Raster Data Model PDF Lecture (1)

Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani C...

Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani Chapter (1) ( 1 of 2) GIS Applications in Civil Eng. INTRODUCTION The vector data model uses the geometric objects of point, line, and polygon to represent spatial features with a clear spatial location and boundary such as streams, land parcels. Although ideal for discrete features with well-defined locations and shapes, the vector data model does not work well with spatial phenomena that vary continuously over the space such as precipitation, elevation, and soil erosion (Figure 1.1). A better option for representing continuous phenomena is the raster data model. The raster data model uses a regular grid to cover the space. The value in each grid cell corresponds to the characteristic of a spatial phenomenon at the cell location. And the changes in the cell value reflect the spatial variation of the phenomenon. Research on the raster data model has instead concentrated on new raster data, data structure, data compression, and integration of raster and vector data. A wide variety of data used in geographic information system (GIS) are encoded in raster format. They include digital elevation data, satellite images, digital orthophotos, scanned maps, and graphic files. This is why the help document of a GIS package typically has a long list of raster data types it supports. Raster data tend to require large amounts of the computer memory. Therefore, issues of data storage and retrieval are important to GIS users. GIS packages can display raster and vector data simultaneously, and can easily convert between these two types of data. In many ways, raster and vector data complement each other. Integration of these two types of data has therefore become a common and desirable feature in a GIS project. 1.1 ELEMENTS OF THE RASTER DATA MODEL A raster is also called a grid or an image in GIS. A raster represents a continuous surface, a raster is divided into rows, columns, and cells. Cells are also called pixels with images. The origin of rows and columns is typically at the upper-left corner of the raster. Rows function as y-coordinates and columns as x- coordinates. Each cell in the raster is explicitly defined by its row and column position. Raster data represent points with single cells, lines with sequences of neighboring cells, and polygons with collections of contiguous cells (Figure 1.2). Although the raster data model lacks the vector model’s precision in representing the location and boundary of spatial features, it has the distinct advantage of having fixed cell locations. In computing algorithms, a raster can be treated as a matrix with rows and columns, and its cell values can be stored in a two-dimensional array and handled as an arrayed variable in code. Raster data are therefore much easier to manipulate, aggregate, and analyze than vector data. 1.1.1 Cell Value Cell values in a raster can be categorical or numeric. A land cover raster, for example, contains categorical data with 1 for urban land use, 2 for forested land, 3 for water body, and so on. The land cover raster is also an example of an integer raster, as its cell values carry no decimal digits. A precipitation raster, on the other hand, contains numeric data such as 20.15, 12.23, and so forth. It is also an example of a floating-point raster, as its cell values include decimal digits. Textbook: Introduction to Geographic Information Systems, Kang-tsung Chang, McGraw-Hill (2019) Page 1 of 6 Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani Chapter (1) ( 1 of 2) GIS Applications in Civil Eng. A floating-point raster requires more computer memory than an integer raster. This difference can become an important factor for a GIS project that covers a large area. There are a couple of other differences. ▪ First, an integer raster has a value attribute table for access to its cell values, whereas a floating-point raster usually does not because of its potentially large number of cell values. ▪ Second, individual cell values can be used to query and display an integer raster but value ranges, such as 12.0 to 19.9, must be used on a floating-point raster. The chance of finding a specific value in a floating-point raster is very small. Where does the cell value register within the cell? The answer depends on the type of raster data operation. Typically, the cell value applies to the center of the cell in operations that involve distance measurements. Examples include resampling pixel values and calculating physical distances. Many other raster data operations are cell-based, instead of point-based, and assume that the cell value applies to the entire cell. 1.1.2 Cell Size The cell size of a raster refers to the size of the area represented by a single cell. If a raster has a cell size of 100 m2, it means each side of its cell is 10 meters in length. The raster is typically called a 10-meter raster. The cell size determines the spatial resolution of a raster. A 10-meter raster has a finer (higher) resolution than a 30-meter raster. A large cell size cannot represent the precise location of spatial features, thus increasing the chance of having mixed features such as forest, pasture, and water in a cell. These problems lessen when a raster uses a smaller cell size. But a small cell size increases the data volume and processing time. 1.1.3 Cell Depth The cell depth of a raster refers to the number of bits for storing cell values. A bit (short for binary digit), the smallest data unit in a computer, has a single binary value of either 0 or 1. A byte is a sequence of bits, with 8 bits equaling 1 byte. A higher cell depth means that the cell can store a wider range of values. For example, an 8-bit raster can store 28 or 256 possible values while a 16-bit or 216 raster can store 65,536 possible values. The way in which the cell values are stored can determine the data volume; specific examples relating cell depth to data volume are offered in Box 1.1. 1.1.4 Raster Bands A raster may have a single band or multiple bands. Each cell in a single-band raster has only one cell value. An example of a single-band raster is an elevation raster, with one elevation value at each cell location. Each cell in a multiband raster is associated with more than one cell value. An example of a multiband raster is a satellite image, which may have five, seven, or more bands at each cell location. Textbook: Introduction to Geographic Information Systems, Kang-tsung Chang, McGraw-Hill (2019) Page 2 of 6 Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani Chapter (1) ( 1 of 2) GIS Applications in Civil Eng. 1.1.5 Spatial Reference Raster data must have the spatial reference information so that they can align spatially with other data sets in a GIS. For example, to superimpose an elevation raster on a vector-based soil layer, we must first make sure that both data sets are based on the same coordinate system. A raster that has been processed to match a projected coordinate system is often called a georeferenced raster. How does a raster match a projected coordinate system? First, the columns of the raster correspond to the x-coordinates, and the rows correspond to the y- coordinates. Because the origin of the raster is at the upper-left corner, as opposed to the lower-left corner for the projected coordinate system, the row numbers increase in the direction opposite that of the y-coordinates. Second, the projected coordinates for each cell of the raster can be computed by using the x-, y- coordinates of the area extent of the raster. The following example is illustrative. Suppose an elevation raster has the following information on the number of rows, number of columns, cell size, and area extent expressed in UTM (Universal Transverse Mercator) coordinates: Rows: 463, columns: 318, cell size: 30 meters x-, y-coordinates at the lower-left corner: 499995, 5177175 m x-, y-coordinates at the upper-right corner: 509535, 5191065 m We can verify that the numbers of rows and columns are correct by using the bounding UTM coordinates and the cell size: Number of rows = (5191065 − 5177175)/30 = 463 Number of columns = (509535 − 499995)/30 = 318 We can also derive the UTM coordinates that define each cell. For example, the cell of row 1, column 1 has the following UTM coordinates (Figure 1.3): 1.3 499995, 5191035 or (5191065 − 30) at the lower-left corner 500025 or (499995 + 30), 5191065 at the upper-right corner 500010 or (499995 + 15), 5191050 or (5191065 − 15) at the cell center 1.2 SATELLITE IMAGES Remotely sensed satellite data are familiar to GIS users. Satellite systems can be divided into passive and active (Table 1.1). Passive systems, often referred to as optical systems, acquire spectral bands from the electromagnetic spectrum reflected or emitted from the Earth’s surface. Measured by wavelength (e.g., micrometer or µm), these spectral bands are recorded in the range of visible light (0.4 − 0.7 µm), near infrared (0.7 − 1.3 µm), and shortwave infrared (1.3 − 3.0 µm). Optical satellite images can be panchromatic or multispectral. Panchromatic images have a single band, which can be displayed in shades of gray, whereas multispectral images have multiple bands, which can be displayed in color composites. Textbook: Introduction to Geographic Information Systems, Kang-tsung Chang, McGraw-Hill (2019) Page 3 of 6 Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani Chapter (1) ( 1 of 2) GIS Applications in Civil Eng. Active systems, commonly referred to as synthetic aperture radar (SAR), provide their energy to illuminate an area of interest and measure the radar waves reflected or scattered back from the Earth’s surface. SAR data can be classified by the wavelength, with L-band the longest, X-band the shortest, and C-band in between. The chief advantage of SAR is that it can work in the presence of clouds, rain, or darkness. For both passive and active systems, the spatial resolution of the satellite image refers to the pixel size. For example, a spatial resolution of 30 meters means that each pixel corresponds to a ground area of 900 m2. The spatial resolution is fixed for optical satellite images, but it can change for SAR images depending on several parameters such as the acquisition mode, wavelength, bandwidth, and incidence angle. Many countries have developed satellite programs since the late 1980s. It is impossible to list all of them. The following sections cover a select set of satellite image examples. 1.2.1 Landsat The U.S. Landsat program, started by the National Aeronautics and Space Administration (NASA) and the U.S. Geological Survey (USGS) in 1972, has produced the most widely used imagery worldwide (http://landsat.usgs.gov/). Landsat 1, 2, and 3 acquired images by the Multispectral Scanner (MSS) with a spatial resolution of about 79 meters. Aboard Landsat 4 in 1982, the Thematic Mapper (TM) scanner obtained images with seven spectral bands (blue, green, red, near infrared, midinfrared I, thermal infrared, and midinfrared II) and with a spatial resolution of 30 meters. A second TM was launched aboard Landsat 5 in 1984. Landsat 6 failed to reach its orbit after launch in 1993. Landsat 7 was launched in 1999, carrying an Enhanced Thematic Mapper Plus (ETM+) sensor designed to seasonally monitor small-scale processes on a global scale, such as cycles of vegetation growth, deforestation, agricultural land use, erosion and other forms of land degradation, and urbanization. Landsat 8 was launched in 2013 with the Operational Land Imager (OLI), which provides similar spectral bands as Landsat 7, along with a new deep blue band (band 1) and a new shortwave infrared band (band 9). Additionally, Landsat 8 carries the thermal infrared sensor, which provides two thermal bands. Landsat 9 was launched in 2021 with OLI2, which is identical to OLI onboard Landsat 8. Name Orbit Type Orbit Height (km) Repeat Cycle (days) Launched in Out of service since Organisation Landsat 1 Sun Synchronous 907 18 1972 1978 NASA - USA Landsat 2 Sun Synchronous 908 18 1975 1982 NASA - USA Landsat 3 Sun Synchronous 915 18 1978 1983 NASA - USA Landsat 4 Sun Synchronous 705 16 1982 2001 NASA - USA Landsat 5 Sun Synchronous 705 16 1984 2013 NASA - USA Landsat 7 Sun Synchronous 705 16 1999 NASA - USA Landsat 8 Sun Synchronous 705 16 2013 NASA - USA Landsat 9 Sun Synchronous 705 16 2021 NASA - USA 1.2.2 SPOT The French SPOT satellite series began in 1986. Each SPOT satellite carries two types of sensors. SPOT 1 to 4 acquire single-band imagery with a 10-meter spatial resolution and multiband imagery with a 20-meter resolution. SPOT 5, launched in 2002, sends back images of 5 and 2.5 meters in single- band, and 10 meters in multiband. SPOT 6 and SPOT 7, launched in 2012 and 2014 respectively, provide panchromatic images with a resolution of 1.5 meters and multispectral images (blue, green red, and near-infrared) with a resolution of 6 meters. Name Orbit Type Orbit Height (km) Repeat Cycle (days) Launched in Out of service since Organisation SPOT 1 Sun Synchronous 832 26 1986 2003 CNES - France SPOT 2 Sun Synchronous 832 26 1990 2009 CNES - France SPOT 3 Sun Synchronous 832 26 1993 1996 CNES - France Textbook: Introduction to Geographic Information Systems, Kang-tsung Chang, McGraw-Hill (2019) Page 4 of 6 Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani Chapter (1) ( 1 of 2) GIS Applications in Civil Eng. SPOT 4 Sun Synchronous 832 26 1998 2013 CNES - France SPOT 5 Sun Synchronous 832 26 2002 2015 CNES - France SPOT 6 Sun Synchronous 694 26 2012 Airbus Defence and Space SPOT 7 Sun Synchronous 694 26 2014 Airbus Defence and Space SPOT images are now parts of products distributed by Airbus Defense and Space (http://www.intelligence-airbusds.com/). Airbus Defense and Space also markets very high-resolution Pleiades satellite images (Table 1.3). 1.2.3 Digital Globe Digital Globe is an American company that specializes in high resolution satellite images (http://www.satimagingcorp.com/). With its purchase of GeoEye in 2013, Digital Globe has expanded the products it carries, including Ikonos, QuickBird, GeoEye-1, and WorldView (1-4). The newest WorldView-4 has a spatial resolution of 31 centimeters for panchromatic and 1.24 meters for multispectral (red, green, blue, and infrared). Table 1.3 shows the spatial resolutions of these products in both panchromatic and multispectral. It should be noted that both Ikonos and QuickBird were decommissioned in 2015, although their archival images are still available. Name Orbit Type Orbit Height (km) Repeat Cycle (days) Launched in Out of service since Organisation GeoEye-1 Sun Synchronous 681 2008 Maxar - USA IKONOS Sun Synchronous 681 3 1999 2015 Maxar - USA Quickbird Sun Synchronous 450 2001 2015 Maxar - USA WorldView-1 Sun Synchronous 496 2007 Maxar - USA WorldView-2 Sun Synchronous 770 1.1 2009 Maxar - USA WorldView-3 Sun Synchronous 617 1 2014 Maxar - USA WorldView-4 Sun Synchronous 617 1 2016 2019 Maxar - USA WorldView-Legion Other 450 2023 Maxar - USA 1.2.4 Sentinel The European Space Agency provides both active and passive satellite data through the Sentinel missions (https://sentinel.esa.int/web/sentinel/home). Like Landsat images, Sentinel data are free for download. Sentinel-1, launched in 2014, acquires C-band SAR images globally at an approximate spatial resolution of 20 meters. Sentinel-1, therefore, joins TerraSAR-X, RADARSAT-2, and COSMOS-SkyMed in offering SAR images. Sentinel-2 launched in 2015, gathers data with 13 bands in the visible, near-infrared, and short-wave infrared ranges: four bands at a spatial resolution of 10 meters, six bands at 20 meters, and three bands at 60 meters. Sentinel-3, launched in 2016, is primarily an ocean mission for such tasks as gathering data on variations in global sea level. Textbook: Introduction to Geographic Information Systems, Kang-tsung Chang, McGraw-Hill (2019) Page 5 of 6 Faculty of Engineering Lecture (1) CE 797: Special Topic Civil Eng. Department Raster Data Model Digital Elevation Models and Dr. Suhail A. Almadani Chapter (1) ( 1 of 2) GIS Applications in Civil Eng. Name Orbit Type Orbit Height (km) Repeat Cycle (days) Launched in Out of service since Organisation Sentinel-1A Sun Synchronous 693 6 2014 ESA Sentinel-1B Sun Synchronous 693 6 2016 ESA Sentinel-1C Sun Synchronous 693 2022 ESA Sentinel-1D Sun Synchronous 693 2023 ESA Sentinel-2A Sun Synchronous 786 5 2015 ESA Sentinel-2B Sun Synchronous 786 5 2017 ESA Sentinel-2C Sun Synchronous 786 2024 ESA Sentinel-2D Sun Synchronous 786 2025 ESA 1.2.5 Terra Satellite In 1999, NASA’s Earth Observing System launched the Terra spacecraft to study the interactions among the Earth’s atmosphere, lands, oceans, life, and radiant energy (heat and light) (http://terra.nasa.gov/About/). Terra carries several instruments, two of them are ASTER and MODIS instruments. ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) is the only high spatial resolution instrument designed for applications in land cover classification and change detection. ASTER’s spatial resolution is 15 meters in the visible and near infrared range, 30 meters in the shortwave infrared band, and 90 meters in the thermal infrared band. MODIS (Moderate Resolution Imaging Spectroradiometer) provides continuous global coverage every one to two days and collects data from 36 spectral bands with spatial resolutions ranging from 250 to 1000 meters. Name Description No of Bands Mission ASTER Advanced Spaceborne Thermal Emission and Reflection Radiometer 14 Terra MODIS Moderate Resolution Imaging Spectroradiometer (PFM on Terra, FM1 on Aqua) 36 Aqua, Terra 1.2.6 Downloading Remote Sensing Data There are many websites worldwide that offer free and/or paid global remote sensing data. Some of these websites belong to government agencies such as the United States geological Survey (USGS) and the National Aeronautics and Space Administration (NASA). The USGS and NASA together offer several web services to find, preview, and download remote sensing data. Three of these web services are EarthExplorer, Glovis, and LookViewer. Remote Sensing data are satellite imageries, airborne imageries, radar imageries, lidar data, and digital elevation data. In this course we will log in to the USGS EarthExplorer website to find, download, and view free imageries of Landsat 8, Landsat 9, Sentinel-2 satellites, and DEM data of the Shuttle Radar Topography Mission SRTM. For information about Landsat see the link https://www.usgs.gov/land- resources/nli/landsat/landsat-satellite-missions. To download Landsat imageries, go to EarthExplorer webservice at https://earthexplorer.usgs.gov/. For information about Sentinel missions see the link https://sentinel.esa.int/web/sentinel/missions. To download Sentinel imageries, go to Copernicus Data Space Ecosystem website at Copernicus Data Space Ecosystem. To download free data from these websites, you must create a personal account in each one. Textbook: Introduction to Geographic Information Systems, Kang-tsung Chang, McGraw-Hill (2019) Page 6 of 6