Remote sensing images represent an objective record of the spectrum relating to the physical properties and chemical composition of Earth surface materials. Image processing is a vital tool for the extraction of thematic and quantitative information from raw image data. In this book, authors have presented a unique combination of tools, techniques, and applications. The book is divided into three parts, with the first part introducing essential image processing techniques for remote sensing. The second part looks at Geographical Information Systems (GIS) and begins with an overview of the concepts, structures and mechanisms by which GIS operate. Finally, the third part introduces Remote Sensing Applications. The authors have described the key concepts and ideas with clarity and in a logical manner and have also included numerous relevant conceptual illustrations. The book contains twenty three chapters, all of which are well written. Each chapter ends with key point remarks and important questions to test the reader’s understanding. 

Part I Image Processing

Chapter 1 Digital Image and Display

This chapter gives an introduction to digital images and the fundamentals of a monochrome and a colour displays. A digital image is a two-dimensional array of numbers. Each cell of a digital image is called a pixel and the number representing the brightness of the pixel is called a digital number (DN). This chapter also describes the technique to display a monochrome image as a colour image—pseudo colour display—where the sequence of gray levels is assigned to colours of increasing spectral wavelength and intensity.

Chapter 2 Point Operations (Contrast Enhancement)

This chapter discusses the contrast enhancement techniques for optimizing the image contrast and brightness for visualization or for highlighting information in particular DN ranges. 

Chapter 3 Algebraic Operations (Multi-image Point Operations)

This chapter talks about the basic algebraic operations upon a multi-spectral image and their applications in image enhancement. Major applications are the selective enhancement of the spectral signatures of intended targets in a multi-spectral image. Also, the formulae should be composed on the basis of spectral or physical principles, and designed for the enhancement of particular targets. This procedure, from spectral analysis to composing an algebraic formula, is generally referred to as supervised enhancement. The solar illumination on a land surface varies with terrain slope and aspect, which results in topographic shadows.

Chapter 4 Filtering and Neighbourhood Processing

This chapter illustrates image filtering techniques for enhancing lineaments (that may represent significant geological structures such as faults, veins, or dykes) and image texture. Digital filtering can be implemented based on convolution in the spatial domain or using the Fourier transform (FT) in the frequency domain. The various low-pass and high-pass filters are also described in this chapter. Gradient is the first derivative of DN change in a direction and gives a measurement of DN slope. Laplacian, as the second derivative, is a scalar that measures the change rate of DN slope. FFT-based, frequency-adaptive filters are also explained introduced for advanced readers.

Chapter 5 RGB–IHS Transformation

This chapter deals with the RGB-IHS and IHS- RGB transformation principles, which allow us to manipulate these qualities with great flexibility. Two decorrelation stretch techniques based on saturation stretch are also discussed. The direct decorrelation stretch technique performs a direct saturation stretch (DDS) without using RGB–IHS and IHS–RGB transformations, which is explained clearly with enough mathematical derivations. The hue of a colour is the spectral property coding. An HRGB colour composite technique is thus introduced that can code the spectral property of up to nine spectral bands into various colours to generate an information-rich colour image without the effects of topographic shadows.

Chapter 6 Image Fusion Techniques

Recently, image fusion techniques have become the most widely used methods for integrating images or raster datasets of different spatial resolutions (or with different properties) to formulate new images. This chapter introduces the three simplest and most popular image fusion techniques:  RGB–IHS transformation, Brovey transform (intensity modulation), and smoothing filter-based intensity modulation.  Both the IHS and Brovey transform image fusion techniques can cause colour distortion. This issue can overcome by smoothing filter-based intensity modulation (SFIM), which also improves spatial resolution.

Chapter 7 Principal Component Analysis

This chapter discusses the principal component analysis (PCA) technique, which is a general method of analysis for correlated multi-variable datasets, and its applications The principal components are the image data representation in the coordinate system formed by the axes of the ellipsoid data cluster, and hence PCA is a coordinate rotation operation to rotate the coordinate system of the original image bands to match the axes of the ellipsoid of the image data cluster. This is explained by looking at methods.

Chapter 8 Image Classification

This chapter describes the most commonly used image classification algorithms and post-classification processing techniques. These methods are essentially (1) multivariable statistical classifications that achieve data partition in the multi-dimensional feature space of multi-layer image data, such as a multi-spectral remotely sensed image, or (2) segmentation based on both statistics and spatial relationships with neighbouring pixels. The statistical classification has two major branches: unsupervised and supervised classifications. Both are explained in this chapter.

Chapter 9 Image Geometric Operations

In remote sensing applications, geometric operations are mainly used for the co-registration of images of the same scene acquired by different sensor systems, at different times, or from different positions, and for rectifying an image to fit a particular coordinate system. In this chapter, the geometric transformation for remotely sensed data is introduced and two major steps of this transformation are illustrated. First, establish the polynomial deformation model, usually done using ground control points (GCPs). Then, resample the image based on the deformation model. This includes resampling image pixel positions (coordinates) and DNs.

Chapter 10 Introduction to Interferometric Synthetic Aperture Radar Techniques

This chapter introduces several Interferometric synthetic aperture radar (InSAR) techniques for 3D terrain representation, for quantitative measurements of terrain deformation, and for the detection of random land surface changes. The differential InSAR (DInSAR) is used for the measurement of terrain deformation. InSAR coherence technique is used for random change detection, and the ratio coherence technique separates the spatial and temporal decorrelation.

Part II Geographical Information Systems

Chapter 11 Geographical Information Systems

This chapter gives an introduction to Geographical Information Systems (GIS) and their applications. Software tools for data processing to acquire different sorts of information are discussed. A detailed differentiation between a cartographic map, a GIS, and thematic mapping is then explained. A map is an analogue spatial database which requires perception and interpretation to extract the embedded information, but once on paper, however, it is fixed and cannot be modified. A GIS display of a map does not require every piece of information to be visible at the same time. It can also change the depiction of a particular object according to the value of one of its attributes.

Chapter 12 Data Models and Structures

The data describing a part of the Earth’s surface or the features found on it is called geographic or spatial data. The objects located on the surface of the Earth are called geographic features, and their positions can be measured and described. This chapter discusses the structures and models for representation of the data. There are two basic types of structures used to represent the features or objects, namely raster and vector data.  Rasters, images or grids consist of a regular array of digital numbers or DNs, representing picture elements or pixels. Vector, or discrete data, store the geometric form and location of a particular feature, along with its attribute information describing what the feature represents.

Chapter 13 Defining a Coordinate Space

This chapter explains the issues of data acquired that are different or from an unknown coordinate system. Assigning the ‘coordinates’ of time and location such that these data can be understood is an important task in GIS functioning. To establish systems describing an object’s location requires the consideration of the shape of the Earth. The coordinates are best described by the latitude and longitude system, while the shape of Earth is approximated by ellipsoid of rotation or spheroid. Map projection is an accepted means for fitting all or part of the curved surface of Earth to the flat surface or plane. This projection cannot be made without distortion of shape, area, distance, direction or scale.

Chapter 14 Operations

This chapter deals with operations to be performed on spatial data. The chapter is focused on operations that assume raster inputs. Map algebra is used for manipulating raster variables defined over a common area. It also describes calculations within and between GIS data layers to produce a new layer. Map algebra operations can also be performed on vector data. A local operation involves the production of an output value as a function of the value(s) at the corresponding locations in the input layer(s).

Chapter 15 Extracting Information from Point Data: Geostatistics

This chapter deals with two topics: gaining a better understanding of the data and dealing with incomplete data. Geostatistics is concerned with the description of patterns in spatial data; each known data point has a geographic location and a value, and the connection between them is exploited to help predict values at the unknown locations. The chapter gives an overview of the main issues and methods involved in extracting and exploiting statistical data. 

Chapter 16 Representing and Exploiting Surfaces

This chapter describes the surface models and the effect of the fractal nature of surface features. Gradient, aspect, and curvature are all phenomena which vary at different scales of topography, and so scales of observation. Measuring any of these phenomena in the field yields very different results according to the distance over which they are being measured.

Chapter 17 Decision Support and Uncertainty

Uncertainty in the GIS data arises because the data we have are never complete, and our knowledge and understanding of a problem are flawed or limited, because of natural variation, measurement error, or out-of-date information. A spatial decision support system (SDSS) is a knowledge-based information system that helps decision makers to identify areas where there are unacceptable levels of risk associated with various predictive outcomes, so that they can then select appropriate courses of action. The methods for reducing uncertainty include defining and standardizing technical procedures, improving education and training (to improve awareness), collecting data more rigorously, increasing spatial/ temporal data resolution during data collection, field checking of observations, better data processing methods and models, and developing an understanding of error propagation in the algorithms used.

Chapter 18 Complex Problems and Multi-Criteria Evaluation

This chapter discusses the procedures by which we deal with complex geospatial problems to which there may be many contributing processes and pieces of evidence. There are a number of different approaches to multi-criteria decision making and analysis, with the aim of estimating suitability or favourability across a region. They are often divided into two broad categories: the knowledge-driven approach and data-driven approach, and of the latter there are two further kinds.

Part III Remote Sensing Applications

Chapter 19 Image Processing and GIS Operation Strategy

This chapter describes how the processing, interpretation, and analysis of multi-source image and map data should be approached to produce thematic maps for a typical project. From defining the project goals, to extracting the real image information, this chapter presents a simple and generic formula and the steps for doing so.

Chapter 20 Thematic Teaching Case Studies in SE Spain

This chapter discusses several teaching case studies on specific themes, using image data of SE Spain, to demonstrate remote sensing applications in earth and environmental sciences. The first case study demonstrates how to design simple and effective image processing techniques to map gypsum outcrops and extract gypsum quarries based on image spectral profile analysis. The second case study demonstrates the application of multispectral and multi-resolution remote sensing for mineral exploration via image processing. The third case study makes an estimation of the extent of vegetation and plasticulture in one small area. The final case study involves the use of multi-spectral imagery to improve the understanding of the regional geology, tectonics, and hydrology in the Tabernas–Andarax Basin of Almeria Province, Spain.

Chapter 21 Research Case Studies

This chapter is based on the author’s published research papers. The cases described [in this chapter] deal with topics such as the design and development of the most effective image processing techniques and strategies for extracting the required thematic information from images and the establishment of the most representative and powerful GIS model to serve the objectives of the project. In the first case study, multi-temporal image data were used to assess the change of vegetation coverage in the three parallel rivers region of Yunnan province, China. The second case study presents a regional assessment of landslide hazard in the Three Gorges area of China, where a multi-variable elimination and characterization model, employing geometric mean and Boolean decision rules, has been applied to a multi-criterion image dataset to categorize the area into a series of potential landslide hazard levels, which are presented in map form. The third case study describes the Land surface change detection in a desert area in Algeria using multi-temporal ERS SAR coherence images. A multi-temporal SAR coherence image presents an objective record of irregular land surface changes between two SAR image acquisitions as decoherence features. The technique is most effective for detecting changes in a largely stable environment.

Chapter 22 Industrial Case Studies

This chapter describes two industrial case studies. The first case describes the data and methodology used to enable a multi-disciplinary assessment of prospectivity for a number of economic commodities, namely nickel, copper, and PGEs (Platinum Group Elements). The work in the second case study, involves evaluation, through remote sensing, of the region around the city of Hargeisa, in Somalia, to identify potential bulk water resources for the city. 

Chapter 23 Concluding Remarks

This chapter gives a summary on essential image processing and GIS techniques for remote sensing applications.

Apendices

Overall, the book covers a variety of topics in the areas of image processing for remote sensing and GIS. Appendix A describes imaging sensor systems and remote sensing satellite details, and Appendix B describes online resources for information about technical details and data sources.

BOOKSBOOKSBOOKS

Essential Image Processing and GIS for Remote Sensing

 

by

Jian Guo Liu and Philippa Mason

Wiley, 2009

 

Reviewed by

Tanish Zaveri (India)

Click on the image above  to go to the publisher’s web page for this book where you will find a description of the book and Hallmark Features.

Click here for

FREE BOOKS!