黄巧义^{1, 2,},
樊小林²,
张木¹,
黄旭¹,
李苹¹,
付弘婷¹,
唐拴虎^1,,
1.广东省农业科学院农业资源与环境研究所/农业部南方植物营养与肥料重点实验室/广东省养分资源循环利用与耕地保育重点实验室 广州 510640
2.华南农业大学/广东高校环境友好型肥料工程技术研究中心 广州 510642
基金项目: 公益性行业（农业）科研专项201503123
广东省科技计划项目2016A020210035
广东省科技计划项目2014B090904068
广州市创新团队项目2016B070701009

详细信息

作者简介:黄巧义, 主要从事植物营养及高效施肥技术研究。E-mail: huangqiaoyi@hotmail.com

通讯作者:唐拴虎, 主要从事新型肥料及植物营养研究。E-mail: 1006339502@qq.com

中图分类号:S126

计量

文章访问数:738
HTML全文浏览量:0
PDF下载量:726
被引次数:0

出版历程

收稿日期:2017-11-01
录用日期:2018-02-05
刊出日期:2018-05-01

Comparative study of image segmentation algorithms for rice canopy

HUANG Qiaoyi^{1, 2,},
FAN Xiaolin²,
ZHANG Mu¹,
HUANG Xu¹,
LI Ping¹,
FU Hongting¹,
TANG Shuanhu^1,,
1. Institute of Agricultural Resources and Environment, Guangdong Academy of Agricultural Sciences/Key Laboratory of Plant Nutrition and Fertilizer in Southern Region, Ministry of Agriculture/Guangdong Key Laboratory of Nutrient Cycling and Farmland Conservation, Guangzhou 510640, China
2. South China Agricultural University/Engineering Research Center of Environment Friendly Fertilizer of Guangdong Province, Guangzhou 510642, China
Funds: the Special Fund for Agro-scientific Research in the Public Interest of China201503123
Guangdong Science and Technology Project, China2016A020210035
Guangdong Science and Technology Project, China2014B090904068
the Innovative Research Team in Guangzhou, China2016B070701009

More Information

Corresponding author:TANG Shuanhu, E-mail: 1006339502@qq.com

摘要
HTML全文
图(3)表(1)
参考文献(30)
相关文章
施引文献
资源附件(0)
访问统计

摘要
摘要:水稻冠层信息在自动化管理上具有重要指导意义，但田间多变光照强度环境显著降低了水稻冠层图像分割和信息提取的精度。为降低光照强度的干扰，本文基于RGB、CIEL^*a^*b^*、HSV色彩空间和多色彩空间（包括RGB、CIEL^*a^*b^*和HSV色彩空间）构建水稻冠层图像的色彩特征组合，然后通过支持向量机（SVM）的线性核函数对水稻冠层图像进行分类识别，其分割方法分别定义为rgb-SVM、lab-SVM、hsv-SVM和Multi-SVM。同时，利用此方法对不同光照强度下的水稻冠层图像进行分割，并与常用的ExG&Otsu分割方法进行对比，比较不同方法的分割效果和光强稳健性。结果表明，rgb-SVM的分割效果优于ExG&Otsu方法，但对晴天条件下获取的水稻冠层图像的分割误差仍较大，光强稳健性低；lab-SVM和hsv-SVM分割方法的分割精确度较低，存在一定的欠分割现象；基于多色彩空间和支持向量机的Multi-SVM分割方法的分割效果最佳，该方法对不同光强下获取的水稻冠层图像的分割误差均控制在4.00%以内，具有较好的光强稳健性。因此，基于多色彩空间和支持向量机的Multi-SVM分割方法能够相对准确地将水稻像元从水稻冠层图像中分割出来，且对田间多变光强条件具有一定的稳健性，可为田间水稻生长发育监测和自动化管理提供一定的技术支持。
关键词:水稻/
冠层图像/
光照强度/
图像分割/
色彩空间/
支持向量机
Abstract:Digital image analysis of rice canopy has widely been used for monitoring rice growth, diagnosing rice nitrogen (N) content, controlling pests and predicting rice yield. But the accuracy, stability and reliability of digital image analysis of rice canopy has greatly relied on assumed segmentation precision of rice pixels. There is current a significant progress in auto-segmentation methods for plant images captured indoor or under controlled light conditions. However, it is still hard to segment images of rice canopy taken in outdoor environments with complex and changing illumination conditions. In this paper, we proposed a segmentation method for rice canopy images taken in outdoor environment that improves the accuracy and robustness of illumination of segmentation based on multi-color spaces and support vector machine (SVM) algorithm. The rice canopy images were taken using a digital camera (NikonD90, Nikon Inc., Tokyo, Japan) in August 11st to September 25th 2016 at the largest double-season rice production area in Pearl River Delta. The camera was mounted on a tripod at 1.5 m above rice canopy with straight downward looking posture. Three typical samples taken under different illumination conditions (which changed from sunny days to cloudy days and to overcast days) were treated as test images. The training data (including rice pixels and background pixels) for modeling the support vector machine classifier was randomly picked from the test images. The color features (r, g, b, L^*, a^*, b^*, H, S, V) defined in 3 ordinarily used color spaces (RGB, CIEL^*a^*b^* and HSV) of each pixel were calculated as training data. The SVM classifiers learned from the training data with the color features from RGB, CIEL^*a^*b^*, HSV and multi-color spaces (including RGB, CIEL^*a^*b^*, HSV) were defined as rgb-SVM, lab-SVM, hsv-SVM and Multi-SVM accordingly. The accuracy and robustness of the proposed methods were examined using the test images, which were next compared with ExG&Otsu (excess green index) performance. With the help of Photoshop image editing software, the ground-truth of the rice canopy images was labeled manually and treated as the reference for segmented error calculation, including false positive rate (the rate where segmentation algorithm falsely classed background pixels as rice pixels) and false negative rate (the rate that the segmented algorithm falsely classed the rice pixels as background pixels). The results showed that rgb-SVM algorithm performed better than ExG&Otsu algorithm. While segmentation errors of rgb-SVM algorithm for the images taken on overcast days and cloudy days were respectively 5.76% and 7.74%, that of rgb-SVM algorithm for the images taken on sunny days reached 16.99%. The accuracies of lab-SVM and hsv-SVM algorithms were unstable and high under-segmentation occurred under lab-SVM and hsv-SVM algorithms for images taken on cloudy days and sunny days. Multi-SVM algorithm had the best segmentation results, which were very close to ground-truth images. Specially, segmentation error of Multi-SVM algorithm for images taken on overcast days, cloudy days and sunny days were as low as 3.11%, 3.28% and 3.95%, respectively, which were lower than that for ExG&Otsu algorithm, especially for images taken on sunny days. The results showed that the accuracy of rice canopy extraction using Multi-SVM algorithm was significantly better than that using the other methods, particularly for images taken under high illumination conditions. The Multi-SVM algorithm based on multi-color spaces and support vector machine proposed in this paper accurately segmented and extracted rice pixels in rice canopy images. It was well-suited to the changing illumination in outdoor environment, thus providing valid data support for monitoring field rice growth under natural field conditions and automated rice farming.
Key words:Rice/
Canopy image/
Illumination condition/
Image segmentation/
Color space/
Support vector machine

HTML全文


图1不同光强条件下水稻和背景的色彩参数分布频率图
Figure1.Frequencies of color indices for the rice and background pixels under different illumination conditions


下载: 全尺寸图片幻灯片


图2获取训练图像中水稻冠层色彩特征构建训练数据集
R、G、B为背景或水稻像元的R、G、B色彩特征值, L^*、a^*、b^*为背景或水稻像元的L^*、a^*、b^*色彩参数值, H、S、V为背景或水稻像元的H、S、V色彩参数值。
Figure2.Example of training dataset acquisition from the training images of rice canopy
R, G, B is the R, G, B value of the background or rice pixels; L^*, a^*, b^* is the L^*, a^*, b^* value of the background or rice pixels; H, S, V is the H, S, V value of the background or rice pixels.


下载: 全尺寸图片幻灯片


图3阴天(a)、多云(b)和晴天(c)下水稻冠层测试图像(A)、Photoshop手工分割的参考图像(B)及rgb-SVM方法(C)、lab-SVM方法(D)、hsv-SVM方法(E)、Multi-SVM方法(F)和ExG & Otsu方法(G)对水稻冠层图像分割效果
Figure3.Tested images (A), reference images manually segmented in Photoshop (B), and the segmented results by rgb-SVM (C), lab-SVM (D), hsv-SVM (E), Multi-SVM (F), and ExG & Otsu (G) methods of rice canopy images captured under overcast day (a), cloudy day (b) and sunny day (c)


下载: 全尺寸图片幻灯片

表1不同分割方法对阴天、多云和晴天获取的水稻冠层图像的分割误差
Table1.Segmentation errors of rice canopy images captured under overcast, cloudy, and sunny days based on different segmentation algorithm

天气条件 Weather condition	分割方法 Segmentation algorithm	假正率 False positive (%)	假负率 False negative (%)	分割误差 Segmentation error (%)	Kappa系数 Kappa coefficient
阴天	ExG & Otsu	1.67±0.66a	8.25±0.65a	9.92±0.36a	0.88±0.05d
Overcast days	rgb-SVM	1.52±0.81ab	4.25±0.55d	5.76±0.77c	0.92±0.01b
(n=30)	lab-SVM	1.26±0.59ab	6.46±0.54c	7.71±0.31b	0.91±0.01c
	hsv-SVM	0.85±0.48ab	7.35±0.42b	8.20±0.43b	0.90±0.01c
	Multi-SVM	0.82±0.25b	2.26±0.61e	3.11±0.66d	0.96±0.01a
多云	ExG & Otsu	0.27±0.27d	25.40±3.15a	25.67±3.05a	0.49±0.10c
Cloudy days	rgb-SVM	2.68±0.96b	5.05±1.05c	7.74±1.25d	0.85±0.04a
(n=30)	lab-SVM	0.13±0.30d	13.17±1.93b	13.09±1.53c	0.69±0.07b
	hsv-SVM	3.81±0.58a	14.14±1.52b	17.95±1.45b	0.67±0.08b
	Multi-SVM	1.63±0.85c	1.64±0.45d	3.28±0.62e	0.94±0.02a
晴天	ExG & Otsu	0.30±0.15c	58.40±6.82a	58.70±6.71a	0.19±0.10c
Sunny days	rgb-SVM	0.60±0.52bc	16.39±3.14b	16.99±3.24b	0.60±0.08b
(n=30)	lab-SVM	1.01±0.53ab	5.23±0.93c	6.09±1.06c	0.82±0.06a
	hsv-SVM	1.30±0.50a	12.58±2.51b	13.88±2.79b	0.66±0.07b
	Multi-SVM	1.26±0.34a	2.69±1.10c	3.95±1.15c	0.89±0.07a
同一天气同列不同小写字母表示分割方法间差异达5%显著水平(LSD-test, P < 0.05)。Different lowercase letters in the same column for the same weather condition mean significant differences among different segmentation algorithms at 0.05 level.

下载: 导出CSV

参考文献(30)

[1]	王晓兵, 许迪, 张砚杰, 等.农场规模、劳动力投入量与技术效率及其相关性问题研究[J].资源科学, 2016, 38(3):476-484 http://www.cqvip.com/QK/90831X/201603/668295428.html WANG X B, XU D, ZHANG Y J, et al. The relevance of farm scale, labor inputs and technical efficiency[J]. Resources Science, 2016, 38(3):476-484 http://www.cqvip.com/QK/90831X/201603/668295428.html
[2]	孙良斌, 方向明.农业生产率增长源泉、瓶颈及影响因素——基于南方五省水稻种植户的实证分析[J].华南理工大学学报:社会科学版, 2017, 19(2):22-30 http://paper.usc.cuhk.edu.hk/Details.aspx?id=6658 SUN L B, FANG X M. A probe into sources, bottlenecks and affecting factors of agricultural productivity growth-A case study based on empirical analysis of the survey data of rice farmers in five southern provinces[J]. Journal of South China University of Technology:Social Science Edition, 2017, 19(2):22-30 http://paper.usc.cuhk.edu.hk/Details.aspx?id=6658
[3]	周宏, 王全忠, 张倩.农村劳动力老龄化与水稻生产效率缺失——基于社会化服务的视角[J].中国人口科学, 2014, (3):53-65 http://www.cnki.com.cn/Article/CJFDTotal-ZKRK201403006.htm ZHOU H, WANG Q Z, ZHANG Q. Research on ageing of rural labor force and efficiency loss of rice production:Based on the perspectives of social service[J]. Chinese Journal of Population Science, 2014, (3):53-65 http://www.cnki.com.cn/Article/CJFDTotal-ZKRK201403006.htm
[4]	GUO W, FUKATSU T, NINOMIYA S. Automated characterization of flowering dynamics in rice using field-acquired time-series RGB images[J]. Plant Methods, 2015, 11:7 doi: 10.1186/s13007-015-0047-9
[5]	方伟, 冯慧, 杨万能, 等.基于可见光成像的单株水稻植株地上部分生物量无损预测方法研究[J].中国农业科技导报, 2015, 17(3):63-69 http://www.cnki.com.cn/Article/CJFDTotal-JGSW201501007.htm FANG W, FENG H, YANG W N, et al. Studies on non-destructive optical method for predicting above-ground biomass of individual rice plant based on visible light imaging[J]. Journal of Agricultural Science and Technology, 2015, 17(3):63-69 http://www.cnki.com.cn/Article/CJFDTotal-JGSW201501007.htm
[6]	FANG H L, LI W J, WEI S S, et al. Seasonal variation of leaf area index (LAI) over paddy rice fields in NE China:Intercomparison of destructive sampling, LAI-2200, digital hemispherical photography (DHP), and AccuPAR methods[J]. Agricultural and Forest Meteorology, 2014, 198/199:126-141 https://www.sciencedirect.com/science/article/pii/S0168192314001968
[7]	LEE K J, LEE B W. Estimation of rice growth and nitrogen nutrition status using color digital camera image analysis[J]. European Journal of Agronomy, 2013, 48:57-65 doi: 10.1016/j.eja.2013.02.011
[8]	MU?OZ-HUERTA R, GUEVARA-GONZALEZ R G, CONTRERAS-MEDINA L M, et al. A review of methods for sensing the nitrogen status in plants:Advantages, disadvantages and recent advances[J]. Sensors, 2013, 13(8):10823-10843 doi: 10.3390/s130810823
[9]	贾良良, 范明生, 张福锁, 等.应用数码相机进行水稻氮营养诊断[J].光谱学与光谱分析, 2009, 29(8):2176-2179 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=gpxygpfx200908037 JIA L L, FAN M S, ZHANG F S, et al. Nitrogen status diagnosis of rice by using a digital camera[J]. Spectroscopy and Spectral Analysis, 2009, 29(8):2176-2179 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=gpxygpfx200908037
[10]	李岚涛, 张萌, 任涛, 等.应用数字图像技术进行水稻氮素营养诊断[J].植物营养与肥料学报, 2015, 21(1):259-268 doi: 10.11674/zwyf.2015.0129 LI L T, ZHANG M, REN T, et al. Diagnosis of N nutrition of rice using digital image processing technique[J]. Journal of Plant Nutrition and Fertilizer, 2015, 21(1):259-268 doi: 10.11674/zwyf.2015.0129
[11]	GUIJARRO M, RIOMOROS I, PAJARES G, et al. Discrete wavelets transform for improving greenness image segmentation in agricultural images[J]. Computers and Electronics in Agriculture, 2015, 118:396-407 doi: 10.1016/j.compag.2015.09.011
[12]	周俊, 王明军, 邵乔林.农田图像绿色植物自适应分割方法[J].农业工程学报, 2013, 29(18):163-170 doi: 10.3969/j.issn.1002-6819.2013.18.020 ZHOU J, WANG M J, SHAO Q L. Adaptive segmentation of field image for green plants[J]. Transactions of the CSAE, 2013, 29(18):163-170 doi: 10.3969/j.issn.1002-6819.2013.18.020
[13]	张志斌, 罗锡文, 臧英, 等.基于颜色特征的绿色作物图像分割算法[J].农业工程学报, 2011, 27(7):183-189 http://d.g.wanfangdata.com.cn/Periodical_nygcxb201107032.aspx ZHANG Z B, LUO X W, ZANG Y, et al. Segmentation algorithm based on color feature for green crop plants[J]. Transactions of the CSAE, 2011, 27(7):183-189 http://d.g.wanfangdata.com.cn/Periodical_nygcxb201107032.aspx
[14]	王远, 王德建, 张刚, 等.基于数码相机的水稻冠层图像分割及氮素营养诊断[J].农业工程学报, 2012, 28(17):131-136 doi: 10.3969/j.issn.1002-6819.2012.17.019 WANG Y, WANG D J, ZHANG G, et al. Digital camera-based image segmentation of rice canopy and diagnosis of nitrogen nutrition[J]. Transactions of the CSAE, 2012, 28(17):131-136 doi: 10.3969/j.issn.1002-6819.2012.17.019
[15]	刘亚东, 崔日鲜.基于可见光光谱和随机森林算法的冬小麦冠层图像分割[J].光谱学与光谱分析, 2015, 35(12):3480-3484 http://industry.wanfangdata.com.cn/dl/Detail/Periodical?id=Periodical_gpxygpfx201512051 LIU Y D, CUI R X. Segmentation of winter wheat canopy image based on visual spectral and random forest algorithm[J]. Spectroscopy and Spectral Analysis, 2015, 35(12):3480-3484 http://industry.wanfangdata.com.cn/dl/Detail/Periodical?id=Periodical_gpxygpfx201512051
[16]	WANG Y, WANG D J, SHI P H, et al. Estimating rice chlorophyll content and leaf nitrogen concentration with a digital still color camera under natural light[J]. Plant Methods, 2014, 10:36 doi: 10.1186/1746-4811-10-36
[17]	GUO W, RAGE U K, NINOMIYA S. Illumination invariant segmentation of vegetation for time series wheat images based on decision tree model[J]. Computers and Electronics in Agriculture, 2013, 96:58-66 doi: 10.1016/j.compag.2013.04.010
[18]	MACFARLANE C, OGDEN G N. Automated estimation of foliage cover in forest understorey from digital nadir images[J]. Methods in Ecology and Evolution, 2012, 3(2):405-415 doi: 10.1111/j.2041-210X.2011.00151.x
[19]	黄林生, 刘文静, 黄文江, 等.小波分析与支持向量机结合的冬小麦白粉病遥感监测[J].农业工程学报, 2017, 33(14):188-195 doi: 10.11975/j.issn.1002-6819.2017.14.026 HUANG L S, LIU W J, HUANG W J, et al. Remote sensing monitoring of winter wheat powdery mildew based on wavelet analysis and support vector machine[J]. Transactions of the CSAE, 2017, 33(14):188-195 doi: 10.11975/j.issn.1002-6819.2017.14.026
[20]	梁栋, 管青松, 黄文江, 等.基于支持向量机回归的冬小麦叶面积指数遥感反演[J].农业工程学报, 2013, 29(7):117-123 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=nygcxb201307017 LIANG D, GUAN Q S, HUANG W J, et al. Remote sensing inversion of leaf area index based on support vector machine regression in winter wheat[J]. Transactions of the CSAE, 2013, 29(7):117-123 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=nygcxb201307017
[21]	VAPNIK V N. Statistical Learning Theory[M]. New York:John Wiley & Sons, 1998
[22]	ROBERTSON A R. The CIE 1976 color-difference formulae[J]. Color Research and Application, 1977, 2(1):7-11 doi: 10.1002/col.1977.2.issue-1
[23]	李运奎, 韩富亮, 张予林, 等.基于CIELAB色空间的红葡萄酒颜色直观表征[J].农业机械学报, 2017, 48(6):296-301 doi: 10.6041/j.issn.1000-1298.2017.06.039 LI Y K, HAN F L, ZHANG Y L, et al. Visualization for representation of red wine color based on CIELAB color space[J]. Transactions of the Chinese Society for Agricultural Machinery, 2017, 48(6):296-301 doi: 10.6041/j.issn.1000-1298.2017.06.039
[24]	肖志云, 刘洪.小波域马铃薯典型虫害图像特征选择与识别[J].农业机械学报, 2017, 48(9):24-31 doi: 10.6041/j.issn.1000-1298.2017.09.003 XIAO Z Y, LIU H. Features selection and recognition of potato typical insect pest images in wavelet domain[J]. Transactions of the CSAM, 2017, 48(9):24-31 doi: 10.6041/j.issn.1000-1298.2017.09.003
[25]	杨红颖, 吴俊峰, 于永健, 等.一种基于HSV空间的彩色边缘图像检索方法[J].中国图象图形学报, 2008, 13(10):2035-2038 doi: 10.11834/jig.20081054 YANG H Y, WU J F, YU Y J, et al. Content based image retrieval using color edge histogram in HSV color space[J]. Journal of Image and Graphics, 2008, 13(10):2035-2038 doi: 10.11834/jig.20081054
[26]	CORTES C, VAPNIK V. Support-vector networks[J]. Machine Learning, 1995, 20(3):273-297 http://onlinelibrary.wiley.com/resolve/reference/XREF?id=10.1007/BF00994018
[27]	MIZUSHIMA A, LU R F. An image segmentation method for apple sorting and grading using support vector machine and Otsu's method[J]. Computers and Electronics in Agriculture, 2013, 94:29-37 doi: 10.1016/j.compag.2013.02.009
[28]	HAMUDA E, GLAVIN M, JONES E. A survey of image processing techniques for plant extraction and segmentation in the field[J]. Computers and Electronics in Agriculture, 2016, 125:184-199 doi: 10.1016/j.compag.2016.04.024
[29]	孙涛, 刘振波, 葛云健, 等.基于数码相片Gamma校正的水稻叶面积指数估算[J].生态学报, 2014, 34(13):3548-3557 http://industry.wanfangdata.com.cn/dl/Detail/Periodical?id=Periodical_stxb201413008 SUN T, LIU Z B, GE Y J, et al. Estimation of paddy rice leaf area index based on photo gamma correction[J]. Acta Ecologica Sinica, 2014, 34(13):3548-3557 http://industry.wanfangdata.com.cn/dl/Detail/Periodical?id=Periodical_stxb201413008
[30]	陆秀明, 黄庆, 孙雪晨, 等.图像处理技术估测水稻叶面积指数的研究[J].中国农学通报, 2011, 27(3):65-68 http://www.cnki.com.cn/Article/CJFDTotal-ZNTB201103014.htm LU X M, HUANG Q, SUN X C, et al. Application of image processing technology in rice canopy leaf area index[J]. Chinese Agricultural Science Bulletin, 2011, 27(3):65-68 http://www.cnki.com.cn/Article/CJFDTotal-ZNTB201103014.htm