Нефть и песок О стали Компрессор - подбор и ошибки Из истории стандартизации резьб Соперник ксерокса - гектограф Новые технологии производства стали Экспорт проволоки из России Прогрессивная технологическая оснастка Цитадель сварки с полувековой историей Упрочнение пружин Способы обогрева Назначение, структура, характеристики анализаторов Промышленные пылесосы Штампованные гайки из пружинной стали Консервация САУ Стандарты и качество Технология производства Водород Выбор материала для крепежных деталей Токарный резец в миниатюре Производство проволоки Адгезия резины к металлокорду Электролитическое фосфатирование проволоки Восстановление корпусных деталей двигателей Новая бескислотная технология производства проката Синие кристаллы Автоклав Нормирование шумов связи Газосварочный аппарат для тугоплавких припоев
Главная --> Промиздат -->  Map principle 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 ( 64 ) 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127

ing the appropriate module, such as r.in.gdal (supports various formats), r.in.tiff (TIFF), r.in.bin (binary data), i.in.erdas (ERDAS/LAN format) etc. For example, to import an ERDAS/LAN file run:

i.in.erdas input=landsat5.Ian output=tm

This imports the landsatS.lan data set (a LAN file usually contains several channels) into the GRASS LOCATION. The output parameter is a prefix, to which channel numbers are added during import.

Automated creation of a LOCATION. A convenient way to generate a new LOCATION from GIS data or an image data set is to use the module r.in.gdal. It is based on the GDAL library and accepts many important GIS and satellite data formats. It optionally allows the user to generate a new LOCATION based on the map metadata from the imported data set. The completeness of metadata depends on the image format, i.e. only some data sets will provide projection information. In terms of raw satellite data, only the xy boundary settings are usually available, so the resulting new LOCATION is generated as xy type. In any case, this method minimizes the efforts to define

a LOCATION.

To generate a new LOCATION and to import a raster data set into it, r.in.gdal has to be used within another GRASS LOCATION (since the module needs a GRASS environment; as an alternative, see Section 11.3 for a script to generate a LOCATION directly from a GIS raster file outside of GRASS). A new LOCATION is only generated when the location parameter is additionally specified. In this case the data stream is written to this new

LOCATION into MAPSET PERMANENT. The current LOCATION, where

the module was started, is not modified. We also recommend specifying the flag -e, which extends a LOCATION if required.

To provide an example, we create a new LOCATION for a LANDSAT-TM7

scene (Spearfish area, 12 July 2002, in UTM/WGS84 projection, GeoTIFF format, available from GLCF Maryland3). Since the map datum and ellipsoid do not match the sample Spearfish LOCATION, we cannot import the satellite scene directly. But we can use the Spearfish UTM/NAD27 LOCATION to run r.in.gdal:

gdalinfo p033r02 9 7t20000712 zl3 nnl0.tif r.in.gdal -e input=p033r029 7t20000712 zl3 nnl0.tif \ output=tm7 2000071.1 location=landsat

Now we have to leave GRASS and restart it with the new LOCATION land-sat to continue the data import for the other channels using r.in.gdal. To use this LANDSAT-TM7 scene in the sample Spearfish location, change back to it and run r.proj to reproject the channels from the new landsat LOCATION into the current LOCATION.



The GeoTIFF format stores the scene projection information, so it will be used by r.in.gdal to define the new LOCATION. In other cases, when the satellite data are not geocoded, ground control points (GCPs) are required to reference the image to a reference map. Sometimes, GCPs are provided by the data vendor and stored in the image metadata. If detected during import, they will be written into a so-called POINTS text file. This GCPs POINTS file can be used later for image rectification. We will explain this in greater detail in Section 9.4.1. Usually GCPs are provided in latitude-longitude coordinates which may not be the desired coordinate system. The optional target parameter of r.in.gdal allows us to transform the GCPs on the fly to another map coordinate system. The required definitions will be read from the desired georeferenced LOCATION, its name must be given by parameter target. An example for a SAR SLC data set in CEOS format:

r.in.gdal input=/cdrom/scenel/dat 01.OQ1 output=sar\ location=sar raw target=gauss boaga

This will import the file dat 01.001 from the mounted CD-ROM into a new LOCATION sar raw with an image prefix sar (which in this case will result in the GRASS image names sar.real and sar.imaginary). The original latitude-longitude GCPs describing the four corners of the image data set are re-projected on the fly to Gauss-Boaga projection as defined in the LOCATION gauss boaga and written to a POINTS file in the new LOCATION sar raw.

Because the HDF format is supported by GDAL, geocoded data sets such as ASTER/TERRA and MODIS/TERRA can be imported with r.in.gdal.

Installing the Imagery sample data set. For the Spearfish region, an image sample data set is available on the GRASS Web site ( Imagery package). The installation is done in the same way as for the Spearfish data set. The procedure is described in Section 3.1.3; you just have to substitute the package names. After extraction, you will find a directory imagery in your GRASS DATABASE. Start GRASS with the LOCATION imagery and specify your name as MAPSET. The following images are available in xy coordinates: gs13.1, gs14.1 (stereo aerial images from 22. August 1971, camera information is in file imagery/user1/gscam), nhap.1, nhap.2, nhap.3 and nhap.enh (multispectral NHAP, National High Altitude Photography4, a set of aerial infrared image in three bands and a color composite), and a SPOT-1 HRV/PAN scene with channels spot.ms.1 (green, HRV1), spot.ms.2 (red, HRV2), spot.ms.3 (NIR, HRV3), spot.p (panchromatic, image scene WRS reference: k=564/j=260, acquisition time: 17:58:50 UTC, date: 27. May 1989). The spectral ranges of SPOT-I HRV are: 0.50-0,59 (green, HRVI), 0.61-0.68 jum (red, HRV2), 0.79-0.89 lm (near infrared, HRV3), all at 20 m spatial resolution. The spectral range of the panchromatic channel is 0.510.73 ;um at 10 m spatial resolution.



9.3. UNDERSTANDING A SATELLITE DATA SET

The descriptions in this section are based on the SPOT-1 images provided in the Imagery data set. Start GRASS with imagery LOCATION and your name as MAPSET. The original image data are stored in MAPSET PERMANENT.

9.3.1 Managing channels and colors

A multispectral data set consists of various channels which represent portions of the spectrum. In the case of LANDSAT-TM5 and TM7, the visible spectrum with base colors blue, green and red is mostly covered as well as part of the infrared and thermal spectrum (see above Section 9.1 for spectrum details). Other satellites such as SPOT and ASTER do not provide the blue channel. To visually explore the imagery, we often need to analyze and modify its colors.

Bits, channels and colors. In general, each channel stores the local brightness level pixel-wise at the observed range of wavelength. Each channel thus describes the spectral response pixel-wise within a small portion of the entire spectrum. When storing this information, the satellite sensors are performing a discretization of the continuous signal received from earths surface into the brightness levels. The number of brightness levels (which is the radiometric resolution) depends on the sensor system. While older sensors like LANDSAT-TM5 internally deliver only 7 bit data (2=128 grey levels per channel), modem systems capture data at 12 bit or higher (2=4096 grey levels per channel). For data distribution the 8 bit, 16 bit, and 32 bit formats are used. Note that level numbering starts with 0 (no signal, usually colored black). Accordingly, an 8 bit image contains 256 levels numbered from 0 (black) to 255 (white) with different grey levels in between.

To view the images, open a GRASS monitor and run slide.show.sh. Some images are not geocoded, so always use g.region with parameter rast and the image name to display an image within this LOCATION. Note that this LOCATION is defined with negative y coordinates due to historical reasons.

9.2.2 Export of multi-channel data sets

The export of one or multiple channels from GRASS is currently only possible into ERDAS/LAN format using the module i.out.erdas. Alternatively you can export single channels into various raster formats, please refer to Chapter 5 for details.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 ( 64 ) 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127