基于激光传感器的稻种轮廓形状测量方法研究  

Study of outline shape measurement method of rice seed based on laser sensor

作　　者：曹鹏[1] 尹文庆[1] 吴林华[1] 王川[1] 范丽[1] 杨志军[1] CAO Peng YIN Wen-qing WU Lin-hua WANG Chuan FAN Li YANG Zhi-jun(Key Laboratory of Intelligent Equipment for Agriculture of Jiangsu Province, College of Engineering, Nanjing Agricultural University, Nanjing 210031, China)

机构地区：[1]南京农业大学工学院江苏省农业智能化装备重点实验室,江苏南京210031

出　　处：《传感器与微系统》2017年第3期29-32,共4页Transducer and Microsystem Technologies

基　　金：江苏省农机三项工程项目(NJ2010-02);南京农业大学青年科技创新基金资助项目(KJ2010032)

摘　　要：稻种轮廓形状测量是稻种基于形状特征识别的前提条件。以激光传感器采集到的稻种表面点云为基础,提出了一种基于主平面剖视图投影的稻种轮廓形状测量方法。通过主成份分析法校正点云模型的坐标系,利用主平面投影法获取稻种的六视图投影模型,应用Alpha Shape算法提取投影点云模型的边缘轮廓,并计算稻种投影模型的周长、面积、长、宽、长宽比、最大半径、最小半径、半径比、圆形度等9个形状特征参数。选取大华香糯、豫粳6号、新稻10号3大类,每种各200粒稻种作为试验样品,采用径向基函数(RBF)神经网络模型对提取到的形状特征进行训练识别,识别率分别达到96%,94%,98%。结果表明:基于激光传感器的稻种轮廓形状测量方法能够较好地适用到稻种识别中。Outline shape measurement of rice seed is key to non-destructive testing of rice seed. On the basis of point cloud on rice seed surface acquired by laser sensor,present an outline shape measuring method of rice seed based on main plane of projection. By principal component analysis method,calibrate coordinate of point cloud model,obtain six main plane of projection of seed rice seed using principal plane projection method,apply Alpha Shape algorithm to extract edge outline of the projected point cloud model,and calculate nine shape feature parameters of perimeter,area,length,width,aspect ratio,maximum radius,minimum radius,radius ratio,circularity of rice seed projection model. Select three categories of rice seed,200 grain per each class as a test sample,using RBF neural network model for training recognition extracted shape feature,recognition rate reach 96 %,94 %,98 %,respectively. The results show that the outline shape measurement method of rice seed based on laser sensor can well be applied to identification of rice seed.

关 键 词：激光传感器 稻种 无损检测 点云 径向基函数神经网络 

分 类 号：TH741[机械工程—光学工程]