Back top previously visited page (this function is disabled)Back to homepage (P)Back to previous page (E)Jump to next page (V)Display terms (this function is disabled)List of terms (this function is disabled)Print this page (this function is disabled)Jump to sitemap (D)Search in current page (this function is disabled)Help (S)

Building Spatial Databases – Theory / Principles of the raster data model

Learners guide


In this unit the basic concepts of raster-based systems will be introduced, such as the electromagnetic spectrum, colour depth, colour models, resolution, etc.


To make the self-test successfully

Principles of the raster data model

Raster-based data model is the other major data model of geoinformation technologies. The major difference between vector and raster-based data models are in the data structure and in the projection of surface objects. The level of details of surface data is significantly different along with the data processing methodologies. For instance, the vector data models are based on relation databases, whereas the raster data models heavily rely on different image processing algorithms. These algorithms rely on mathematical methods like Fourier transformation or multivariable statistical knowledge such as cluster analysis or main-component-analysis.

The electromagnetic spectrum

Figure 47. Range of the electromagnetic spectrum. RGB defines the visible range, whereas near IR and middle IR and far IR are infra-red ranges. As described in Figure 47, electromagnetic ranges are used in remote sensing. Raster-based GIS (geoinformation systems) uses these ranges. Light emitted from the sun and reflected from the surface of the earth are sensed and stored in image format, which are stored on hard disks. Let us examine the method of specifying a digital image.

Pixel, spectral resolution, wavelength

The most basic building blocks of a digital image are called pixels. This is the smallest image point that can be created by image creation tools. An image creation tool can be satellites, scanners or a digital camera. A pixel is optically homogeneous, namely its colour and intensity are the same. The captured area in pixels has real measurements, namely scale.

The term spatial resolution is closely related to the size of pixels and the size of the captured image. The resolution is proportionally equivalent to the number of pixels it contains on a captured area. For example, an image with the resolution of n*m pixels (n rows and m columns) on a smaller area is considered as a high resolution image, however if the captured area is vast then the resolution is quite small, although in physical size it does not change at all.

We must also know the spectral range in which the image was taken (by a digital camera). Simple digital cameras are working in the spectral range of the visible light, whereas other devices are working in the spectral range of infra-red. In the era of satellites for sensing, various tools have been developed for working in different spectral ranges. Some of them are working in visible spectral ranges like SPOT and ICONOS, whereas others are operating in a wide spectral range like LANDSAT TM, which operates on RGB and four infra-red bands. A device spectral resolution number is proportional to the number of bands it can sense. The most advanced devices are hyperspectral cameras, which can process hundreds of electromagnetic spectra into bands. There are also many devices operating in radio frequency ranges, based on radar concept for sensing surfaces. The major difference between the radar and the above mentioned devices is that radar-based devices emit waves to detect and sense data, whereas others are using emissions from the sun.

Colour depth and transparency

The representation of colours has evolved with the development of computers. In the beginning, colours were represented in 4 bits, later in 8 bits. Coloured images were poor in tones, which is rather suitable for presenting as graphic. Nowadays colours are represented in 24, moreover in 36 bits on computers; therefore, digital images are capable of displaying millions of tones. Figure 48 describes the bits with representable colours.

Figure 48. Colours by pixel in numeric representation.

Spatial images are using the above mentioned colour representations; moreover, during the pre-processing states the algorithms usually transform images from one model to another (e.g. thematic representation). It cannot be overemphasized that a black and white image represented in 1 bit is not equivalent to the black-and-white image we know of. Namely, in the photos we are familiar with the greyscale of 256 grades (8 bits) represents the objects in tones, while the 1-bit representation model means only black or white. This 1-bit colour model is applied in special cases, such as edge detection, where the edge is represented by 1 value, and if there are no edges, the pixel value is 0. The 1-bit model can be applied if Boolean values should be represented, such as in the case of masking or handling target areas.

Colour models

RGB model

Let us take a three dimensional coordinate system where the axis represents the colours.

Figure 49. RGB colour cube.

The RGB colour model is a composite of Red, Green and Blue colours, which has been created by:

Colour = a *Red + b * Green + c * Blue

where a,b and c coefficients stand for the ratio of RGB components in the resulted colour. By applying the 24-bit based colour representation mode, we will get 2 on the power of 24 shades of colours, which means that 2^8 of Red, Green and Blue shades can be represented, that is 256 intensity values can be set for each base colour.

HSI model

The HSI colour model is fundamentally different from the RGB model. Whereas RGB uses the Cartesian coordinate system, HSI uses the polar coordinate system for representing colours as described on Figure 50.

Figure 50. HSI colour cone.

HSI is the abbreviation of Hue, Saturation and Intensity words. Remote sensing software programs support both colour models. There are many more colour models like CMYK (mostly used in desktop publishing and in the printing industry, although these kinds of models are not used in remote sensing.

Back to table of contents (J)

Új Széchenyi terv
A projekt az Európai Unió támogatásával, az Európai Szociális Alap társfinanszirozásával valósul meg.

A Társadalominformatika: moduláris tananyagok, interdiszciplináris tartalom- és tudásmenedzsment rendszerek fejlesztése az Európai Unió támogatásával, az Európai Szociális Alap társfinanszírozásával, az ELTE TÁMOP 4.1.2.A/1-11/1-2011-0056 projekt keretében valósult meg.
Generated with ELTESCORM framework