商空间理论下面向对象的遥感影像分类

传统的分类方法仅仅基于像素光谱特征,不适合于高分辨率遥感影像.本文提出了一种新的基于商空间理论,面向对象的高分辨率遥感影像分类方法,即综合云模型、模糊支持向量机和决策树的分层合成分类技术.针对决定分类效果的两个因素,影像分割和分类算法,分别做出了一些改进.第一,本文提出了一个自适应的基于云模型的区域增长分割策略.第二,...     

面向对象的高分辨遥感影像信息提取技术

面向对象的影像信息提取技术已经成为了解决高分辨率遥感影像计算机解译的主要手段。本文从面向对象影像信息提取技术中图像多尺度分割这一关键步骤出发,对其基本思想和主要算法FENA进行了叙述。最后本文以eCognition软件中的多尺度分割算法为例,进行了相关实验,证明了面向对象的遥感影像分析技术的可行性及在计算机解译中的优势性。     

一种高分辨率城区遥感影像分割方法

高分辨率遥感影像具有丰富的地物细节信息。本文将均值漂移算法应用到高分辨率城市区域遥感影像的分割中,并对其实验结果进行了分析,为下一步的研究提供理论支持。     

面向对象的高分辨率遥感影像分类在卫片执法中的应用研究

近年来,随着我国城市化进程不断向前迈进,城市用地日趋紧张起来,在这样的背景之下,出现了诸多的违法用地、建设等事端。近年来,为了有效地打击此类的违法行为,确保我国土地使用更加科学、合理,有关部门逐渐在对土地动态监测的过程中采用了高分辨率遥感影像分类技术。本文基于此,分析探讨了面向对象的高分辨率遥感影像分类在卫片执法中的应用。     

高分辨率数字成像技术及其在遥感中的应用

高分辨率技术对于数字图像特征非常重要。尤其近20年来,数字成像技术大大推动了空间信息领域的发展。随着遥感应用需求的增加,获取高分辨率图像的任务非常紧迫。为了达到这个目标,我们的工作从四方面展开,分别是:空间分辨率、时间分辨率、光谱分辨率和辐射分辨率。根据不同的分辨率提出不同的成像方式,并设计了相应的原型系统。在此基础上,进行了大量的实验来验证各个成像方式的可行性。     

基于ScSPM-Reranking的高分辨率遥感影像的检索

为了从高分辨率遥感影像中获取详细的地表地物信息,为城市规划、环境监测以及灾情分析提供可靠的数据,进行了高分辨率遥感影像的检索研究,包括对图像的特征提取和图像之间相似度的描述。为了提高图像检索精度,运用了采用稀疏编码(Sc)的空间塔式匹配(Sc SPM)技术和重排序(Reranking)技术,提出了基于Sc SPM结合Reranking(ScSPM-Reranking)的遥感高分辨率影像的检索方法。该方法首先使用Sc SPM提取空间场景的特征,然后结合这些特征使用cityblock距离进行初步检索,最后对初步检索的结果进行Reranking排序,获得高精度的检索结果。同其他检索方法进行了对比实验,实验结果证明,该方法具有较高的检索精度。     

基于子空间字典偶学习的高光谱图像分类

针对高光谱高分辨率带来巨大数据量和空间分辨率引起混合像元的问题,提出了基于子空间(subspace)的字典偶学习(DPL)算法,简称DPLsub算法。DPL算法是对字典学习的改进,它通过学习得到综合字典和分析字典,在模式识别中体现了高效性,而子空间投影的方法能更好地表征噪声和高度混合的像元。将光谱和空间特征融合的方法用于分类研究试验。实验数据是两幅高光谱影像,比较了子空间字典偶学习(DPLsub)模型和其他三种分类器即最小二乘支持向量机(LS-SVM)、稀疏多分类回归(SMLR)和字典学习(DL-OMP)的分类结果。实验结果显示,DPLsub算法无论在时间上还是精度上都优于其他算法,证明了这种子空间字典偶学习方法对高光谱图像分类的可行性与高效性。     

DCI为什么要求数字影院放映系统的分辨率为4K或2K

上世纪90年代初,美国电影电视工程师学会"数字影像层级结构"特别工作小组对成像系统的空间分辨率要求进行了研究.该小组从分辨率的角度将显示器分为低分辨率、常规分辨率、高分辨率和超高分辨率等四个层级.该小组基于以下几个基本概念,提出了空间分辨率层级的设计理念:--基于常用的数字处理和存储体系结构的一种层级整数递进;--支持基于影像贴片(image tile)的各种宽高比和空间分辨率;--能够利用适宜分辨率的贴片,针对任何应用要求来构建显示器.通过使用贴片和仅仅四个分辨率层级,便有可能构建出适用于任何用途的新显示器,而且这种显示器还可以用于显示来自任何其它层级的影像.美国数字电影倡导组织(DCI)要求数字影院放映系统的分辨率为4K或2K,属超高分辨率或高分辨率的级别.     

基于多残差网络的遥感图像语义分割方法

高分辨率遥感图像含有许多较为复杂的地物信息,对其进行的语义分割存在分割精度低、分割边界模糊等问题.本文提出一种新型的多尺度语义分割网络模型,旨在提高遥感图像语义分割精度.该模型为编码—解码(Encoder-Decoder)网络结构,编码器利用残差网络对图像特征进行提取;解码器利用反卷积进行上采样;残差连接将提取到的高级语义特征与残差连接层提取到的多尺度特征进行融合;同时使用Dice损失函数代替传统的交叉熵损失函数,以处理多类语义分割任务中的类别数量不平衡和难分样本问题.实验可得:与其它经典分割模型相比,本文算法对遥感图像具有较高的分割精度,所提出的方法在"CCF卫星影像的AI分类与识别竞赛"的数据集上均交并比(Mean Intersection over Union, MIoU)值达到了0.823 5,召回率Recall达到0.891 4.     

Large scale compound classification for multi-resolution satellite data based on conditional random fields models

针对如何在算法层次上利用不同空间分辨率遥感数据提高地表分类精度的问题,提出了一种基于条件随机场模型的全新的多分辨率复合分类算法.该算法针对同一地区、不同覆盖范围的两种高低分辨率遥感图像,以广域低分辨率图像的高精度地表分类为目的,利用高低分辨率图像间的空间分辨率多对一关系,基于云理论构建“真实”似然特征映射,由用来描述光谱特征与类别关系的“真实”似然特征序列以及像元间上下文关系构建条件随机场模型的两类势函数,并在此基础上对广域低分辨率图像进行全局地表分类.该算法不仅提供了对多分类特征的支持,而且考虑了地物分布的空间连续性.多组高低分辨率图像组合下的复合分类及不同算法间的分类精度对比分析结果表明,该算法可有效提高广域低分辨率图像的分类精度,并具有良好的鲁棒性.     

Texture augmented analysis of high resolution satellite imagery in detecting invasive plant species

Abstract

During recent decades, a considerable number of alien species have been brought into Taiwan and have caused significant impacts to local ecosystems and biodiversity. High resolution satellite imagery can provide detailed spatial characteristics over a large area and has a great potential for accurate vegetation mapping. However, most traditional multispectral image classification techniques focus on spectral discrimination of ground objects and may overlook useful spatial information provided by high resolution images. To achieve the best result, analysis of high resolution imagery should also incorporate spatial variations of the data. Therefore, this paper has looked into using a texture augmented procedure to analyze a high resolution satellite (QuickBird) image in order to detect an invasive plant species (Leucaena leucocephala) in southern Taiwan. Samples of primary vegetation covers were selected from the image to determine suitable texture analysis parameters for extracting texture features helpful for classification. Validation with ground truth data showed that the analysis produced high accuracies in detecting the target plant species and overall classification for primary vegetation types within the study site.     

A prototype expert system for interpretation of remote sensing image data

Automated image interpretation systems of remotely sensed images are of great help in the present scenario of growing applications. In this paper, we have critically studied visual interpretation processes for urban land cover and land use information. It is observed that the core activity of interpretation can be described as plausible combinations of pieces of evidential information from various sources such as images, collateral data, experiential knowledge and pragmatics. Interpretation keys for the interpretation of standard false colour composites are considered to be tone/colour, pattern, texture, size, shape, association, relief and season. These interpretation keys encompass the spectral, spatial and temporal knowledge required for image interpretation. Our focus is on a knowledge-based approach for interpretation of standard false colour composites (fcc). Basic information required for a knowledge-based approach is of four types viz., spectral, spatial, temporal and heuristic. Generic classes and subclasses of image objects are identified for the land use/land cover theme. Logical image objects are conceptualised as region/area, line and point objects. An object-oriented approach for the representation of spectral and spatial knowledge has been adopted. Heuristic information is stored in rules. The Dempster-Shafer theory of evidence is used to combine evidence from various interpretation keys for identification of generic class and subclass of a logical image object. Analysis of some Indian Remote Sensing Satellite images has been done using various basic probability assignments in combination with learning. Explanation facility is provided by tracing the rules fired in the sequence.     

Opportunities and pitfalls in characterisation of nanoscale features

Abstract

Thermally assisted field emission gun TEM (FEGTEM), with associated electron energy loss spectroscopy (EELS) and electron spectroscopic imaging (ESI) is used to characterise nanoscale features in technologically important materials. The exceptionally high spatial and temporal coherence of the FEG source allows high resolution image information down to 0.1 nm to be routinely imaged, but the complexity of the atomic contrast makes interpretation of interface structure more difficult than with LaB₆ sources. A spatial resolution of 0.9 nm for through thickness features and 2 nm for embedded particles is demonstrated for ESI, illustrating the remarkable sensitivity of this technique. The energy resolution of the FEG source (0.65 eV) is sufficient to resolve the fine near edge structure in EELS spectra to allow bonding information to be obtained with a spatial resolution of 1 nm or possibly better. X-ray energy dispersive spectroscopy can be performed on a similar scale, sufficiently small to identify submonolayer segregates at interfaces. The limitations of each technique are discussed in relation to the critical microstructural information required for each material investigated.     

Effect of image resolution on intensity based scene illumination classification using neural network

In this paper, a framework for testing scene illumination classification with different image resolutions is proposed. The testing aims to provide the researchers with valuable information about the effect of image resolution on scene illumination classification using a neural network. The experiment is done by extracting three types of features from the images. These three types consist of statistical features, physic based features and histogram based features. It has been demonstrated that scene illumination classification can be affected by changing the image resolution. Despite the popular belief that high resolution images lead to better results, scene illumination classification by the proposed method performed best using low resolution images. At the second part of discussion, the reason behind this phenomenon is mathematically analysed and explained.     

Digitized holography: modern holography for 3D imaging of virtual and real objects

Recent developments in computer algorithms, image sensors, and microfabrication technologies make it possible to digitize the whole process of classical holography. This technique, referred to as digitized holography, allows us to create fine spatial three-dimensional (3D) images composed of virtual and real objects. In the technique, the wave field of real objects is captured in a wide area and at very high resolution using the technique of synthetic aperture digital holography. The captured field is incorporated in virtual 3D scenes including two-dimensional digital images and 3D polygon mesh objects. The synthetic field is optically reconstructed using the technique of computer-generated holograms. The reconstructed 3D images present all depth cues like classical holograms but are digitally editable, archivable, and transmittable unlike classical holograms. The synthetic hologram printed by a laser lithography system has a wide viewing zone in full-parallax and give viewers a strong sensation of depth, which has never been achieved by conventional 3D systems. A real hologram as well as the details of the technique is presented to verify the proposed technique.     

Multiscale gradient domain compression for astronomical high dynamic range imaging

Visualisation of high dynamic range images requires compression of the data to be properly displayed on media with more limited dynamic ranges. Astronomical images pose a difficult challenge for dynamic range compression algorithms, due to the nature of the imaged objects and to the lack of a reflectance illumination model based on spatial frequencies. As a result, most of the algorithms commonly used for daylight high dynamic range compression fail in achieving an optimal visualisation of astronomical targets. We propose an extended multiscale algorithm based on compression of the dynamic range in the gradient domain. Our algorithm effectively compresses the dynamic range, enhances local contrast and avoids noise amplification. This is achieved with a multiscale representation of the image and the use of luminance information. Our results show a significantly improved visualisation of astronomical images compared to the standard gradient domain compression, as well as more robustness to noise and better artefact suppression.     

Gauss Gradient and SURF Features for Landmine Detection from GPR Images

Recently, ground-penetrating radar (GPR) has been extended as a well-known area to investigate the subsurface objects. However, its output has a low resolution, and it needs more processing for more interpretation. This paper presents two algorithms for landmine detection from GPR images. The first algorithm depends on a multi-scale technique. A Gaussian kernel with a particular scale is convolved with the image, and after that, two gradients are estimated; horizontal and vertical gradients. Then, histogram and cumulative histogram are estimated for the overall gradient image. The bin values on the cumulative histogram are used for discrimination between images with and without landmines. Moreover, a neural classifier is used to classify images with cumulative histograms as feature vectors. The second algorithm is based on scale-space analysis with the number of speeded-up robust feature (SURF) points as the key parameter for classification. In addition, this paper presents a framework for size reduction of GPR images based on decimation for efficient storage. The further classification steps can be performed on images after interpolation. The sensitivity of classification accuracy to the interpolation process is studied in detail.     

A multiresolution‐based method for the determination of the relative resolution between images: First application to remote sensing and medical images

Spatial resolution is a key parameter of all kind of images. This is of particular importance in fields as, for example, medicine or remote sensing. The nominal resolution of a positron emission tomography (PET) or nuclear magnetic resonance (NMR) scanners are directly related to the size, number, and position of the detectors in the scanner ring. Also, the nominal spatial resolution of the remote sensing satellites is a well‐known characteristic because it is directly related to the area in ground that represents a pixel in the detector. Nevertheless, in practice, the actual resolution of a medical scanner image or of an image obtained from a satellite is difficult to know precisely because it depends of many other factors. However, if we have two or more images of the same region of interest, obtained using similar or different instruments, it is possible to compare the relative resolution between them. In this paper we propose a wavelet‐decomposition‐based method for the determination of the relative resolution between two images of the same area. The method can be applied, in principle, to any kind of images. As example, we applied the method to pairs of remote sensing and medical images. In the case of remote sensing, we computed the relative resolution between SPOT‐3, LANDSAT‐5 and LANDSAT‐7 panchromatic and multispectral images taken under similar as well as under very different conditions. In the case of medical imaging, we computed the relative resolution between a pair of simultaneously obtained PET and NMR images of the same object. On the other hand, if we know the true absolute resolution of one of the images of the pair, we can compute the resolution of the other. Thus, in the last part of this paper, we describe a spatial calibrator that we have designed and constructed to help compute the absolute resolution of a single remotely sensed image, presenting an example of its use. © 2006 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 15, 225–235, 2005     

Quantitative chemical imaging of element diffusion into heterogeneous media using laser ablation inductively coupled plasma mass spectrometry, synchrotron micro-X-ray fluorescence, and extended X-ray absorption fine structure spectroscopy

Quantitative chemical imaging of trace elements in heterogeneous media is important for the fundamental understanding of a broad range of chemical and physical processes. The primary aim of this study was to develop an analytical methodology for quantitative high spatial resolution chemical imaging based on the complementary use of independent microanalytical techniques. The selected scientific case study is focused on high spatially resolved quantitative imaging of major elements, minor elements, and a trace element (Cs) in Opalinus clay, which has been proposed as the host rock for high-level radioactive waste repositories. Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS), providing quantitative chemical information, and synchrotron radiation based micro-X-ray fluorescence (SR-microXRF), providing high spatial resolution images, were applied to study Cs migration into Opalinus clay rock. The results indicate that combining the outputs achievable by the two independent techniques enhances the imaging capabilities significantly. The qualitative high resolution image of SR-microXRF is in good agreement with the quantitative image recorded with lower spatial resolution by LA-ICPMS. Combining both techniques, it was possible to determine that the Opalinus clay sample contains two distinct domains: (i) a clay mineral rich domain and (ii) a calcium carbonate dominated domain. The two domains are separated by sharp boundaries. The spatial Cs distribution is highly correlated to the distribution of the clay. Furthermore, extended X-ray absorption fine structure analysis indicates that the trace element Cs preferentially migrates into clay interlayers rather than into the calcite domain, which complements the results acquired by LA-ICPMS and SR-microXRF. By using complementary techniques, the quantification robustness was improved to quantitative micrometer spatial resolution. Such quantitative, microscale chemical images allow a more detailed understanding of the chemical reactive transport process into and within heterogeneous media to be gained.