays the technical foundation for batch sorting.
出处
《传感技术学报》
CAS
CSCD
北大核心
2016年第1期136-140,共5页
Chinese Journal of Sensors and Actuators
基金
河南省科技攻关计划项目(142102110054)
河南省教育厅科学技术研究重点项目(12A510014)
关键词
机器视觉
颜色特征
纹理特征
概率神经网络
花椒
computer vision
color feature
texture feature
probabilistic neural network
Zanthoxylurn Bungeanum Maxim
