利用融合高度与单目图像特征的支持向量机模型识别杂草

被引：32

作者：

王璨

李志伟

机构：

[1] 山西农业大学工学院

来源：

农业工程学报 | 2016年 / 32卷 / 15期

关键词：

双目视觉; 支持向量机; 特征提取; 杂草识别; 双目图像; 特征融合;

D O I：

暂无

中图分类号：

TP391.41 [];

学科分类号：

080203 ;

摘要：

除草是保证农作物高产的必要工作。针对机械化除草和智能喷药中存在的杂草识别问题,以2～5叶苗期玉米及杂草为研究对象,进行了融合高度特征与单目图像特征的杂草识别方法研究。首先从单目图像中提取16个形态特征和2个纹理特征;然后基于双目图像,提出了针对植株的高度特征提取方法,所得高度特征与实际测量值间误差在±12 mm以内;利用max-min ant system算法对形态特征进行优化选择,将形态特征减少到6个,有效减少数据量62.5%,并与纹理和高度特征进行融合;将2～5叶玉米幼苗的可除草期划分为3个阶段,分别构建融合高度特征与单目图像特征的SVM识别模型,并与相应不含高度特征模型进行对比。经测试,3个阶段模型的识别准确率分别为96.67%,100%,98.33%;平均识别准确率达98.33%。不含高度特征模型的识别准确率分别为93.33%,91.67%,95%;平均识别准确率为93.33%。结果表明,融合高度特征与单目图像特征的SVM识别模型优于不含高度特征模型,平均识别准确率提高了5百分点。该方法实现了高准确率的杂草识别,研究结果为农业精确除草的发展提供参考。

引用

页码：165 / 174

页数：10

共 24 条

[11] Exploiting affine invariant regions and leaf edge shapes for weed detection[J] . Wajahat Kazmi,Francisco Garcia-Ruiz,Jon Nielsen,Jesper Rasmussen,Hans J?rgen Andersen.Computers and Electronics in Agriculture . 2015
[12] Texture Feature Extraction Method Combining Nonsubsampled Contour Transformation with Gray Level Co-occurrence Matrix[J] . Xiaolan He,Yili Wu,Yiwei Wu.Journal of Multimedia . 2013 (6)
[13] Modular-based classification system for weed classification using mixture of features[J] . W.K. Wong,Ali,Chekima,Choo,Chee,Wee,Khoo,Brendon,Muralindran,Marriappan.Int. J. of Computational Vision and Robotics . 2013 (4)
[14] Weed Recognition Using Image-Processing Technique Based on Leaf Parameters[J] . Kamal N. Agrawal1,Karan Singh1,Ganesh C. Bora2 and Dongqing Lin2.Journal of Agricultural Science and Technology B . 2012 (8)
[15] A Modified Gray level Co-occurrence Matrix based Thresholding for Object Background Classification[J] . A. Dash,P. Kanungo,B.P. Mohanty.Procedia Engineering . 2012
[16] Speeded-Up Robust Features (SURF)[J] . Herbert Bay,Andreas Ess,Tinne Tuytelaars,Luc Van Gool.Computer Vision and Image Understanding . 2007 (3)
[17] A vision-based method for weeds identification through the Bayesian decision theory[J] . Alberto Tellaeche,Xavier P. Burgos-Artizzu,Gonzalo Pajares,Angela Ribeiro.Pattern Recognition . 2007 (2)
[18] Autonomous robotic weed control systems: A review[J] . D.C. Slaughter,D.K. Giles,D. Downey.Computers and Electronics in Agriculture . 2007 (1)
[19] Efficient RANSAC for point-cloud shape detection
Schnabel, R.
Wahl, R.
Klein, R.
[J]. COMPUTER GRAPHICS FORUM, 2007, 26 (02) : 214 - 226
[20] Colour and shape analysis techniques for weed detection in cereal fields
Pérez, AJ
López, F
Benlloch, JV
Christensen, S
[J]. COMPUTERS AND ELECTRONICS IN AGRICULTURE, 2000, 25 (03) : 197 - 212

← 1 2 3 →