基于导函数层序聚类的光谱成分预测方法研究(英文)

多光谱成像技术通过增加颜色通道的维数,成功的实现了基于光谱的颜色复制。然而,由于其颜色信息维数较高,此方法在提高色度精度的同时引入了较大的计算及存储压力。为此,最常用的方法就是通过对光谱数据进行分组,并利用每组光谱数据集中的主成分向量来对各个光谱曲线进行线性表示,从而实现数据的降维处理。提出了1种新的针对光谱数据导函数曲线的聚类分析方法,并利用伪逆算法进行光谱重建;本研究采用孟塞尔光泽色卡及无光泽色卡作为实验数据集,并将提出的导函数聚类分析法与现有的主成分分析法、聚类分析法以及色相角分类法相比较,实验结果证明其颜色预测精度在色度匹配及光谱匹配方面均优于现有方法。     

基于BP神经网络的多光谱空间表示方法研究

多光谱成像技术通过增加颜色通道的维数,克服了传统颜色复制方式在同色异谱方面的缺陷,成功的实现了基于光谱的颜色复制。然而,由于其颜色信息维数较高,此方法在提高色度精度的同时引入了较大的计算及存储压力。为此,最常用的方法就是通过特定的光谱分组度量对光谱数据进行分组,并利用每组光谱数据集中的主成分向量来对各个光谱曲线进行线性表示,从而实现数据的降维处理。本研究提出了一种全新的,基于主波长分组及BP神经网络寻址的光谱空间表示方法,并通过对具体光谱颜色数据集的向量表示,证明此方法在光谱颜色表达的方面的准确性。     

基于BP神经网络的多光谱空间表示方法研究(英文)

多光谱成像技术通过增加颜色通道的维数,克服了传统颜色复制方式在同色异谱方面的缺陷,成功的实现了基于光谱的颜色复制。然而,由于其颜色信息维数较高,此方法在提高色度精度的同时引入了较大的计算及存储压力。为此,最常用的方法就是通过特定的光谱分组度量对光谱数据进行分组,并利用每组光谱数据集中的主成分向量来对各个光谱曲线进行线性表示,从而实现数据的降维处理。本研究提出了一种全新的,基于主波长分组及BP神经网络寻址的光谱空间表示方法,并通过对具体光谱颜色数据集的向量表示,证明此方法在光谱颜色表达的方面的准确性。     

基于RBFNN-FDA模型对塑料拖鞋鞋底的拉曼光谱研究

为建立一种快速无损检验区分塑料拖鞋鞋底的方法,利用显微共聚焦激光拉曼光谱仪采集了43个不同来源的塑料拖鞋鞋底样本的拉曼光谱图。拉曼数据经主成分分析降维后提取特征矩阵,对得到的特征矩阵进行系统聚类,建立Fisher判别函数对系统聚类的结果进行评价。最终构建径向基函数神经网络(RBFNN)实现对样本的鉴别分类,并绘制接受者操作特征曲线用以评估诊断价值。结果表明：拉曼数据提取出的特征矩阵经系统聚类被分为4组,Fisher判别分析经交叉验证后准确率为97.7%,径向基函数神经网络的准确率为100%。该方法实现了对样本快速无损的分类及预测,模型结构准确,可以为公安实际办案提供一种新思路。     

红外光谱法结合K⁃均值聚类与神经网络对饮料瓶的检验研究

利用傅里叶变换红外光谱法对41个不同品牌的塑料饮料瓶进行快速无损检测.谱图数据在经过预处理后可将样品分为聚对苯二甲酸乙二醇酯和聚乙烯两类.每一类内部的各个样品红外特征峰存在差异.对于数量最多的一类样品,通过主成分分析将样品光谱数据降维并提取主成分,然后结合K-均值聚类对样品进一步分组.最后以聚类结果作为因变量,构建神经...     

有序的KD-tree在图像特征匹配上的应用

针对用KD-tree实现高维空间点匹配中存在的错误匹配问题进行讨论,分析其存在的原因;接着,使用PCA,根据各维数之间的协方差,求出它们的主成分奉献率,再按主成分奉献率进行维数优先级排序,并在该基础上增加了KD-tree各节点的权重;最后,将改进前后的KD-tree应用于Sift特征点匹配。实验证明,改进后的KD-tree能在保持实时性的前提下,大大提高匹配的准确率。     

差分拉曼光谱结合机器学习对汽车灯罩的可视化研究

为实现对汽车灯罩的快速无损、高效便捷的分类及预测,提出一种基于差分拉曼光谱结合机器学习对灯罩的可视化鉴别方法。利用差分拉曼光谱仪对32个品牌、9种车型共计46种汽车灯罩样品进行了检测,得到了样品的差分拉曼光谱图。对光谱数据先进行主成分分析降维,用提取到的5个PCA对样品进行自组织映射(SOM)聚类,同时对聚类结果可视化处理,再结合多元无序logistic回归、可优化支持向量机(SVM)对样品进行分类预测。46个样品被分为7类,多元无序logistic回归、可优化支持向量机的准确率皆为100%,实现了对样品的分类及预测。差分拉曼光谱准确高效,谱峰尖锐清晰,可以分析样品的主要成分及填料。SOM函数分类效果较好,所建立的分类模型可以精准地对不同来源汽车灯罩样品进行区分及归属预测,为汽车灯罩物证溯源提供了一种新的思路和技术手段。     

拉曼光谱结合化学计量学在不同品位磷矿快速分类中的应用

采用拉曼光谱技术结合化学计量学探讨不同品位磷矿快速鉴别和分类的可行性。采用共聚焦显微拉曼光谱系统分析了高、中、低三类不同品位的4种磷矿样品在200～1 950 cm~(-1)范围内的拉曼光谱特性,并对经过自适应迭代重加权惩罚最小二乘(airPLS)算法校正、一阶导和二阶导3种光谱预处理方法处理后的拉曼光谱结合主成分分析(PCA)和系统聚类分析(HCA)建立判别模型。结果显示,在主成分分析(PCA)中,经过3种预处理方法后的拉曼光谱均能实现对4种磷矿样本的聚类,且前两种预处理方式中,在第1主成分上,4种样品随品位值呈规律分布。使用PCA降维后的一阶导数光谱结合系统类分析(HCA)对4种磷矿样品进行分类,准确率为98.75%。结果表明,利用拉曼光谱技术结合化学计量学能够实现不同品位磷矿的快速鉴别和分类,为磷矿品位现场快速检测和评估打下基础。     

拉曼光谱结合化学计量学在不同品位磷矿快速分类中的应用

采用拉曼光谱技术结合化学计量学探讨不同品位磷矿快速鉴别和分类的可行性。采用共聚焦显微拉曼光谱系统分析了高、中、低三类不同品位的4种磷矿样品在200～1 950 cm^(-1)范围内的拉曼光谱特性,并对经过自适应迭代重加权惩罚最小二乘(airPLS)算法校正、一阶导和二阶导3种光谱预处理方法处理后的拉曼光谱结合主成分分析(PCA)和系统聚类分析(HCA)建立判别模型。结果显示,在主成分分析(PCA)中,经过3种预处理方法后的拉曼光谱均能实现对4种磷矿样本的聚类,且前两种预处理方式中,在第1主成分上,4种样品随品位值呈规律分布。使用PCA降维后的一阶导数光谱结合系统类分析(HCA)对4种磷矿样品进行分类,准确率为98.75%。结果表明,利用拉曼光谱技术结合化学计量学能够实现不同品位磷矿的快速鉴别和分类,为磷矿品位现场快速检测和评估打下基础。     

光谱、光度、色度计量用玻璃的研究

研制成光谱、光度与色度计量用的单一玻璃滤光片,模拟颜色匹配函数(?)(λ)、(?)(λ)、(?)(λ)。 采用铝磷酸盐、铝硼酸盐、硼硅酸盐玻璃作为基质玻璃。研究了着色剂的合适浓度及其x效应;系统地考察了氧化还原条件、碱度、玻璃成分以及基质玻璃组合对于玻璃光谱特性的影响。     

Measurement and Analysis of Object Reflectance Spectra

Recent algorithms developed in the field of color vision make assumptions based on the spectral reflectance curves of Munsell chips and natural materials. Some of them rely on data collected many years ago. which is partially incomplete in the visible spectrum. or contains many occurrences of the same material in it. In this article. we present a set of new measurements of different materials. In particular. we measured the spectral reflectance of Munsell chips, paints. and various natural materials in the 390–730-nm range. In addition, we have analyzed. through principal-component analysis, the possibility of representing the data collected with a set of basis functions. We show the implications of varying the number of principal components used (from 7 down to 3) on the errors introduced using this method.     

Reconstruction of reflectance spectra using weighted principal component analysis

The weighted principal component analysis technique is employed for reconstruction of reflectance spectra of surface colors from the related tristimulus values. A dynamic eigenvector subspace based on applying certain weights to reflectance data of Munsell color chips has been formed for each particular sample and the color difference value between the target, and Munsell dataset is chosen as a criterion for determination of weighting factors. Implementation of this method enables one to increase the influence of samples which are closer to target on extracted principal eigenvectors and subsequently diminish the effect of those samples which benefit from higher amount of color difference. The performance of the suggested method is evaluated in spectral reflectance reconstruction of three different collections of colored samples by the use of the first three Munsell bases. The resulting spectra show considerable improvements in terms of root mean square error between the actual and reconstructed reflectance curves as well as CIELAB color difference under illuminant A in comparison to those obtained from the standard PCA method. © 2008 Wiley Periodicals, Inc. Col Res Appl, 33, 360–371, 2008     

A residual modified transformation formula from munsell to sRGB color system

A residual modified transformation formula from Munsell to sRGB color system is presented in this article. The development of the transformation formula is based on the 1625 Munsell color chips in the Munsell Renotation Data that could be displayed on the sRGB monitor. The developed transformation formula consists of two models, one is named as the corresponding matrix model and the other is the residual modified model. The corresponding matrix model was obtained using numerical analysis methods to map each chip color attribute values from Munsell to sRGB and then its corresponding matrix for each Hue was constructed. The residual modified model was obtained using the discrete cosine transform to construct a residual modified function, which was used to modify the transformation error of each chip after applying the corresponding matrix model. The transformed accuracy rate for the corresponding matrix model is 88.4% and for the residual modified model can be enhanced to 96.6% for all of the chips. The developed transformation formula can be applied to research in which Munsell colors are presented on the sRGB monitor. With the aid of these formulas, designers can show the advanced real‐time results on a sRGB monitor for the product's color planning based on Munsell color system. Therefore, this research has a great contribution on the practical application for color planning in product design. © 2014 Wiley Periodicals, Inc. Col Res Appl, 40, 243–255, 2015     

Munsell reflectance spectra represented in three‐dimensional Euclidean space

In this article we describe the results of an investigation into the extent to which the reflectance spectra of 1269 matt Munsell color chips are well represented in low dimensional Euclidean space. We find that a three dimensional Euclidean representation accounts for most of the variation in the Euclidean distances among the 1269 Munsell color spectra. We interpret the three dimensional Euclidean representation of the spectral data in terms of the Munsell color space. In addition, we analyzed a data set with a large number of natural objects and found that the spectral profiles required four basis factors for adequate representation in Euclidean space. We conclude that four basis factors are required in general but that in special cases, like the Munsell system, three basis factors are adequate for precise characterization. © 2003 Wiley Periodicals, Inc. Col Res Appl, 28, 182–196, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.10144     

Colour category foci of Munsell colour spectra revealed by two computational methods

Colour names and psychophysical colour categories play an important role on human communication. For several application areas from computer vision to Internet shopping, it would be useful to manage colour information using methods of computational colour naming in a similar manner as people do in their everyday life. In this study, we applied two computational methods, the nonnegative matrix factorization and self‐organizing maps, to derive basic colours from a spectral database of Munsell colours, and a subset of it. The subset was generated to include only the most saturated samples of each Munsell Hue and Munsell Value pair of the original database. Using both the methods and both the databases, we calculated the sets of 3, 4, 6, and 8 basis vectors to represent the focal colours of colour categories. Colour names of the calculated focal colours were investigated using the results by Sturges and Whitfield as a reference. Nonnegative matrix factorization yields calculated colours more compatible with human basic colours, but the spectra generated by self‐organizing maps are more similar to natural spectra as their shape is smoother. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2010     

Reconstruction of reflectance data by modification of Berns' Gaussian method

Berns' method for the synthesis of spectral reflectance curve from the tristimulus color coordinates is modified. Firstly, the Gaussian bell shape red primary is replaced with a sigmoidal one to solve the dissimilarity between the spectral curves at the end region of spectrum. Secondly, three predetermined Gaussian primaries used in the original Berns' method are replaced by the adaptive ones which their half‐height bandwidths vary with the tristimulus values of the desired color. The mentioned modifications are applied for the recovery of the reflectance curves of 1409 surface colors (including 1269 Munsell color chips and 140 samples of Colorchecker SG) and also 204 textile samples. Results of recovery are evaluated by the mean and the maximum color difference values under other standard light sources. The mean as well as the maximum of root mean squares between the reconstructed and the actual spectra are also calculated. The modifications are compared with the common principal component analysis (PCA) as well as Hawkyard's methods for recovery of reflectance factor. Although the PCA leads to the best results, the modifications significantly improve the recovery outcomes in comparison with the original Berns method. © 2008 Wiley Periodicals, Inc. Col Res Appl, 34, 26–32, 2009.     

Estimation of CIE tristimulus values under various illuminants

In this work, a procedure is described to estimate the CIE tristimulus values of Munsell color chips under new illuminants from CIE tristimulus values under specific illuminant. The estimation was performed by multiple linear regression (MLR) and principal component analysis (PCA) techniques. In addition, the effects of illuminants combination and Munsell hue, value and chroma on reconstruction of CIE tristimulus values were evaluated in both methods. The results obtained indicate that the performance of principal component method is better than multiple regression method. The accuracy of estimation depends on the initial and target illuminants and hue, chroma and value of samples. © 2009 Wiley Periodicals, Inc. Col Res Appl, 34, 100–107, 2009