0 引言
高光谱成像技术是20世纪80年代从遥感技术发展而来的一种快速、无损的检测技术,该技术可在不破坏样品的条件下进行观测,不仅可以迅速测定样品中的各个成分,而且可通过图像反映出样品的形状及纹理特征,故高光谱成像技术在无损检测方面具有较大的应用前景[1,2,3]。种子是农业生产中最重要的生产资料,是现代农业发展的核心要素,各国都把种子的改良、繁育、推广作为发展农业生产的战略措施,未来国际市场农产品的竞争在很大程度上是技术竞争,而种子生产是技术含量较高的生产因素。 近年来,高光谱技术在农业中的应用研究迅速发展,在土壤方面主要应用于对水分、重金属、土壤有机质以及营养元素含量的测定上[4,5,6],农作物的研究集中在作物的品质方面较多[7,8,9],而在种子方面,该技术则重点在其品种鉴别、活力测定及品质检测等方面[10,11],基于高光谱技术在种子应用上的波谱范围、模型方法及作物种子样本量等有所不同,为推进该技术在作物种子方面的研究与应用,本文从波段范围(可见光、可见光—短波近红外及长波近红外)、种子活力(出苗、出芽)及种子品质(外在品质和内在品质)等方面进行了较为详尽的叙述,总结诸多学者研究成果的异同,以期为加快高光谱技术在现代种植业的生产与实践步伐,为提升种子市场竞争力提供有益的参考。
1 种子品种鉴别
我国主要三大粮食作物为水稻、小麦、玉米,根据中国统计年鉴分析,2017年水稻、小麦和玉米的产量占全国粮食产量的32.15%,20.30%和39.16%。因此粮食产量对稳定社会、发展经济的作用显得尤为重要,除此之外,某些经济作物同样影响着农村的经济收入,并且是人类生存最基本、最必需的生活资料。但近年来随着杂交品种的日益增多,市面上掺假种子越来越多,因此其品种的快速、准确识别对生产来说就具有至关重要的意义,传统的种子品种鉴别方法主要有人工检测、田间检测、形态学方法、荧光扫描鉴定法、化学鉴定法和电泳鉴定法,但由于这些方法存在测定过程较为繁琐、耗时长、重复性差等缺点[12,13,14,15],故需要寻找一种新的技术——高光谱成像技术来进行鉴别。
1.1 基于偏最小二乘判别法的种子品种鉴别
在研究初期,主要利用高光谱技术对三大主要粮食作物进行研究,在小麦种子的鉴别研究方面, Miralbés[16]利用高光谱技术对欧洲12种小麦采用WINISIⅢ软件进行偏最小二乘法(partial least squares,PLS)判别分析,得到模型测试集的识别精度为99.5%,从而实现对小麦特定品种的鉴别。
对玉米种子的鉴别研究同样很多,初期主要研究该模型是否有利于玉米种子品种的鉴别,黄敏等[17]利于全波段结合PLS对9个品种的玉米种子进行分类,此模型可实现玉米品种的准确分类,在探究了模型的适应性之后,现阶段主要研究如何提高模型的精度; 吴翔等[18]对4个品种共384个玉米种子样本建立了连续投影算法(successive projections algorithm,SPA)与偏最小二乘判别分析(PLS-DA)相结合的分类模型,建模集和预测集的总体识别率分别为78.5%和70.8%; 魏立峰[19]在此基础上对图像的去噪方法进行了更新,提出了基于Contourlet变换和阈值函数的高光谱图像去噪模型,使玉米种子的训练集和测试集精度分别达到97.77%和90.80%; 王庆国等[20]对玉米种子产地和年份进行鉴别,得出训练集和测试集精度分别为99.11%和98.39%,为未来的研究提供了一种新的方向。
可见,高光谱成像技术在农作物种子品种鉴别方面研究较多,尤其是PLS-DA模型的应用较广,而在经济作物和小宗作物种子上的应用较少,大宗作物的研究应用上因样本量、试验条件和后期处理方法的不同也有所不同。
1.2 基于机器学习算法的种子品种鉴别
1.2.1 粮食作物种子
随着计算机技术的发展,现阶段的研究主要集中在利用机器学习算法对种子品种进行鉴别,不同模型在种子品种鉴别的适宜性研究中,丁秋[25]利用全波段、可见光、近红外短波波段3种不同光谱信息对逐步判别分析(stepwise discriminant analysis,SDA)、K均值聚类分析(K-means cluster analysis,K-means)、支持向量机(support vector machine,SVM)小麦籽粒品种鉴别模型进行了对比,结果表明SDA模型效果明显较优,鉴别的识别率达到了88.7%; 张航等[10]进一步研究了相同模型在不同品种数的适宜性,基于主成分分析法(principal component analysis,PCA)- SVM分类模型对多个品种的小麦进行识别,结果显示,在3个、4个、6个小麦品种间的识别率分别达到了95%,80%和56%。
1.2.2 其他作物种子
在模型适宜性研究方面,张初等[29]研究4种不同的模型对西瓜种子进行分类,发现利用极限学习机(extreme learning machine,ELM)模型判别的效果最佳,建模集和预测集的判别正确率均为100%。在模型优化方面,主要分为对图像和光谱的优化处理。在图像的研究方面,邹伟等[30]利用PCA提取出3个特征波长,并对其图像进行纹理特征提取,模型的建模集和预测集精度分别为93.75%和91.67%。在光谱的研究方面,Tan等[31]利用PCA对大豆种子的光谱进行降维,并结合ANN建模方法对种子进行分类,精度可达到97.5%; Hashim等[32]发现图像处理与光谱处理相结合对3种橡胶树种子进行分类可得到较好的精度,且利用PCA方法对38个特征进行降维后,识别精度提高了14个百分点; 程术希等[33]通过对大白菜种子光谱数据的采集,发现了采用载荷系数的特征波长选择方法剔除大量冗余数据可取得很好的效果,并在此基础上通过Ada-Boost,ELM,RF和SVM这4种分类算法进行分类判别,4种模型鉴别精度均达到了90%以上,其中ELM和RF模型识别效果达到了100%。
综上所述,上述研究为常见的粮食作物及经济作物种子品种鉴别提供了最佳模型参考,在未来的研究中,研究者可根据不同研究对象从而选用最佳模型,但以上模型在以后的发展中仍需要不断更新完善,以保证模型的精度与适用性。
2 种子活力检测
种子活力是评价种子质量的重要指标,高活力的种子在田间表现出出芽率高、植株强壮、产量高等优点,但是传统的活力检测方法为四唑染色实验、电导率实验和幼苗生长测定实验等,存在操作过程复杂繁琐、耗时长、重复性差且对种子有破坏性等缺点。因此,快速无损的高光谱成像技术检测将是未来的发展趋势。
2.1 可见光波段
在可见光波段下,前人研究主要集中在能否区分出无活力种子,Mo等[34]采用高光谱成像技术对辣椒种子的光谱进行了采集,结果表明,在600~700 nm波段范围下使用红光光源,并且结合PLS-DA模型对辣椒种子的活力识别效果最好,对无活力种子的识别率达到了100%; 在684 nm波段下采用荧光光谱图像对有活力黄瓜种子的识别率达到了99%[35]。张婷婷等[36]在430~970 nm波段范围下采集小麦种子的光谱图像,并对数据预处理方法进行了筛选,建立全波段PLS-DA模型,结果得出,均值中心化(mean centering,MC)预处理方法建立的PLS-DA模型表现最优,经过此模型筛选后,可使小麦种子的最终发芽率达到93.1%。
2.2 可见光—短波近红外波段
在可见光—短波近红外波段下,利用高光谱成像技术对植物的种子活力进行了大量研究, Tigabu等[37]通过人工加速老化的方法获得活力不同的松树种子,对老化与未老化的松树种子识别率高达100%; 李美凌等[38]在Tigabu等研究基础上,对400~1 000 nm波段范围下的水稻种子的光谱图像进行采集,通过PCA分析获得主成分图像,并确定特征波段,结果表明,SVM模型预测的判别率可达100%。除了探究模型的适用性之外,前人还针对不同模型进行了一系列的优化,尤佳[39]对不同活力的脱绒棉种采用PCA进行降维,建立判别分析和SVM模型,其中SVM中采用K折交叉验证(K-fold cross-validation,K-CV)和粒子群算法(particle swarm optimization, PSO)优化参数,结果表明判别分析模型效果最优,其活力判别精度可达到89.70%; 黄蒂云[40]提取出长、宽、面积和圆形度等12个脱绒棉种的形态特征,并利用PLS-DA模型进行建模,将该模型的识别精度从93%提升到 96%。因此可以看出,后者的研究方法对不同活力的脱绒棉种分类适宜性较好; Soltani等[41]应用近红外检测技术进行种子活力检测,研究结果表明,近红外光谱技术在鉴定单粒山毛榉树种子是否有活力方面的精确度达到 100%; Ambrose等[42]在400~2 500 nm波段范围下鉴别微波热处理后及未处理的2种活力水平玉米种子,所建立的PLS-DA模型校准集和预测集的鉴别率分别高达97.6%和95.6%; Zhang等[43]对单个小麦种子的两侧进行高光谱成像,得到了4个数据集分别为腹沟侧、反向侧、平均值(每粒种子两侧光谱的平均值)和混合数据集(每粒种子两侧光谱),并利用PLS-DA和SVM技术,建立了种子的分类模型,结果表明,标准正态变量SPA-PLS-DA模型对整粒种子(>85.2%)和活种子(>89.5%)具有较高的分类精度。
2.3 长波近红外波段
研究者对玉米种子也进行了许多研究,杨冬风等[46]利用近红外光谱和BP神经网络对玉米胚乳区域进行光谱采集,并建立玉米种子活力智能检测模型,模型识别的准确率为95.0%; Wakholi等[47]对玉米种子的光谱图像进行了采集并采用线性判别分析(linear discriminant analysis,LDA)、PLS-DA和SVM这3种分类模型进行了建模,结果表示,SVM分类模型得到的最高识别率高达100%,比以往研究的基于PLS的方法提升了5百分点; Ambrose等[48]基于傅立叶变换近红外和拉曼光谱技术采用PCA和PLS对玉米种子活力进行评估,结果表明,傅立叶变换近红外光谱法可有效检测种子活力,其准确率高达100%,预测能力超过95%。
综上,本文探讨了在不同波段范围内对于检测不同作物种子是否具有活力及活力等级分类可选取的最佳模型,旨在未来的研究中,研究者可根据自身的条件选取不同的研究波段(表1)。但由于上述研究均采用的是人工老化的方法,因此在未来的研究中,可根据自然条件筛选出更多老化程度的种子,用于模型的完善与更新。
表1 高光谱成像在种子活力检测中的应用研究概况
Tab.1
| 波段 | 种子类别 | 波段采集 范围/nm | 建模方法 | 参考文献 |
|---|---|---|---|---|
| 可见光 波段 | 辣椒 | 600~700 | PLS-DA | Mo等[34] |
| 黄瓜 | 425~700 | 多光谱算法 | Mo等[35] | |
| 小麦 | 430~970 | PLS-DA | 张婷婷等[36] | |
| 可见光—短波近红外波段 | 松树 | 850~1 048 | PLS-DA | Tigabu等[37] |
| 水稻 | 400~1 000 | SVM | 李美凌等[38] | |
| 脱绒棉 | 400~1 000 | SVM | 尤佳[39] | |
| 脱绒棉 | 400~1 000 | PLS-DA | 黄蒂云[40] | |
| 毛榉树种 | 400~2 498 | PLS-DA | Soltani等[41] | |
| 玉米 | 400~2 500 | PLS-DA | Ambrose等[42] | |
| 小麦 | 400~1 000 | PLS-DA | Zhang等[43] | |
| 长波近红外波段 | 大麦、小麦、高粱 | 1 000~2 498 | PLS-DA | McGoverin等[44] |
| 甜瓜 | 948~2 494 | PLS-DA | Kandpal等[45] | |
| 玉米 | 833~2 500 | 神经网络 | 杨冬风等[46] | |
| 玉米 | 1 000~2 500 | SVM | Wakholi等[47] | |
| 玉米 | 1 000~2 500 | PLS-DA | Ambrose等[48] | |
| 水稻 | 874~1 740 | PLS-DA | 许思等[49] | |
| 水稻 | 874~1 734 | SVM | 吴小芬等[50] |
3 种子品质检测
3.1 种子外在品质
在1976年,Norris等[53]首先采用近红外反射光谱对农产品的含水率开始研究,从而推广出近红外反射光谱技术在种子的生化物质含量、真实性、缺陷种子等方面的研究。在种子是否具有杂质且是否损伤方面,Wallays等[54]在2009年对大麦、小麦和玉米进行了杂质检测。通过遗传算法,在400~950 nm波段之间可以较好地检测出其中的杂质,并且成功建立了这3种谷物的杂质检测系统; Agelet[55]对采集的玉米籽粒的近红外光谱进行分析,对高温损伤和冷冻伤害的玉米种子进行了模型的筛选,结果表明采用PLS模型对热损伤的玉米检测效果最好,结果达到了99%; Kiratiratanapruk等[56]在Agelet的基础上结合颜色和纹理特征对玉米种子进行了识别,并采用SVM方法对种子进行了检测,对正常种子和缺陷种子的识别度可达到95.6%和80.6%; 王超鹏等[57]提取全表面颜色特征参数训练分类器进行霉变玉米种子检测,并使用朴素贝叶斯分类器进行霉变种子识别准确率达到了98%,提取种子尖端颜色特征参数训练分类器,并使用朴素贝叶斯分类器检测的准确率达到了99.3%。
3.2 种子内在品质
在检测种子是否被侵染方面, Williams等[60]采用了新的探索性主成分分析方法(PCAL)降噪和阴影,并采用PLS-DA模型进行建模,此方法可以很好地识别出感染细菌的种子; Xing等[61]通过研究得出,在720 nm以上的波段可鉴别小麦种子是否霉变,在728~878 nm波段范围可将发芽的小麦种子从健康的种子中识别出来,识别率可达100%; Del Fiore等[62]首次采用了高光谱成像技术识别真菌侵染玉米籽粒,证明了可见光—近红外高光谱技术用于检测真菌侵染玉米的可行性; Singh等[59]在Del Fiore的基础上,采用多变量图像分析方法对高光谱数据进行降维处理,利用PCA选择有效波长,对被真菌侵染的小麦的识别精度达到了100%。
在检测种子是否被特定的真菌侵染方面,Yao等[63]发现紫外线可以使感染黄曲霉毒素的玉米种子向蓝光波段产生不同程度的偏移,且种子中黄曲霉毒素的浓度越高,荧光峰值越低,为未来的研究提供了科学的参考; 袁莹等[64]同样利用高光谱成像技术对玉米籽粒的黄曲霉毒素进行了检测,采用PCA分析对高光谱数据进行降维,并采用因子判别分析(factor discriminant analysis,FDA)进行建模,训练集和验证集分别达到了95%和86%; Shahin等[65]利用400~1 000 nm波段范围高光谱成像系统对小麦种子是否被镰刀菌侵染检测研究,并建立线性判别分析预测模型,识别率可达到92%,对受侵染种子的预测可达到86%。在此基础上, Barbedo等[66]提出了一种基于高光谱的小麦籽粒中镰刀菌的自动检测算法,在528~1 785 nm波段范围下采用阈值分割法对图像进行了预处理,使籽粒与背景分离,最后判断其感染赤霉病的程度,该算法不仅能检测赤霉病,也可以评估脱氧雪腐镰刀菌含量。
3.3 种子元素含量
Velasco等[67]采集了451个油菜籽粒的近红外光谱用于籽粒蛋白质含量的检测,并将原始光谱进行一阶导数、标准正态变量校正、去趋势散射校正之后进行建模分析,结果可知,近红外光谱与油菜籽粒蛋白质含量之间的确定系数为0.94; Cogdill等[68]采用PCA和PLS等分析方法对玉米种子的含油率和含水率进行了检测,得到了较好的结果; Weinstock等[69]利用950~1 700 nm波段的近红外光谱图像和多变量分析,采用PLS方法预测玉米籽粒的油分和油酸含量,油分含量预测均方根误差为0.7,油酸含量的预测偏差为14%; Font等[70]采集了芥菜、油菜和鹰嘴豆等种子籽粒的高光谱数据用于测定种子的纤维、脂肪酸组成、蛋白质和油脂含量,并且采用PLS法进行建模,模型有较好的预测结果,但某些化学成分的预测结果较差; Tallada等[71]采集玉米籽粒的光谱特征对籽粒中的粗蛋白、色氨酸、赖氨酸、油分和可溶性糖含量通过PLS方法进行建模,结果表明,该模型对蛋白质含量的检测效果较好,R2达到了0.75,但是该模型对色氨酸、赖氨酸、油分含量和可溶性糖的预测效果较差。
4 讨论
高光谱技术在种子方面有较多应用,本文主要总结出高光谱技术在种子品种鉴别、种子活力检测以及种子品质检测方面的应用(表2),通过总结发现研究主要集中在玉米、小麦和水稻等粮食作物,但对咖啡、甜菜、烟叶等经济作物以及人参、枸杞等药类研究较少。
表2 作物种子的测定内容及使用光谱仪类型
Tab.2
| 年份 | 测定内容 | 光谱仪 | 光谱仪 产地 | 波长/nm |
|---|---|---|---|---|
| 2008年 | 小麦的品种鉴别 | Foss-NIRSystem 6500 | 美国 | 400~2 500 |
| 2017年 | 小麦的品种鉴别 | ICL-B1610M-SC000 | 美国 | 388~1 000 |
| 2019年 | 小麦的品种鉴别 | PFD-65-V10E | 芬兰 | 900~1 700 |
| 2012年 | 玉米的品种鉴别 | 1003A-10140HyperspcTMVNIRC-Series | 美国 | 563.6~911.4 |
| 2014年 | 玉米的品种鉴别 | 1003A-10140HyperspcTMVNIRC-Series | 美国 | 400~1 000 |
| 2008年 | 水稻的品种鉴别 | Handheld FieldSpec | 美国 | 350~1 075 |
| 2013年 | 水稻的品种鉴别 | ImspectorN17E | 芬兰 | 1 039~1 612 |
| 2015年 | 水稻的品种鉴别 | 1003A-10140HyperspcTMVNIRC-Series | 美国 | 400~1 000 |
| 2016年 | 水稻的品种鉴别 | RT100-HP激光诱导击穿光谱仪 | 美国 | 190~1 040 |
| 2018年 | 水稻的品种鉴别 | XDS Rapid Content TM | 丹麦 | 400~2 498 |
| 2011年 | 油菜籽的品种鉴别 | ImSpectorV10E-QE | 芬兰 | 400~1 000 |
| 2013年 | 西瓜种子的品种鉴别 | ImSpectorN17E-QE | 芬兰 | 800~1 800 |
| 2014年 | 大豆的品种鉴别 | ImSpectorV25E | 芬兰 | 1 000~2 500 |
| 2014年 | 大白菜的品种鉴别 | ImSpectorN17E-XLNIR | 芬兰 | 874~1 734 |
| 2011年 | 大麦、小麦、高粱的种子活力 | ImSpectorN17E-QE | 芬兰 | 1 000~2 498 |
| 2018年 | 小麦的种子活力 | ImSpectorV10E | 芬兰 | 400~1 000 |
| 2019年 | 小麦的种子活力 | ImSpectorV10E | 芬兰 | 400~970 |
| 2013年 | 玉米的种子活力 | VECTOR22/N | 德国 | 833~2 500 |
| 2016年 | 玉米的种子活力 | HeadwellPhotonics,Fitchburg,MA | 美国 | 400~2 500 |
| 2018年 | 玉米的种子活力 | HeadwallPhotonics,Fitchburg,MA | 美国 | 1 000~2 500 |
| 2018年 | 玉米的种子活力 | 赛默飞(Antaris)II型傅立叶变换近红外光谱仪 | 美国 | 1 000~2 500 |
| 2015年 | 水稻的种子活力 | ImSpectorV10E-QE | 芬兰 | 400~1 000 |
| 2016年 | 水稻的种子活力 | ImSpectorN17E-QE | 芬兰 | 874~1 740 |
| 2017年 | 水稻的种子活力 | ImSpectorN17E | 芬兰 | 874~1 734 |
| 2017年 | 脱绒棉种的种子活力 | ImSpectorV10E -QE | 芬兰 | 400~1 000 |
| 2018年 | 脱绒棉种的种子活力 | ImSpectorV10E -QE | 芬兰 | 400~1 000 |
| 2003年 | 松树的种子活力 | FOSS Tecator | 瑞典 | 850~1 048 |
| 年份 | 测定内容 | 光谱仪 | 光谱仪 产地 | 波长/nm |
| 2003年 | 毛榉树的种子活力 | FOSS NIR Systems | 美国 | 400~2 498 |
| 2014年 | 辣椒的种子活力 | VNIR,HeadwallPhotonics,Fitchburg,MA | 美国 | 400~700 |
| 2015年 | 黄瓜的种子活力 | VNIR,HeadwallPhotonics,Fitchburg,MA | 美国 | 425~700 |
| 2016年 | 甜瓜的种子活力 | SWIR,HeadwallPhotonics,Fitchburg,MA | 美国 | 948~2 494 |
| 2009年 | 大麦、小麦、玉米的杂质检测 | SpecimV10 spectrograph | 芬兰 | 400~950 |
| 2010年 | 小麦的受损种子 | ImSpectorV10E | 芬兰 | 400~1 000 |
| 2011年 | 小麦的镰刀菌 | VNIR100E | 美国 | 400~1 000 |
| 2015年 | 小麦的镰刀菌 | EV/NIR HyperspecModel 1003B-10151 | 美国 | 528~1 785 |
| 2015年 | 小麦的灰绿曲霉、青霉、赭曲霉毒素 | Model No.SU640-1.7RT-D | 美国 | 1 000~1 600 |
| 2012年 | 玉米的冻害种子 | PertenDA7200 | 瑞典 | 860~1 640 |
| 2016年 | 玉米的受损种子 | ImSpectorV10E-QE,ImSpectorN25E | 芬兰 | 400~1 000 |
| 2009年 | 玉米的镰刀菌 | SWIR | 芬兰 | 960~1 662 |
| 2010年 | 玉米的种子霉变 | ImSpector-SpecimV10 | 芬兰 | 400~1 000 |
| 2010年 | 玉米的黄曲霉毒素 | AvaSpec-ULS2048-USB2 | 荷兰 | 400~1 000 |
| 2014年 | 玉米的黄曲霉毒素 | SpecimV10MspectrographOulu | 芬兰 | 400~1 000 |
| 2017年 | 玉米的种子真菌侵染 | ImspectorV10E | 芬兰 | 400~1 000 |
| 2004年 | 玉米的含油率和含水率 | ApogeeKX-260 Monochrome Scientific Camera | 美国 | 750~1 090 |
| 2006年 | 玉米的油分和油酸 | Matrix NIR | 美国 | 950~1 700 |
| 2009年 | 玉米的粗蛋白、色氨酸、赖氨酸、油分和可溶性糖 | CDI Spectrometer | 美国 | 904~1 685 |
| 2015年 | 水稻的种子霉变 | JFE,Techno-ResearchCorporation | 日本 | 400~1 000 |
| 2006年 | 油菜、鹰嘴豆、芥菜的纤维、脂肪酸组成、蛋白质和油脂 | NIR Systems Model6500 Spectrophotometer | 美国 | 400~2 500 |
| 2014年 | 油茶的脂肪酸、油酸、亚油酸和棕榈酸 | FieldSpecHH2 | 美国 | 325~1 075 |
近年来,随着基因技术和杂交的不断发展,高光谱技术在快速识别转基因种子和杂交种子的应用中得到越来越多的关注,翟亚锋等[72]通过实验得出近红外反射光谱技术鉴别转基因小麦,采用PCA结合仿生模式识别准确率达到95%以上; Feng等[73]利用高光谱成像技术鉴定转基因玉米的种类,利用全光谱或最佳波长建立判别模型,结果表明,高光谱成像技术可用于鉴定转基因和非转基因玉米籽粒; 林萍等[74]提出一种利用近红外光谱技术快速鉴别Bt基因水稻种子及其亲本的新方法,采用等距映射流形降维法对采集到的光谱数据进行非线性降维,提取出45个特征波长,并且利用最小二乘SVM方法进行建模,其中预测的准确度高达94.67%。因此,转基因及杂交作物种子的各项指标的研究将成为新的研究趋势。
5 结论与展望
在数据获取方面,主要有获取光谱特征和图像特征2种方式,光谱特征相对于图像特征而言,获取方式较为方便,根据不同特征建立的模型性能存在一定差异,但差异不大。在种子分类方面,主要应用偏最小二乘法(partial least squaress,PLS)、Ada-Boost算法、极限学习机(extreme learning machine,ELM)、随机森林(random forest,RF)、支持向量机(support vector machine,SVM)和人工神经网络(artificial neural network,ANN)等模型进行建模,这些模型都取得了较好的成效。在未来,为了保证模型的有效性和通用性,需要对模型进行不断更新。
除此之外,应采用大量样本,建立种子品种分类和分级的数据库,在未来的鉴别过程中,可根据不同的实际情况,选择出最佳的光谱采集波段与建模模型。在种子活力研究方面,主要研究将健康种子与老化种子和低活力种子分开,但是由于自然衰老种子较难获取,故研究常采用人工方法老化种子,而且在实践应用中,自然老化样品的采集时间跨度较大,不同的自然老化条件也会影响样品的特性。因此,建立一个使用自然老化样品进行种子活力和活力检测的通用数据库难度较大,离实际应用还有一定距离。除此之外,在研究高光谱技术检测种子是否被真菌损伤时,上述研究仍有部分缺陷,如某些研究的假定真菌浓度远远低于实际浓度,且研究需要大量样本来建立一个通用的模型。
参考文献
高光谱成像技术及其在农产品检测中的应用
[J].
Hyperspectral imaging technology and its application in agricultural product detection
[J].
Indoor positioning modeling by visible light communication and imaging
[J].
Computer image analysis of seed shape and seed color for flax cultivar description
[J].DOI:10.1016/j.compag.2007.10.001 URL [本文引用: 1]
高光谱小波能量特征估测土壤有机质含量
[J].
Determination of soil organic matter content based on hyperspectral wavelet energy features
[J].
基于高光谱特征波长的冬小麦水分含量估测模型
[J].
Study on water content estimation model of winter wheat based on hyperspectral characteristic wavelength
[J].
华南地区土壤全钾含量高光谱反演模型研究
[J].
Hyperspectral inversion model of soil total potassium content in south China
[J].
基于高光谱遥感的农作物分类研究进展
[J].
Research advances on crop identification using hyperspectral remote sensing
[J].
基于无人机高光谱数据的玉米叶面积指数估算
[J].
Estimation of corn leaf area index based on UAV hyperspectral image
[J].
利用叶面积指数优化冬小麦高光谱水分预测模型
[J].
Optimizing winter wheat hyperspectral moisture prediction model using leaf area index
[J].
基于高光谱图像技术的小麦种子分类识别研究
[J].
Research on wheat seed classification and recognition based on hyperspectral imaging
[J].
种子活力测定方法研究进展
[J].
Research progress of seed vigor test methods
[J].
Use of multispectral imaging in varietal identification of Tomato
[J].
Assessment of seed quality using non-destructive measurement techniques:A review
[J].DOI:10.1017/S0960258516000234 URL [本文引用: 1]
基于电子鼻传感器阵列优化的甜玉米种子活力检测
[J].
Vigor detection of sweet corn seeds by optimal sensor array based on electronic nose
[J].
基于近红外光谱的单粒水稻种子活力快速无损检测
[J].
Qualitative analysis of single rice seed vigor using near infrared reflectance spectroscopy
[J].
Discrimination of European wheat varieties using near infrared reflectance spectroscopy
[J].
基于高光谱图像的玉米种子特征提取与识别
[J].
Morphological characteristics of maize seed extraction and identification based on the hyperspectral image
[J].
基于高光谱技术的玉米种子可视化鉴别研究
[J].
Study on visual identification of corn seeds based on hyperspectral imaging technology
[J].
玉米种子高光谱图像品种检测方法研究
[D].
Research on detection method of maize variety based on hyperspectral image
[D].
基于高光谱图像的玉米种子产地与年份鉴别
[J].
Geographical origin and years identification of maize seeds based on the hyperspectral image
[J].
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].
Variety discrimination for single rice seed by integrating spectral,texture and morphological features based on hyperspectral image
[J].
水稻种子鉴别的近红外光谱快速无损分析
[D].
Rapid and nondestructive analysis for the identification of rice seeds with near infrared spectroscopy
[D].
马铃薯黑心病和单薯质量的透射高光谱检测方法
[J].
Transmission hyperspectral detection method for weight and black heart of potato
[J].Potatos are one of the world's major food crops. It not only has medicinal value and food value, but also has industrial value. The quality of potatos is directly related to their commodity level, benefits, and market competitiveness. Therefore, its quality testing is an important part of potato processing. Currently, common non-destructive testing techniques (near infrared spectroscopy and machine vision technology) are unable to achieve simultaneous detection of a potato's internal and external quality. Transmission hyperspectral imaging technology has some penetrating ability, when the light passes through the agricultural products, spectral and image of hyperspectral imaging data will change according to the differences in their internal characteristics. Therefore, the transmission hyperspectral imaging technology not only can detect the internal quality of agricultural products, but also can detect some external qualities. Since the single detection technology cannot simultaneously detect the internal and external quality of potatoes, the internal black heart and external weight of potatoes are detected by the transmission hyperspectral imaging technology and fusing spectrum and image information. In this study, 266 hyperspectral images (400-1 000 nm) were collected by the transmission hyperspectral imaging system, and then the spectrum and the image information were extracted. Using a Monte Carlo cross-validation method to exclude the data of two abnormal black heart samples, and variable selection methods of uninformative variable elimination (UVE) and successive projections algorithm (SPA) were used to do the variable selection for the spectrum of the black heart sample. The eventual adoption of 9 spectral variables were used to establish the detection model of black heart by a partial least squares discriminant analysis (PLS-DA); variable selection methods of competitive adaptive reweighed sampling (CARS) and successive projections algorithm (SPA) were used to do variable selection for a weight testing sample spectrum, the eventual adoption of 9 variables established a detection model of weight testing by partial least-squares regression (PLS); the Area information of transmission hyperspectral image was extracted, which combined with the 9 spectral variables to set up an PLS model for weight detection based on spectral - image information. The research demonstrates that the accurate recognition rate of black heart is 100%, and the minimum shoddy area which could be identified was 1.88 cm2. The performance of the weight detection model based on the spectrum-image (10 variables) is much better than the one based on the spectrum (9 variables), the prediction correlation coefficient (Rp) was 0.99, and the forecast root mean square error (RMSEP) was 10.88. The results indicate that using the transmission hyperspectral imaging technology with the fusion of image and spectrum information to detect potatoes' internal black heart and external weight simultaneously is feasible.
基于近红外高光谱成像技术的长枣含水量无损检测
[J].
Non-destructive determination of moisture in jujubes based on near-infrared hyperspectral imaging technique
[J].
基于高光谱成像技术小麦籽粒品种鉴别研究
[D].
Studies on varieties identification of wheat grain based on hyperspectral imaging technique
[D].
基于可见/近红外光谱的水稻品种快速鉴别研究
[J].
Discrimination of varieties of paddy based on VIS/NIR spectroscopy combined with chemometrics
[J].A simple, fast and non-destructive method based on visible/near infrared reflectance (Vis/NIR) spectroscopy and chemometrics was put forward for discriminating varieties of paddy. Firstly, A field spectroradiometer was used for collecting spectra in the wavelength range from 325 to 1 025 nm. The Vis/NIR spectra were acquired from 150 samples of five varieties of paddy. Secondly, original spectral data were decomposed as low-frequency wavelet coefficients and high-frequency wavelet coefficients by wavelet transform (WT) at first level. High-frequency wavelet coefficients were deleted as they contained too many noise, so the reconstructed signals from low-frequency wavelet coefficients were used as replacer of original spectral data. Thirdly, principal component analysis (PCA) compressed the above data into several new variables that were the linear combination of original spectral variables. The analysis suggested that the first four PCs (principle components) could account for 99.89% of the original spectral information, it means that the four PCs could explain most variation of original variables. In order to set up the model for discriminating varieties of paddy, the four diagnostic PCs were applied as inputs of back propagation artificial neural network (BP-ANN), and the values of varieties of different paddy were applied as the outputs of BP-ANN. The threshold of error was set as 0.2, the optimal structure of BP-ANN was three layers with nodes as 4-9-3. The whole 150 samples were randomly divided into two parts, one of which that consisted of 100 samples was used to model, and the other one containing 50 samples was used to predict. This model has been used to predict the varieties of 50 unknown samples, and the discrimination rate 96% has been obtained. It proved that the model was very reliable and practicable. In short, it is feasible to discriminate varieties of paddy based on visible/near infrared reflectance (Vis/NIR) spectroscopy and chemometrics.
基于激光诱导击穿光谱的水稻品种鉴别研究
[J].
Identification of rice seed varieties based on LIBS
[J].
Rice seed cultivar identification using near-infrared hyperspectral imaging and multivariate data analysis
[J].
DOI:10.3390/s130708916
URL
PMID:23857260
[本文引用: 1]
A near-infrared (NIR) hyperspectral imaging system was developed in this study. NIR hyperspectral imaging combined with multivariate data analysis was applied to identify rice seed cultivars. Spectral data was exacted from hyperspectral images. Along with Partial Least Squares Discriminant Analysis (PLS-DA), Soft Independent Modeling of Class Analogy (SIMCA), K-Nearest Neighbor Algorithm (KNN) and Support Vector Machine (SVM), a novel machine learning algorithm called Random Forest (RF) was applied in this study. Spectra from 1,039 nm to 1,612 nm were used as full spectra to build classification models. PLS-DA and KNN models obtained over 80% classification accuracy, and SIMCA, SVM and RF models obtained 100% classification accuracy in both the calibration and prediction set. Twelve optimal wavelengths were selected by weighted regression coefficients of the PLS-DA model. Based on optimal wavelengths, PLS-DA, KNN, SVM and RF models were built. All optimal wavelengths-based models (except PLS-DA) produced classification rates over 80%. The performances of full spectra-based models were better than optimal wavelengths-based models. The overall results indicated that hyperspectral imaging could be used for rice seed cultivar identification, and RF is an effective classification technique.
利用近红外高光谱图像技术快速鉴别西瓜种子品种
[J].为了研究采用近红外高光谱图像技术对西瓜种子品种快速无损鉴别的可行性，该文采用近红外高光谱图像技术，通过提取西瓜种子的光谱反射率，结合Savitzky-Golay (SG)平滑算法，经验模态分解算法（empirical mode decomposition，EMD）和小波分析（wavelet transform，WT）对提取出的光谱数据进行去除噪声处理，采用连续投影算法（successive projections algorithm，SPA）和遗传-偏最小二乘法（genetic algorithm-partial least squares，GA-PLS）进行特征波长选择。基于全波段光谱建立了偏最小二乘判别分析（partial least squares-discriminant analysis，PLS-DA），基于特征波长建立了反向传播神经网络（back-propagation neural network，BP NN) 判别模型和极限学习机（extreme learning machine，ELM）判别模型。试验结果表明，基于特征波长的BPNN模型和ELM模型的结果优于基于全部波长的PLS-DA模型，基于SG预处理光谱提取的特征波长建立的ELM模型取得最优的判别效果，建模集和预测集的判别正确率均为100%。结果表明应用近红外高光谱成像技术对西瓜种子品种鉴别是可行的，为西瓜种子的品种快速鉴别提供了一种新方法。
Fast identification of watermelon seed variety using near infrared hyperspectral imaging technology
[J].Watermelon seed variety selection plays a vital role in watermelon planting, and the variety of watermelon seeds directly affect the yield and quality of watermelons. In this study, we aimed to identify the cultivars of watermelon seeds by using a novel, rapid, non-invasive, and low cost technique named hyperspectral imaging. 121 samples of four different cultivars of watermelon seeds were investigated, and a near-infrared hyperspectral imaging system (874-1734 nm with 256 bands) was established to acquire the hyperspectral images of the samples. A region of interest (ROI) with 15×15 pixels of the hyperspectral image of each sample was defined, and the average reflectance spectrum of the ROI were extracted. To remove the absolute noises of the spectra, only the spectral range 1 042-1 646 nm was used for analysis, and to reduce the noises existed in spectral range 1 042-1 646 nm, the extracted 121 reflectance spectra were preprocessed by Savitzky-Golay smoothing (SG), Empirical Mode Decomposition (EMD), and Wavelet Transform (WT) methods. The preprocessed spectra were then used to select sensitive wavelengths by Successive Projections Algorithm (SPA) and Genetic Algorithm-partial least squares (GA-PLS) methods. Different numbers of sensitive wavelengths were selected by different variable selection methods with different preprocessing methods. 24, 16, and 15 sensitive wavelengths were selected by SPA with spectra preprocessed by SG, EMD, and WT, respectively. Moreover, 38, 33. and 32 sensitive wavelengths were selected by GA-PLS with spectra preprocessed by SG, EMD. and WT, respectively. Partial least squares - discriminant analysis (PLS-DA) was used to build discriminant models with the full spectra, and back-propagation neural network (BPNN) and extreme learning machine (ELM) were applied to build discriminant models with the selected wavelength variables. A PLS-DA model with spectra preprocessed by EMD obtained the best identification rate among all PLS-DA models, with an identification rate of 91.57% in the calibration set and 78.95% in the prediction set. SPA-BPNN models showed relatively worse results than GA-PLS-BPNN models with the same spectral preprocessing methods. The SG-GA-PLS-BPNN model obtained the best performance among all BPNN models, with an identification rate of 92.77% in the calibration set and 86.84% in the prediction set. Compared with the PLS-DA models and the BPNN models, ELM models obtained the best results. All ELM models obtained an identification rate over 90% in the calibration set and the prediction set, and the SG-SPA-ELM model, SG-GA-PLS-ELM model, and WT-SPA-ELM model obtained the identification rate of 100% of calibration and prediction. The overall results showed that BPNN and ELM models performed better than PLS-DA models, and the ELM models with the selected wavelengths based on SG preprocessed spectra obtained the best results, with 100% classification accuracy for both the calibration set and the prediction set. The SG preprocessing method showed the best performance in all PLS-DA, BPNN, and ELM models. The results indicated that it was feasible to use near-infrared hyperspectral imaging to identify the watermelon seed varieties, and near-infrared hyperspectral imaging provided an alternate way of rapid identification of watermelon seed variety. ELM, as a single hidden layer feed-forward network, was an effective classification method in watermelon seed cultivar identification. Moreover, the results in this paper showed the great potential of hyperspectral imaging in the seed industry for on-line identification of seed cultivars and detection of the seed quality parameters.
基于高光谱图像技术的油菜籽品种鉴别方法研究
[J].提出了一种采用高光谱图像技术结合人工神经网络对油菜籽品种进行鉴别的方法 . 采集多个品种油菜籽400~1000nm 范围的高光谱图像数据 , 通过主成分分析法 (PCA) 获得主成分图像 , 确定特征波长 ; 采用基于灰度直方图和灰度共生矩阵联合的统计方法从特征图像中提取纹理特征参数 , 应用人工神经网络建立油菜籽品种鉴别模型 . 结果表明 , 模型训练时品种判别率为93.75%, 预测的判别率为91.67%. 说明高光谱图像技术对油菜籽品种具有较好的分类和鉴别作用 .
Identification of rapeseed varieties based on hyperspectral imagery
[J].
Identification of soybean seed varieties based on hyperspectral image
[J].Different soybean seed varieties have different components (oil, protein, fat etc.) content. Identification of soybean seed varieties is a critical factor that improves the quality of produced soybean. In this study, hyperspectral image technique was applied in order to classify soybean seeds based on their varieties. The spectral reflectance data was collected using the optical sensor system with spectral region of 1000-2500nm. Principal component analysis (PCA) was performed to reduce the dimensionality of the data and remove the redundancy. Scores of four PCs were used as input features in the classification algorithm. Four texture feature parameters (angular second moment, energy, entropy and correlation) were extracted from each feature image selected by PCA. For the extraction of specific features, four significant feature parameters were computed from the 16 characteristic variables. Artificial neural network (ANN) classifier was employed for classification using top selected features. The obtained average training accuracy rate was 97.50% and the average testing accuracy rate was 93.88%. Thus, the results confirmed that hyperspectral image technique in-conjunction with BP neural network could be useful for soybean seed varieties classification.
An intelligent classification model for rubber seed clones based on shape features through imaging techniques
[C]//
高光谱与机器学习相结合的大白菜种子品种鉴别研究
[J].DOI:10.3964/j.issn.1000-0593(2014)09-2519-04 URL PMID:25532356 [本文引用: 1]
Variety recognition of chinese cabbage seeds by hyperspectral imaging combined with machine learning
[J].
Non-destructive quality evaluation of pepper (Capsicum annuum L.) seeds using LED-induced hyperspectral reflectance imaging
[J].
DOI:10.3390/s140407489
URL
PMID:24763251
[本文引用: 2]
In this study, we developed a viability evaluation method for pepper (Capsicum annuum L.) seeds based on hyperspectral reflectance imaging. The reflectance spectra of pepper seeds in the 400-700 nm range are collected from hyperspectral reflectance images obtained using blue, green, and red LED illumination. A partial least squares-discriminant analysis (PLS-DA) model is developed to classify viable and non-viable seeds. Four spectral ranges generated with four types of LEDs (blue, green, red, and RGB), which were pretreated using various methods, are investigated to develop the classification models. The optimal PLS-DA model based on the standard normal variate for RGB LED illumination (400-700 nm) yields discrimination accuracies of 96.7% and 99.4% for viable seeds and nonviable seeds, respectively. The use of images based on the PLS-DA model with the first-order derivative of a 31.5-nm gap for red LED illumination (600-700 nm) yields 100% discrimination accuracy for both viable and nonviable seeds. The results indicate that a hyperspectral imaging technique based on LED light can be potentially applied to high-quality pepper seed sorting.
Multispectral fluorescence imaging technique for discrimination of cucumber seed viability
[J].
高光谱技术无损检测单粒小麦种子生活力的特征波段筛选方法研究
[J].
Wavelength variable selection methods for non-destructive detection of the viability of single wheat kernel based on hyperspectral imaging
[J].
Discrimination of viable and empty seeds of Pinus patula Schiede & Deppe with near-infrared spectroscopy
[J].
基于高光谱图像的水稻种子活力检测技术研究
[J].
Study on detection technology of rice seed vigor based on hyperspectral image
[J].
基于高光谱图像的脱绒棉种活力检测方法研究
[D].
The detection method research on delinted cottonseeds’ vigor based on hyperspectral imaging
[D].
基于高光谱图像技术的脱绒棉种品种鉴别方法研究
[D].
Study on identification method of delinted cottonseeds varieties based on hyperspectral image technology
[D].
Prediction of viability of oriental beechnuts,fagus orientalis,using near infrared spectroscopy and partial least squares regression
[J].
High speed measurement of corn seed viability using hyperspectral imaging
[J].
A reliable methodology for determining seed viability by using hyperspectral data from two sides of wheat seeds
[J].
Characterisation of non-viable whole barley,wheat and sorghum grains using near-infrared hyperspectral data and chemometrics
[J].
DOI:10.1007/s00216-011-5291-x
URL
PMID:21842198
[本文引用: 2]
Undesired germination of cereal grains diminishes process utility and economic return. Pre-germination, the term used to describe untimely germination, leads to reduced viability of a grain sample. Accurate and rapid identification of non-viable grain is necessary to reduce losses associated with pre-germination. Viability of barley, wheat and sorghum grains was investigated with near-infrared hyperspectral imaging. Principal component analyses applied to cleaned hyperspectral images were able to differentiate between viable and non-viable classes in principal component (PC) five for barley and sorghum and in PC6 for wheat. An OH stretching and deformation combination mode (1,920-1,940 nm) featured in the loading line plots of these PCs; this water-based vibrational mode was a major contributor to the viable/non-viable differentiation. Viable and non-viable classes for partial least squares-discriminant analysis (PLS-DA) were assigned from PC scores that correlated with incubation time. The PLS-DA predictions of the viable proportion correlated well with the viable proportion observed using the tetrazolium test. Partial least squares regression analysis could not be used as a source of contrast in the hyperspectral images due to sampling issues.
Near-infrared hyperspectral imaging system coupled with multivariate methods to predict viability and vigor in muskmelon seeds
[J].
玉米种子活力近红外光谱智能检测方法研究
[J].
Research on maize vigor intelligent detection based on near infrared spectroscopy
[J].
Rapid assessment of corn seed viability using short wave infrared line-scan hyperspectral imaging and chemometrics
[J].
Comparative nondestructive measurement of corn seed viability using Fourier transform near-infrared (FT-NIR) and Raman spectroscopy
[J].
基于高光谱的水稻种子活力无损分级检测
[J].
Research on detection technology of rice seed vigor based on hyperspectral
[J].
高光谱技术在常规水稻种子活力检测中的应用
[J].
Inspect rice seed vigor of conventional rice by hyperspectral imaging with chemometric methods
[J].
Advanced applications of hyperspectral imaging technology for food quality and safety analysis and assessment:A review — Part II:Applications
[J].
Hyperspectral imaging applications in agriculture and agro-food product quality and safety control:A review
[J].
Predicting forage quality by infrared replectance spectroscopy
[J].DOI:10.2527/jas1976.434889x URL [本文引用: 1]
Hyperspectral waveband selection for on-line measurement of grain cleanness
[J].DOI:10.1016/j.biosystemseng.2009.05.011 URL [本文引用: 1]
Feasibility of near infrared spectroscopy for analyzing corn kernel damage and viability of soybean and corn kernels
[J].
DOI:10.1016/j.jcs.2011.11.002
URL
[本文引用: 1]
The current US corn grading system accounts for the portion of damaged kernels, measured by time-consuming and inaccurate visual inspection. Near infrared spectroscopy (NIRS), a non-destructive and fast analytical method, was tested as a tool for discriminating corn kernels with heat and frost damage. Four classification algorithms were utilized: Partial least squares discriminant analysis (PLS-DA), soft independent modeling of class analogy (SIMCA), k-nearest neighbors (K-NN), and least-squares support vector machines (LS-SVM). The feasibility of NIRS for discriminating normal or viable-germinating corn kernels and soybean seeds from abnormal or dead seeds was also tested. This application could be highly valuable for seed breeders and germplasm-preservation managers because current viability tests are based on a destructive method where the seed is germinated. Heat-damaged corn kernels were best discriminated by PLS-DA, with 99% accuracy. The discrimination of frost-damaged corn kernels was not possible. Discrimination of non-viable seeds from viable also was not possible. Since previous results in the literature contradict the current damage-discrimination results, the threshold of seed damage necessary for NIRS detection should be analyzed in the future. NIRS may accurately classify seeds based on changes due to damage, without any correlation with germination. (C) 2011 Elsevier Ltd.
Color and texture for corn seed classification by machine vision
[C]//
基于高光谱成像技术与CARS算法的玉米种子含水率检测
[J].
Moisture content detection of maize kernels based on hyperspectral imaging technology and CARS
[J].
Detection of sprouted and midge-damaged wheat kernels using near-infrared hyperspectral imaging
[J].
Fungal damage detection in wheat using short-wave near-infrared hyperspectral and digital colour imaging
[J].
Indirect detection of Fusarium verticillioides in maize (Zea maize L.) kernels by NIR hyperspectral imaging
[J].
Detection of sprout damage in Canada Western Red Spring wheat with multiple wavebands using visible/near-infrared hyperspectral imaging
[J].DOI:10.1016/j.biosystemseng.2010.03.010 URL [本文引用: 1]
Early detection of toxigenic fungi on maize by hyperspectral imaging analysis
[J].
DOI:10.1016/j.ijfoodmicro.2010.08.001
URL
PMID:20869132
[本文引用: 1]
Fungi can grow on many food commodities. Some fungal species, such as Aspergillus flavus, Aspergillus parasiticus, Aspergillus niger and Fusarium spp., can produce, under suitable conditions, mycotoxins, secondary metabolites which are toxic for humans and animals. Toxigenic fungi are a real issue, especially for the cereal industry. The aim of this work is to carry out a non destructive, hyperspectral imaging-based method to detect toxigenic fungi on maize kernels, and to discriminate between healthy and diseased kernels. A desktop spectral scanner equipped with an imaging based spectrometer ImSpector- Specim V10, working in the visible-near infrared spectral range (400-1000 nm) was used. The results show that the hyperspectral imaging is able to rapidly discriminate commercial maize kernels infected with toxigenic fungi from uninfected controls when traditional methods are not yet effective: i.e. from 48 h after inoculation with A. niger or A. flavus.
Correlation and classification of single kernel fluorescence hyperspectral data with aflatoxin concentration in corn kernels inoculated with Aspergillus flavus spores
[J].
DOI:10.1080/19440040903527368
URL
PMID:20221935
[本文引用: 1]
or=100 ng g(-1) (parts per billion), were significantly different from each other at the 0.01 level of alpha. Classification accuracy under a two-class schema ranged from 0.84 to 0.91 when a threshold of either 20 or 100 ng g(-1) was used. Overall, the results indicate that fluorescence hyperspectral imaging may be applicable in estimating aflatoxin content in individual corn kernels.]]>
基于高光谱成像技术和因子判别分析的玉米黄曲霉毒素检测研究
[J].
Detection of corn aflatoxin based on hyperspectral imaging technology and factor discriminant analysis
[J].
Detection of Fusarium damaged kernels in Canada Western Red Spring wheat using visible/near-infrared hyperspectral imaging and principal component analysis
[J].
Detecting Fusarium head blight in wheat kernels using hyperspectral imaging
[J].
Nondestructive assessment of protein content in single seeds of rapeseed (Brassica napus L.) by near-infrared reflectance spectroscopy
[J].
Single-kernel maize analysis by near-infrared hyperspectral imaging
[J].
Prediction of oil and oleic acid concentrations in individual corn (Zea mays L.) kernels using near-infrared reflectance hyperspectral imaging and multivariate analysis
[J].
DOI:10.1366/000370206775382631
URL
PMID:16454902
[本文引用: 1]
Due to their heterogeneous structure and variability in form, individual corn (Zea mays L.) kernels present an optical challenge for nondestructive spectroscopic determination of their chemical composition. Increasing demand in agricultural science for knowledge of specific traits in kernels is driving the need to find high-throughput methods of examination. In this study macroscopic near-infrared (NIR) reflectance hyperspectral imaging was used to measure small sets of kernels in the spectroscopic range of 950 nm to 1700 nm. Image analysis and principal component analysis (PCA) were used to determine kernel germ from endosperm regions as well as to define individual kernels as objects out of sets of kernels. Partial least squares (PLS) analysis was used to predict oil or oleic acid concentrations derived from germ or full kernel spectra. The relative precision of the minimum cross-validated root mean square error (RMSECV) and root mean square error of prediction (RMSEP) for oil and oleic acid concentration were compared for two sets of two hundred kernels. An optimal statistical prediction method was determined using a limited set of wavelengths selected by a genetic algorithm. Given these parameters, oil content was predicted with an RMSEP of 0.7% and oleic acid content with an RMSEP of 14% for a given corn kernel.
The use of near-infrared spectroscopy (NIRS) in the study of seed quality components in plant breeding programs
[J].
Prediction of maize seed attributes using a rapid single kernel near infrared instrument
[J].
基于仿生模式识别和近红外光谱的转基因小麦快速鉴别方法
[J].
Fast discrimination of varieties of transgene wheat based on biomimetic pattern recognition and near infrared spectra
[J].A new method for the fast discrimination of varieties of transgene wheat by means of near infrared spectroscopy and biomimetic pattern recognition (BPR) was proposed and the recognition models of seven varieties of transgene wheat and two varieties of acceptor wheat were built. The experiment adopted 225 samples, which were acquired from nine varieties of wheat. Firstly, a field spectroradiometer was used for collecting spectra in the wave number range from 4 000 to 12 000 cm(-1). Secondly, the original spectral data were pretreated in order to eliminate noise and improve the efficiency of models. Thirdly, principal component analysis (PCA) was used to compress spectral data into several variables, and the cumulate reliabilities of the first ten components were more than 97.28%. Finally, the recognition models were established based on BPR For the every 25 samples in each variety, 15 samples were randomly selected as the training set. The remaining 10 samples of the same variety were used as the first testing set, and all the 200 samples of the other varieties were used as the second testing set. As the 96.7% samples in the second set were correctly rejected, the average correct recognition rate of first testing set was 94.3%. The experimental results demonstrated that the recognition models were effective and efficient. In short, it is feasible to discriminate varieties of transgene wheat based on near infrared spectroscopy and BPR.
Discrimination of transgenic maize kernel using NIR hyperspectral imaging and multivariate data analysis
[J].
