基于高光谱图像技术的小麦种子分类识别研究
|
摘要
为了探讨高光谱图像技术在小麦种子分类识别中应用的可行性,采集了河南地区主要种植的7个小麦品种的种子高光谱图像及900～1 700nm范围的光谱信息,建立了主成分分析法(PCA)-支持向量机(SVM)分类模型。运用PCA对光谱数据进行降维处理,结合SVM模型比较了不同实验条件下小麦种子分类准确率以及在最佳条件下3个、4个和6个品种种子的分类准确率。结果显示,3个品种间种子分类准确率除个别外平均达到95%以上,4个品种间种子分类准确率在80%左右,6个品种间种子分类准确率在66%左右。这说明充分利用光谱信息可以对3个或4个小麦品种进行多籽粒分类。
In order to apply the hyperspectral imagine technology in classification and recognition of wheat seed,hyperspectral images and spectral information in the range of 900-1 700 nm from wheat seed of seven varieties were collected and extracted in Henan province.A principal component analysis(PCA)-support vector machine(SVM)classification model was established.The spectral data was processed by reducing dimension based on PCA,classification accuracy in different experimental conditions and its optimized classification accuracy in three,four and six varieties were compared by combining with SVM model.The experiment results showed that the average classification accuracy among the three different varieties is above 95% except for some individuals.The classification accuracy among the four varieties is about 80%.The classification accuracy among the six varieties is about66%.The results showed that it is effective and feasible for multi-grain classification of three or four wheat seed varieties by spectral information.
In order to apply the hyperspectral imagine technology in classification and recognition of wheat seed,hyperspectral images and spectral information in the range of 900-1 700 nm from wheat seed of seven varieties were collected and extracted in Henan province.A principal component analysis(PCA)-support vector machine(SVM)classification model was established.The spectral data was processed by reducing dimension based on PCA,classification accuracy in different experimental conditions and its optimized classification accuracy in three,four and six varieties were compared by combining with SVM model.The experiment results showed that the average classification accuracy among the three different varieties is above 95% except for some individuals.The classification accuracy among the four varieties is about 80%.The classification accuracy among the six varieties is about66%.The results showed that it is effective and feasible for multi-grain classification of three or four wheat seed varieties by spectral information.
引文
[1]郭天财.中国北方专用小麦[M].北京:气象出版社,2004:78-84.GUO T C.Wheat in Northern China[M].Beijing:China Meteorological Press,2004:78-84.
[2]宋家永,阎耀礼,周新宝.优质小麦产业化[M].北京:中国农业科学技术出版社,2002:17-23.SONG J Y,YAN Y L,ZHOU X B.Industrialization of high quality wheat[M].Beijing:China Agricultural Science and Technology Press,2002:17-23.
[3]QUAN J G,BAI B,JIN S,et al.Indoor positioning modeling by visible light communication and imaging[J].Chinese Optics Letters,2014,12(5):052201-3.
[4]KHATCHATOURIAN O,PADILHA F R R.Soybean varieties recognition through the digital image processing using artificial neural network[J].Engenharia Agricola,2008,28(4):767.
[5]DANA W,IVO W.Computer image analysis of seed shape and seed color for flax cultivar description[J].Computers and Electronics in Agriculture,2007,61(2):129.
[6]吴静珠,刘倩,陈岩,等.基于近红外与高光谱技术的小麦种子多指标检测方法[J].传感器与微系统,2016,35(7):42.WU J Z,LIU Q,CHEN Y,et al.Multi-index detection method of wheat seed based on NIR and hyperspectrum technology[J].Transducer and Microsystem Technologies,2016,35(7):42.
[7]邓小琴,朱启兵,黄敏.融合光谱、纹理及形态特征的水稻种子品种高光谱图像单粒鉴别[J].激光与光电子学进展,2015,52(2):122.DENG X Q,ZHU Q B,HUANG M.Variety discrimination for single rice seed by integrating spectral,texture and morphological features based on hyperspectral image[J].Laser&Optoelectronics Progress,2015,52(2):122.
[8]朱启兵,冯朝丽,黄敏,等.基于高光谱图像技术和SVDD的玉米种子识别[J].光谱学与光谱分析,2013,33(2):519.ZHU Q B,FENG Z L,HUANG M,et al.Maize seed identification using hyperspectral imaging and SVDD algorithm[J].Spectroscopy and Spectral Analysis,2013,33(2):519.
[9]朱启兵,冯朝丽,黄敏,等.基于图像熵信息的玉米种子纯度高光谱图像识别[J].农业工程学报,2012,28(23):271.ZHU Q B,FENG Z L,HUANG M,et al.Maize seed classification based on image entropy using hyperspectral imaging technology[J].Transactions of the Chinese Society of Agricultural Engineering,2012,28(23):271.
[10]YANG S,ZHU Q B,HUANG M.Application of joint skewness algorithm to select optimal wavelengths of hyperspectral image for maize seed classification[J].Spectroscopy and Spectral Analysis,2017,37(3):994.
[11]许思,赵光武,邓飞,等.基于高光谱的水稻种子活力无损分级检测[J].种子,2016,35(4):38.XU S,ZHAO G W,DENG F,et al.Research on detection technology of rice seed vigor based on hyperspectral[J].Seed,2016,35(4):38.
[12]董高,郭建,王成,等.基于近红外高光谱成像及信息融合的小麦品种分类研究[J].光谱学与光谱分析,2015,35(12):3372.DONG G,GUO J,WANG C,et al.The classification of wheat varieties based on near infrared hyperspectral imaging and information fusion[J].Spectroscopy and Spectral Analysis,2015,35(12):3372.
[13]周望,沈为民,周健康,等.Binning技术在光谱仪中的实验研究[J].激光与光电子学进展,2010,47(4):96.ZHOU W,SHEN W M,ZHOU J K,et al.Experimental research for binning technique in spectroscope[J].Laser&Optoelectronics Progress,2010,47(4):96.
[14]李美凌,邓飞,刘颖,等.基于高光谱图像的水稻种子活力检测技术研究[J].浙江农业学报,2015,27(1):3.LI M L,DENG F,LI Y,et al.Rice seed vigor detection technology based on hyperspectral imagery[J].Acta Agriculturae Zhejiangensis,2015,27(1):3.
[15]余辉,赵晖.支持向量机多类分类算法新研究[J].计算机工程与应用,2008,44(7):188-189.YU H,ZHAO H.New research on multi-classification based on support vector machines[J].Computer Engineering and Applications,2008,44(7):188.
[16]王革丽,杨培才,毛宇清.基于支持向量机方法对非平稳时间序列的预测[J].物理学报,2008,57(2):717.WANG G L,YANG P C,MAO Y Q.On the application of non-stationary time series prediction based on the SVM method[J].Acta Physica Sinica,2008,57(2):717.
[17]邹少奎,殷贵鸿,唐建卫,等.小麦品种周麦22号的分子遗传基础及其特异引物筛选[J].麦类作物学报,2017,37(4):472.ZOU S K,YIN G H,TANG J W,et al.Molecular and genetic basis of wheat variety Zhoumai 22 and specific primers screening[J].Journal of Triticeae Crops,2017,37(4):472.
[18]唐建卫,殷贵鸿,韩玉林,等.栽培措施对周麦27号主要农艺性状及品质特性的影响[J].作物杂志,2013(4):111.TANG J W,YIN G H,HAN Y L,et al.Effect of cultivation measures on agronomic characteristics and quality of Zhoumai 27[J].Crops,2013(4):111.
[2]宋家永,阎耀礼,周新宝.优质小麦产业化[M].北京:中国农业科学技术出版社,2002:17-23.SONG J Y,YAN Y L,ZHOU X B.Industrialization of high quality wheat[M].Beijing:China Agricultural Science and Technology Press,2002:17-23.
[3]QUAN J G,BAI B,JIN S,et al.Indoor positioning modeling by visible light communication and imaging[J].Chinese Optics Letters,2014,12(5):052201-3.
[4]KHATCHATOURIAN O,PADILHA F R R.Soybean varieties recognition through the digital image processing using artificial neural network[J].Engenharia Agricola,2008,28(4):767.
[5]DANA W,IVO W.Computer image analysis of seed shape and seed color for flax cultivar description[J].Computers and Electronics in Agriculture,2007,61(2):129.
[6]吴静珠,刘倩,陈岩,等.基于近红外与高光谱技术的小麦种子多指标检测方法[J].传感器与微系统,2016,35(7):42.WU J Z,LIU Q,CHEN Y,et al.Multi-index detection method of wheat seed based on NIR and hyperspectrum technology[J].Transducer and Microsystem Technologies,2016,35(7):42.
[7]邓小琴,朱启兵,黄敏.融合光谱、纹理及形态特征的水稻种子品种高光谱图像单粒鉴别[J].激光与光电子学进展,2015,52(2):122.DENG X Q,ZHU Q B,HUANG M.Variety discrimination for single rice seed by integrating spectral,texture and morphological features based on hyperspectral image[J].Laser&Optoelectronics Progress,2015,52(2):122.
[8]朱启兵,冯朝丽,黄敏,等.基于高光谱图像技术和SVDD的玉米种子识别[J].光谱学与光谱分析,2013,33(2):519.ZHU Q B,FENG Z L,HUANG M,et al.Maize seed identification using hyperspectral imaging and SVDD algorithm[J].Spectroscopy and Spectral Analysis,2013,33(2):519.
[9]朱启兵,冯朝丽,黄敏,等.基于图像熵信息的玉米种子纯度高光谱图像识别[J].农业工程学报,2012,28(23):271.ZHU Q B,FENG Z L,HUANG M,et al.Maize seed classification based on image entropy using hyperspectral imaging technology[J].Transactions of the Chinese Society of Agricultural Engineering,2012,28(23):271.
[10]YANG S,ZHU Q B,HUANG M.Application of joint skewness algorithm to select optimal wavelengths of hyperspectral image for maize seed classification[J].Spectroscopy and Spectral Analysis,2017,37(3):994.
[11]许思,赵光武,邓飞,等.基于高光谱的水稻种子活力无损分级检测[J].种子,2016,35(4):38.XU S,ZHAO G W,DENG F,et al.Research on detection technology of rice seed vigor based on hyperspectral[J].Seed,2016,35(4):38.
[12]董高,郭建,王成,等.基于近红外高光谱成像及信息融合的小麦品种分类研究[J].光谱学与光谱分析,2015,35(12):3372.DONG G,GUO J,WANG C,et al.The classification of wheat varieties based on near infrared hyperspectral imaging and information fusion[J].Spectroscopy and Spectral Analysis,2015,35(12):3372.
[13]周望,沈为民,周健康,等.Binning技术在光谱仪中的实验研究[J].激光与光电子学进展,2010,47(4):96.ZHOU W,SHEN W M,ZHOU J K,et al.Experimental research for binning technique in spectroscope[J].Laser&Optoelectronics Progress,2010,47(4):96.
[14]李美凌,邓飞,刘颖,等.基于高光谱图像的水稻种子活力检测技术研究[J].浙江农业学报,2015,27(1):3.LI M L,DENG F,LI Y,et al.Rice seed vigor detection technology based on hyperspectral imagery[J].Acta Agriculturae Zhejiangensis,2015,27(1):3.
[15]余辉,赵晖.支持向量机多类分类算法新研究[J].计算机工程与应用,2008,44(7):188-189.YU H,ZHAO H.New research on multi-classification based on support vector machines[J].Computer Engineering and Applications,2008,44(7):188.
[16]王革丽,杨培才,毛宇清.基于支持向量机方法对非平稳时间序列的预测[J].物理学报,2008,57(2):717.WANG G L,YANG P C,MAO Y Q.On the application of non-stationary time series prediction based on the SVM method[J].Acta Physica Sinica,2008,57(2):717.
[17]邹少奎,殷贵鸿,唐建卫,等.小麦品种周麦22号的分子遗传基础及其特异引物筛选[J].麦类作物学报,2017,37(4):472.ZOU S K,YIN G H,TANG J W,et al.Molecular and genetic basis of wheat variety Zhoumai 22 and specific primers screening[J].Journal of Triticeae Crops,2017,37(4):472.
[18]唐建卫,殷贵鸿,韩玉林,等.栽培措施对周麦27号主要农艺性状及品质特性的影响[J].作物杂志,2013(4):111.TANG J W,YIN G H,HAN Y L,et al.Effect of cultivation measures on agronomic characteristics and quality of Zhoumai 27[J].Crops,2013(4):111.