Visual enhancement of old documents with hyperspectral imaging

Hyperspectral imaging (HSI) of historical documents is becoming more common at national libraries and archives. HSI is useful for many tasks related to document conservation and management as it provides detailed quantitative measurements of the spectral reflectance of the document that is not limited to the visible spectrum. In this paper, we focus on how to use the invisible spectra, most notably near-infrared (NIR) bands, to assist in visually enhancing old documents. Specifically, we demonstrate how to use the invisible bands to improve the visual quality of text-based documents corrupted with undesired artifacts such as ink-bleed, ink-corrosion, and foxing. For documents of line drawings that suffer from low contrast, we use details found in the invisible bands to enhance legibility. The key components of our framework involve detecting regions in the document that can be enhanced by the NIR spectra, compositing the enhanced gradient map using the NIR bands, and reconstructing the final image from the composited gradients. This work is part of a collaborative effort with the Nationaal Archief of the Netherlands (NAN) and Art Innovation, a manufacturer of hyperspectral imaging hardware designed specially for historical documents. Our approach is evaluated on historical documents from NAN that exhibit degradations common to documents found in most archives and libraries.     

SODA: A new method of in-scene atmospheric water vapor estimation and post-flight spectral recalibration for hyperspectral sensors: Application to the HyMap sensor at two locations

A new method of estimating per-pixel atmospheric column water vapor (ACWV) and potential differences in the reported band center wavelengths of the HyMap sensor has been developed. The new method uses variations of a second order derivative algorithm (SODA) to assess the impact of atmospheric residual features on calculated surface reflectance spectra after atmospheric compensation. The SODA method provides an alternative to the current band ratio techniques of ACWV estimation and also allows the same form of algorithm to be used for the estimation of possible band shifts. A comparison of in-situ measured surface reflectance at two field sites in Western Australia demonstrates improvement in the resulting spectra when post-flight updates are made to the reported HyMap band center wavelengths and applied during the atmospheric compensation process. The same SODA methodology was varied to estimate the ACWV on a per-pixel basis and found to significantly reduce the appearance of the underlying surface structure on the resulting ACWV images as well as improve the overall accuracy of the estimation. The ACWV estimated from the HyMap imagery at the two field sites was found to agree with in-situ atmospheric ACWV measurements to within 2% and represented a two fold increase in accuracy over a 3 band ratio Continuum Interpolated Band Ratio (CIBR) technique of ACWV estimation.     

Kernel-based regularized-angle spectral matching for target detection in hyperspectral imagery

Target detection is one of the most important applications of hyperspectral imagery in the field of both civilian and military. In this letter, we firstly propose a new spectral matching method for target detection in hyperspectral imagery, which utilizes a pre-whitening procedure and defines a regularized spectral angle between the spectra of the test sample and the targets. The regularized spectral angle, which possesses explicit geometric sense in multidimensional spectral vector space, indicates a measure to make the target detection more effective. Furthermore Kernel realization of the Angle-Regularized Spectral Matching (KAR-SM, based on kernel mapping) improves detection even more. To demonstrate the detection performance of the proposed method and its kernel version, experiments are conducted on real hyperspectral images. The experimental tests show that the proposed detector outperforms the conventional spectral matched filter and its kernel version.     

基于自适应脊波网络的高光谱遥感图像分类

神经网络是遥感地物自动分类的重要工具之一。利用多尺度几何分析中的眷波基函数建立了一种自适应眷波网络模型。在传统自适应粒子群算法的基础上,提出一种引入粒子密度因子的自适应粒子群优化算法作为网络训练算法。为验证其性能,利用互信息约简技术对22。波段AVIRIS 92AV3C高光谱数据进行约简,并将它们作为网络输入实现对高光谱遥感地物的自动分类。仿真试验表明:引入粒子密度因子的粒子群算法与传统粒子群算法相比,不易出现早熟问题,在处理高维非线性组合优化问题时具有一定优势;由于眷波函数对高维奇异性的表征能力,相比于传统的RBF和SVM分类器,脊波神经网络分类器对具有明显边界特征的地物分类问题具有较高的精度,同时网络规模小,结构简单。     

从意象思维到图式表达——论艺术设计意图的清晰化

设计活动是一个由模糊到清晰的过程,它是建立在人对事物的观察、感知与思考的基础上。在这个过程中思维意象要不断地呈现出来,呈现的方式有两种,一种是以图的形式,一种是以实物或模型的形式。图式表达相对来说是一种便捷、基础的方式。从这个角度来看,图像化的思维与表达对于研究形态的创造与变化来说是不可或缺的。本论文通过对意象思维的分析,针对形态创造的问题,提出一系列观察思考与表达的方法,以发现新的设计灵感源泉。     

Non-Linear Scale-Spaces Isomorphic to the Linear Case with Applications to Scalar,Vector and Multispectral Images

A basic requirement of scale-space representations in general is that of scale causality, which states that local extrema in the image should not be enhanced when resolution is diminished. We consider a special class of nonlinear scale-spaces consistent with this constraint, which can be linearised by a suitable isomorphism in the grey-scale domain so as to reproduce the familiar Gaussian scale-space. We consider instances in which nonlinear representations may be the preferred choice, as well as instances in which they enter by necessity. We also establish their relation to morphological scale-space representations based on a quadratic structuring function.     

新型高光谱图像的超分辨率制图方法

由于高光谱图像的应用在很大程度上受限于其较低的空间分辨率,为此提出了一种结合支持向量机和小波变换的高光谱图像超分辨率制图方法．先对高光谱图像进行光谱解混得到分量图,然后对分量图进行一级小波分解．各局域窗内中心像元的3个高频系数与邻域像元低频系数之间的对应关系表示为训练样本,用于支持向量机的学习．训练好的模型用来对低分辨率图像即分量图进行超分辨率制图．实验表明,这种借助小波变换来获取训练样本的学习方法无需先验信息,相比采用BP神经网络学习的方法,支持向量机的超分辨率制图效果更佳．     

一种改进的基于正交子空间投影的高光谱图像异常检测算法

针对背景与目标能量差异较小的高光谱图像异常检测技术需求,提出了一种基于正交子空间投影的异常检测方法：首先利用正交子空间投影抑制背景信息;然后调整噪声,根据空间密度选取决策半径实现异常检测,并对结果进行形态滤波,消除大面积虚警。实验结果表明,本文算法能够检测到能量与背景差异较小的异常,且计算效率较高。     

Quantitative comparison between full-spectrum and filter-based imaging in hyperspectral fluorescence microscopy

We implement a filterless illumination scheme on a hyperspectral fluorescence microscope to achieve full-range spectral imaging. The microscope employs polarisation filtering, spatial filtering and spectral unmixing filtering to replace the role of traditional filters. Quantitative comparisons between full-spectrum and filter-based microscopy are provided in the context of signal dynamic range and accuracy of measured fluorophores' emission spectra. To show potential applications, a five-colour cell immunofluorescence imaging experiment is theoretically simulated. Simulation results indicate that the use of proposed full-spectrum imaging technique may result in three times improvement in signal dynamic range compared to that can be achieved in the filter-based imaging.     

An approach for visual attention based on biquaternion and its application for ship detection in multispectral imagery

This paper proposes an approach for visual attention based on biquaternion, and investigates its application for ship detection in multispectral imagery. The proposed approach describes high-dimensional data in the form of biquaternion and utilizes the phase spectrum of biquaternion Fourier transform to generate a required saliency map that can be used for salient target detection. In our method, the multidimensional data is processed as a whole, and the features contained in each spectral band can be extracted effectively. Compared with traditional visual attention approaches, our method has very low computational complexity. Experimental results on simulated and real multispectral remote sensing data have shown that the proposed method has excellent performance in ship detection. Furthermore, our method is robust against white noise and almost meets real-time requirements, which has great potentials in engineering applications.