 |
|
|
|
| Data capture |
|
|
Data capture is a method used for the collection of confirmed information. In order to understand the properties of confirmed data and their areas of application, one ought to know about methods of data capture.
|
By the introduction of GIS the collection of data will usually be an expensive process. Rational data capture poses great demands on the user and the software. This is due to the fact that the original material is often unstructured with many errors. Furthermore, the graphical basis of the map often originates from different sources with different claims of accuracy.
Digital map data may, for instance, be obtained from different sources and in different technical ways. Traditional analogue maps (folio, paper) may be used after having been through a digitalising process via scanning or desk digitalising. Photogrammetric construction may also give digital basic data. In addition, pre-digitalised data may be retrieved from a number of different map databases.
Anybody who has made use of GIS has experienced that there are many problems connected to the collection of data. An integration of all types of confirmed information demands very good functions for data capture and quality control. |
| | | Factors being considered at the collection of data |
Before a data capture is put into effect there are many factors that have to be considered. Of these, the following may be mentioned: need, costs, availability and time perspective.
A balance has to be made between quality desires and costs has to be made. Low quality data may easily be quickly obtained. Providing data of high quality may well be a long-term process (field work, airial photography, adaptation) as this type of data is to a small extent available in the market today. Data quality will be discussed in a separate article. |
| | | Current data sources |
GIS may use all data that may be localised. The following list gives a brief survey of the most important data sources:
• Direct from surveying (coordinate fixed points)
• Manual digitalising of maps (vector based digitalising form existing maps)
• Scanning of maps, drawings, figures, photos, aerial images etc.
• Digital raster data from satellites or aeroplanes
• GPS and other localising- or navigational systems
• Table data from existing databases/archives
• Word processing information
• Video
The next paragraphs will give a short explanation as to how the digital data in the above survey may be established for the use of GIS. |
| | | Land surveying |
Land surveying is a science with traditions stretching back to The Old Ages. A development of the subject of land surveying has taken place over thousands of years parallel to the development of other sciences such as physics, mathematics, mathematical statistics and astronomy. In recent time the development within data processing (information technology) has created challenges and possibilities for land surveyors.
The land surveyor works with advanced instruments. They are instruments of a high degree of accuracy and the most advanced technology. Telescopes are used in measuring angles and differences in altitude. After the Second World War range finding has become common. Telescopes in combination with electronic distance meters and computers are called a total station. An automated total station, which focuses by itself on a prism and carries out the measurements, is called a robot station.
Surveyors produce figure coordinates. Land surveying is the collective term for measuring area and altitude ratio for reproduction in a map version. Traditional surveying is carried out by means of a theodolite. This is an instrument used to measure or mark off directions or angles in a horizontal or vertical plane.
Land survey gives data of a high degree of accuracy.
Land survey is usually used for objects demanding frequent updating, such as property boundaries, buildings, cisterns etc., for details demanding great accuracy or for the measuring of small areas. |
| | | Manual digitalizing of existing map material |
|
| 1. Point by point, 2. Automatic with intervals of time and 3. Automatic with intervals of distance. Click for a larger version. |
|
A map is no longer only a paper. Digital maps are becoming more and more widespread. When we have a digital map, we may present it on a screen and attach properties to the map objects and make the map interactive. Then one may, for instance, click on a town on the map in order to bring out a table showing population figures and other information.
Manual digitalization of maps is suitable for, among other things, maps and documents that cannot be scanned, originals with a large quantity of information (maps where the scanner may have problems with differentiating one object from another) and for special and unusual formats.
Manual digitalization of maps is usually carried out by using a digitalization desk and a cursor or marker. The markers usually consist of a magnifying glass and a reticule cross, and also function buttons so that some commands may be given directly without using the keyboard on the computer. The size and accuracy of the digitalizing desk have to be evaluated against the demands of accuracy of the digitalization
Before digitalization starts, the transformation has to be decided between desk- and map coordinates. This is usually done by choosing 4-6 points on the map with known coordinates. It may, for instance, be grid intersections, areal boundaries and the like. The map coordinates are keyed in together with the digitalization of the fixed data. On this basis the programmer may decide the transformation and make out an average error. If this average error is too great, the process is repeated. For two right-angled coordinate systems two points are necessary in order to decide the transformation.
The registration of points itself may be carried out via different methods:
Single-points registration: Points are registered by pressing on one of the function buttons. This type of registration is best suited for single points and for lines that are combinations of many smaller right lines, for instance, property boundaries or buildings. A disadvantage of this method may be an unnecessary large quantity of data if the parameters of registration density are not correctly set.
Sequence registration: Points are registered continuously at fixed intervals, with a fixed distance or from the bending of the curve.
The figure above shows different methods of point registration.
During digitalization it is common to give theme codes and different other commands such as the closing of polygons and the creation of rectangular symbols. The digitalized curve should be written out direct on a graphic screen so that errors that have been discovered during digitalization may be corrected. Such a display on a screen will also give an overview of which parts of the map are digitalized.
In connection with digitalization of vector data, it is often desirable to reduce the quantity of points, either in connection with the digitalization or as a post-processing. Most of the filtering methods try to find a selection of the original set of points representing as much as possible of the original curve, while removing points representing unnecessary data in relation to the main features of the curve.
Among the advantages of digitalization the following could be mentioned: simple technology, cheap equipment and human interpretation. Among the disadvantages one ought to mention are that the methods are labor-intensive, may involve the loss of accuracy and the possibility of gross errors. |
| | | Scanning |
Scanning is an automatic process. The points, lines and areas, i.e. the elements of the paper map, are transferred to digital raster data by means of the laser sensor of the scanner. While manual digitalization gives vector data, digitalization in a scanner gives raster data. Raster data gives us maps consisting of a finely meshed grid. The squares are called pixels (= picture elements). The size of the pixels, what we call the resolution, may vary. If the resolution becomes too low in relation to the details of the map – i.e. too large pixels – much information will get lost.
When the laser head scans the original map, the laser registers grey shades of tones. The laser may be adjusted to only register certain values of grey shades of tone. This is something we may take advantage of when it is a question of digitalizing (scanning) one theme at a time. The scanner quickly produces a large quantity of encoded raster data. Raster data to be used in analyses have to be coded. It is also possible to convert raster data to vector form. At present there are computer programs for automatic vectorization of scanned material.
If a folio separated (the different themes are divided among several folios) analogue map already exists, scanning is much faster than manual digitalization. The map is read automatically into a computer where a raster image of the map will be created. If the computer program has, in addition, automatic functions for the correction of errors, recognition of themes etc., maps may be digitalized and corrected for errors up to seven times faster than by manual digitalization. More intelligent software of pattern recognition of raster maps may probably be able to reduce this time additionally. Scanning is also actual for the digitalization of analogue images, for instance, aerial images.
Rapidity, together with high and stable accuracy, are obvious advantages when choosing scanning as a method. Among the disadvantages one could mention low “intelligence” in the data. The method demands a great deal of follow-up work in order to improve the data and form structures.
Scanning is best suited when:
• the map mainly consists of lines (for instance, contour lines, water outlines)
• the map has many similar symbols and few graphical conflicts
• the map is of high quality and lines and symbols are distinct
The fact that this method is still used rather than automatic scanning is due to the advantages of manual digitalization. The method may be used with small investments in equipment, it is easy to learn and together with modern methods of data control and checking of errors the quality becomes satisfactory. Besides, scanners satisfying the demands of accuracy for maps are very expensive. Normally, such equipment only exists in larger government offices or enterprises. |
| | | Digital raster data from satellites or aeroplanes |
The information arriving from earth observation satellites is described in “Satellites supervise the Earth”, see Earth observation. Aeroplanes are also equipped with scanners quickly scanning wide areas in a short time. Optical data are observed within different areas of the electronic magnetic spectrum. These data may be handled and interpreted in the same way as data from satellites. The resolution in aerial images is very high, usually ca. 0.025 mm. Aerial images therefore contain a great deal of information.
The scanner installed in aeroplanes works as a telescope covering stripes across the aerial direction.
The figure shows a schematic diagram of a scanner installed in an aeroplane (1 shows the aerial direction).
Aerial images and maps are formed by different projections. They have therefore the same geometry only if the actual terrain is flat and horizontal, and the direction of photography is exactly horizontal. This will almost never occur in practice. Registrations made in aerial images must therefore be transferred to a map-correct form in order to estimate lengths and areas and in order to be able to be used together with other map data in GIS. |
|
|
|
|
| |
|
|
| |
|
|
|
|
|
|
