高级检索

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

叶片传统特征和距离矩阵与角点矩阵相结合的树种识别算法

陈明健 陈志泊 杨猛 莫琴

陈明健, 陈志泊, 杨猛, 莫琴. 叶片传统特征和距离矩阵与角点矩阵相结合的树种识别算法[J]. 北京林业大学学报, 2017, 39(2): 108-116. doi: 10.13332/j.1000-1522.20160351
引用本文: 陈明健, 陈志泊, 杨猛, 莫琴. 叶片传统特征和距离矩阵与角点矩阵相结合的树种识别算法[J]. 北京林业大学学报, 2017, 39(2): 108-116. doi: 10.13332/j.1000-1522.20160351
CHEN Ming-jian, CHEN Zhi-bo, YANG Meng, MO Qin. Research on tree species identification algorithm based on combination of leaf traditional characteristics and distance matrix as well as corner matrix[J]. Journal of Beijing Forestry University, 2017, 39(2): 108-116. doi: 10.13332/j.1000-1522.20160351
Citation: CHEN Ming-jian, CHEN Zhi-bo, YANG Meng, MO Qin. Research on tree species identification algorithm based on combination of leaf traditional characteristics and distance matrix as well as corner matrix[J]. Journal of Beijing Forestry University, 2017, 39(2): 108-116. doi: 10.13332/j.1000-1522.20160351

叶片传统特征和距离矩阵与角点矩阵相结合的树种识别算法

doi: 10.13332/j.1000-1522.20160351
基金项目: 

国家自然科学基金项目 61402038

详细信息
    作者简介:

    陈明健。主要研究方向:数据库技术、模式识别。Email:mingjianok@163.com  地址: 100083 北京市海淀区清华东路35号北京林业大学信息学院

    责任作者:

    陈志泊,教授,博士生导师。主要研究方向:数据库技术、计算机软件与理论。Email:zhibo@bjfu.edu.cn  地址:同上

  • 中图分类号: TP391.4;S718.3

Research on tree species identification algorithm based on combination of leaf traditional characteristics and distance matrix as well as corner matrix

  • 摘要: 针对基于叶片特征进行树种识别的问题,本文在结合叶片纹理、不变矩以及传统形状共25维传统特征的基础上,自定义了叶尖角、边角均值等2个叶片轮廓特征,并以相似多边形定义及其推论作为理论依据,提出了一种基于叶片轮廓构建距离矩阵与角点矩阵进行树种识别的分类方法。该方法首先对树木叶片图像进行预处理,提取出归一化的叶片特征向量,然后利用KNN最近邻分类器筛选出相似度最高的前20个结果集(Top 20),然后构建距离矩阵和角点矩阵进行更为精确的识别匹配。在图像预处理阶段,为获取更为准确的叶片轮廓特征,利用叶片在HSV颜色空间中饱和度特征以及色度特征方面的显著差异性,设计了一种消除叶片阴影的图像预处理算法。在识别匹配阶段,利用Douglas Peucker approximation算法提取叶片轮廓的近似多边形,定义了距离矩阵、角点矩阵、矩阵中元素间相似度、矩阵相似度以及综合相似度计算方法,设计了全局匹配与局部匹配相结合的算法。该算法在Android系统的手机平台上进行了实现和运行验证,结果表明:在Flavia数据集中,对32种共1 907个正常叶片样本的识别准确率为99.61%,对32种共851个残叶样本的准确率为94.92%;在Leafsnap数据集中,对185种共23 147个Lab样本前5个结果集(Top 5)的识别准确率为98.26%。相对其他算法,该算法识别准确率更高,对叶片外形描述能力更强,对残叶、扭曲叶、阴影叶具有更好的鲁棒性,算法的实用性和适应性更强。

     

  • 图  1  叶片图像去除阴影流程

    Figure  1.  Process of removing blade image shadow

    图  2  去除阴影效果图

    Figure  2.  Effect image of removing shadow

    图  3  多边形拟合效果图

    Figure  3.  Fitting effect image of polygon

    图  4  叶尖角示意图

    Figure  4.  Sketch map of leaf angles

    图  5  叶尖角计算流程

    Figure  5.  Calculation process of leaf angles

    图  6  整体识别流程

    Figure  6.  Overall identification process

    图  7  Flavia数据库的一个子集

    Figure  7.  A subset of the Flavia database

    图  8  Leafsnap数据库的一个子集

    Figure  8.  A subset of the Leafsnap database

    图  9  准确率随范围变化趋势

    Figure  9.  Accuracy rate with the changing trend of top range

    表  1  Flavia数据集实验结果

    Table  1.   Experimental results of Flavia datasets

    %

    选择的特征
    Chosen feature
    正常叶片Entire leaf 残缺叶片Incomplete leaf
    排序第1结果集的
    识别准确率
    Top 1 accuracy
    排序前5结果集的
    识别准确率
    Top 5 accuracy
    排序第1结果集的
    识别准确率
    Top 1 accuracy
    排序前5结果集的
    识别准确率
    Top 5 accuracy
    27维特征+KNN 27 feature+KNN 94.11 97.29 83.12 87.06
    形状特征+KNN Shape feature +KNN 81.6 86.30 73.11 76.44
    形状+纹理+KNN Shape+texture+KNN 84.3 87.61 81.32 85.69
    27维特征+SVM 27 feature+SVM 96.77 85.97
    27维特征+DBNs 27 feature+DBNs 98.70 91.72
    27维特征+Fourier+LBP+DBNs 27 feature+Fourier+LBP+DBNs 99.26 93.69
    27维特征+KNN+本文匹配算法27 feature+KNN+algorithm in this paper 99.61 99.79 94.92 95.20
    多尺度HoCS特征+KNN Multi-scale HoCS characteristics+KNN 75.66 93.28
    轮廓点与质心相对距离直方图+KNN Histogram of relative distance between contour points and centroid+KNN 90.50 97.81 86.57 90.63
    下载: 导出CSV

    表  2  Leafsnap数据集实验结果

    Table  2.   Experimental results on Leafsnap datasets

    %

    选择的特征
    Chosen feature
    Lab子集第1个
    结果集的准确率
    Lab-top 1
    accuracy
    Lab子集前5个
    结果集的准确率
    Lab-top 5
    accuracy
    Flavia子集第1个
    结果集的准确率
    Field-top 1
    accuracy
    Flavia子集前5个
    结果集的准确率
    Field-top 5
    accuracy
    27维特征+K最近邻+本文预处理
    27 feature+KNN+ shadow elimination
    85.37 89.42 84.16 87.33
    27维特征+K最近邻+未消阴影预处理
    27 feature+KNN+did not eliminate the shadow
    70.79 77.38 65.14 69.43
    形状特征+ KNN Shape feature +KNN 71.62 74.05 69.82 73.21
    形状+纹理+ KNN Shape + texture+KNN 83.14 86.78 80.01 84.70
    27维特征+SVM 27 feature+SVM 86.42 - 86.76 -
    27维特征+ DBNs 27 feature+DBNs 93.51 - 90.37 -
    27维特征+Fourier+LBP+DBNs
    27 feature+Fourier+LBP+DBNs
    96.82 - 93.14 -
    27维特征+KNN+本文匹配算法
    27 feature+KNN+algorithm in this paper
    97.12 98.26 95.48 96.30
    多尺度HoCS特征+ KNN
    Multi-scale HoCS characteristics+KNN
    72.62 93.05 70.48 92.11
    轮廓点与质心相对距离直方图+ KNN
    Histogram of relative distance between contour points and centroid+KNN
    86.32 90.81 84.09 88.87
    下载: 导出CSV
  • [1] 王丽君, 淮永建, 彭月橙.基于叶片图像多特征融合的观叶植物种类识别[J].北京林业大学学报, 2015, 37(1):55-61. doi: 10.13332/j.cnki.jbfu.2015.01.006

    WANG L J, HUAI Y J, PENG Y C.Follow-leaf plant species identification based on multi-feature fusion of leaf images[J]. Journal of Beijing Forestry University, 2015, 37(1):55-61. doi: 10.13332/j.cnki.jbfu.2015.01.006
    [2] 袁津生, 姚宇飞.基于分形维度的叶片图像识别方法[J].计算机工程与设计, 2012, 33(2):670-673. doi: 10.3969/j.issn.1000-7024.2012.02.050

    YUAN J S, YAO Y F.Face image recognition method based on fractal dimension[J]. Computer Engineering and Design, 2012, 33(2): 670-673. doi: 10.3969/j.issn.1000-7024.2012.02.050
    [3] YAHIAOUI I, MZOUGHI O, BOUJEMAA N. Leaf shape descriptor for tree species identification[C]//2012 IEEE International Conference on Multimedia and Expo. Amsterdam: IEEE, 2012: 254-259. https: //www.researchgate.net/publication/261125339_Leaf_Shape_Descriptor_for_Tree_Species_Identification
    [4] 李洋, 李岳阳, 罗海驰, 等.基于形状特征的植物叶片在线识别方法[J].计算机工程与应用, 2017, 53(2):162-165. doi: 10.3778/j.issn.1002-8331.1504-0233

    LI Y, LI Y Y, LUO H C, et al. On-line identification method of plant leaf based on shape feature[J]. Computer Engineering and Applications, 2017, 53(2):162-165. doi: 10.3778/j.issn.1002-8331.1504-0233
    [5] SATTI V, SATYA A, SHARMA S. An automatic leaf recognition system for plant identification using machine vision technology[J]. International Journal of Engineering Science and Technology, 2013, 5(4): 874-879. http://d.old.wanfangdata.com.cn/OAPaper/oai_doaj-articles_83ccc22bcf4760049fb096a6365da736
    [6] BACKES A R, CASANOVA D, BRUNO O M. Plant leaf identification based on volumetric fractal dimension[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2009, 23(6): 1145-1160. doi: 10.1142/S0218001409007508
    [7] 谢莉, 成运, 曾接贤, 等.基于颜色和梯度方向共生直方图的图像检索[J].计算机工程与应用, 2016, 52(10):181-186. doi: 10.3778/j.issn.1002-8331.1406-0366

    XIE L, CHENG Y, ZENG J X, et al. Image retrieval based on color and gradient directional symbiotic histogram[J]. Computer Engineering and Applications, 2016, 52(10):181-186. doi: 10.3778/j.issn.1002-8331.1406-0366
    [8] MALLAH C, COPE J, ORWELL J. Plant leaf classification using probabilistic integration of shape, texture and margin features[J/OL].[2014-01-06]. http://www.actapress.com/Abstract.aspx?paperId=455022.
    [9] WANG X F, HUANG DS, DU J X, et al. Classification of plant leaf images with complicated background[J]. Applied Mathematics & Computation, 2008, 205(2):916-926. https://www.sciencedirect.com/science/article/abs/pii/S0096300308003718
    [10] NOVOTNY P, SUK T. Leaf recognition of woody species in Central Europe[J]. Biosystems Engineering, 2013, 115(4): 444-452. doi: 10.1016/j.biosystemseng.2013.04.007
    [11] LE T L, NGUYEN Q K, PHAM N H. Leaf based plant identification system for android using surf features in combination with bag of words model and supervised learning[C]//2013 International Conference on Advanced Technologies for Communications (ATC 2013). Ho Chi Minh City: IEEE, 2013: 404-407. https: //www.researchgate.net/publication/259644407_Leaf_based_plant_identification_system_for_Android_using_SURF_features_in_combination_with_Bag_of_Words_model_and_supervised_learning
    [12] SUMATHI C S, KUMAR A V S. Neural network based plant identification using leaf characteristics fusion[J]. International Journal of Computer Applications, 2014, 89(5):31-35. doi: 10.5120/15499-4141
    [13] 刘念, 阚江明.基于多特征融合和深度信念网络的植物叶片识别[J].北京林业大学学报, 2016, 38(3):110-119. http://d.old.wanfangdata.com.cn/Periodical/bjlydxxb201603014

    LIU N, KAN J M. Establishment of plant leaf based on multi-feature fusion and depth-of-concept network[J]. Journal of Beijing Forestry University, 2016, 38(3): 110-119. http://d.old.wanfangdata.com.cn/Periodical/bjlydxxb201603014
    [14] SHAYAN H, SAJEEVAN G. Plant recognition from leaf image through artificial neural network[J]. International Journal of Computer Applications, 2013, 62(17):15-18. doi: 10.5120/10172-4897
    [15] HU R X, JIA W, LING H B, et al. Multiscale distance matrix for fast plant leaf recognition[J]. IEEE transactions on image processing, 2012, 21(11): 4667-4672. doi: 10.1109/TIP.2012.2207391
    [16] GWO C Y, WEI C H. Plant identification through images: using feature extraction of key points on leaf contours1[J]. Applications in Plant Sciences, 2013, 1(11):1-9 http://med.wanfangdata.com.cn/Paper/Detail/PeriodicalPaper_PM25202493
    [17] KUMAR N, BELHUMENR P N, BISWAS A, et al. Leafsnap: a computer vision system for automatic plant species identification[EB/OL].[2016-08-20]. http://neerajkumar.org/base/papers/nk_eccv2012_leafsnap.pdf.
    [18] 董红霞, 郭斯羽.一种结合形状与纹理特征的植物叶片分类方法[J].计算机工程与应用, 2014, 50(23):185-188. doi: 10.3778/j.issn.1002-8331.1301-0138

    DONG H X, GUO S Y. Classification of plant leaf based on shape and texture features[J]. Computer Engineering and Applications, 2014, 50 (23): 185-188. doi: 10.3778/j.issn.1002-8331.1301-0138
    [19] 朱黎辉, 李晓宁.基于多特征组合的球形果蔬智能分选方法[J].计算机工程与应用, 2016, 52(5):173-178. doi: 10.3778/j.issn.1002-8331.1403-0280

    ZHU L H, LI X N.Intelligent sorting method of spherical fruits and vegetables based on multi-feature combination[J]. Computer Engineering and Applications, 2016, 52(5): 173-178. doi: 10.3778/j.issn.1002-8331.1403-0280
    [20] VISVALINGAM M, WHYATT D. Line generalisation by repeated elimination of points[J]. The Cartographic Journal, 1993, 30(1):46-51. doi: 10.1179/caj.1993.30.1.46
    [21] 曾接贤, 刘秀朋, 符祥.角点距离矩阵和同心圆划分的曲线描述与匹配[J].中国图象图形学报, 2012, 17(8):1011-1020. http://d.old.wanfangdata.com.cn/Periodical/zgtxtxxb-a201208016

    ZENG J X, LIU X P, FU X. Curve description and matching of corner distance matrix and concentric circle partition[J]. Chinese Journal of Image and Graphics, 2012, 17(8): 1011-1020. http://d.old.wanfangdata.com.cn/Periodical/zgtxtxxb-a201208016
    [22] WU S G, BAO F S, XU E Y, et al. A leaf recognition algorithm for plant classification using probabilistic neural network[C]//2007 IEEE International Symposium on Signal Processing and Information Technology.Cairo: IEEE, 2007: 11-16. http: //www.oalib.com/paper/4021582
    [23] WU S G, BAO F S, XU E Y, et al. A leaf recognition algorithm for plant classification using PNN (probabilistic neural network)[EB/OL].[2016-09-10]. http://flavia.sourceforge.net/.
  • 加载中
图(9) / 表(2)
计量
  • 文章访问数:  774
  • HTML全文浏览量:  232
  • PDF下载量:  11
  • 被引次数: 0
出版历程
  • 收稿日期:  2016-10-27
  • 修回日期:  2016-12-21
  • 刊出日期:  2017-02-01

目录

    /

    返回文章
    返回