The raster graphics approach in mapping

Comout. & Grapll~cs Vol. 9, No. 4, PlY.373-381, 1985 Printed in Great Britain. 0097-8493/85 $3.00 + .00 © 1986 Perllamon Press Ltd. Arctic Views o n C o m p u t e r G r a p h i c s THE RASTER GRAPHICS APPROACH IN MAPPING MIKAEL JERN U N I R A S A/S, Nerregade 7, DK-2800 Lyngby, Denmark Abstract--Graphics hardware technology is expanding from pen plotters and vector screens to raster. The decade for outstanding advancements in raster output devices, in terms of greater resolution, more colors and greatly reduced costs, is upon us. Ever since the first graphic output devices were developed, discussion as to whether the raster or vector technique is better has occupied computer graphics analysts. There are problems which are easier to solve in one system, but almost impossibleto attack in the other. For example, high-precisionengineeringdrafting consists mainly of short line segments,an operation very cumbersome for a raster structure, but simple with the vector technique. On the other hand, solid shaded areas, colored lines of varying thickness and satellite image processingare impossible with vector systems, but natural for a raster display. Recent years developments have concentrated on devices which do not draw straight lines, but instead use a matrix of dots to build up the picture. This method is the raster technique. 1. VECTOR GRAPHICS VS. RASTER GRAPHICS Printer graphics Computer generated pictures may be divided into two classes; vector graphics (line drawings) raster graphics (continuous-tone images) Not only are these two classes of techniques very different in appearance, but they require different techniques for their generation. The vector graphics technique is unique in its ability to draw from one arbitrary (X, Y) location on the display to another (X, Y) location. Early computer graphics sought to imitate the actions of a person drawing with a pen (Vector technique). Mechanical devices were constructed to move a pen across paper in straight lines: digital plotters. The line drawings are in most respects easier to create because the algorithms for their generation are simpler, the amount of information required to represent them is less, and they can be displayed on equipment which has until recently been more readily available. Example of a complex line drawing is shown in Fig. 1. Raster technique A computer generated raster image is a picture which is based on a rectangular array of digital information. Each element of this array is known as a picture element or pixel. The most primitive example of a raster image is a black-and-white "dot picture" which, for instance, may be produced on a matrix printer, the digital information defining a pixel is in this case a single bit. The two values of a bit, 1 or 0, will then correspond to the two possible state of the dot, namely on or off, i.e. black or white, respectively. Black-and-white images are fine for some applications, but grossly unsatisfactory for others. The addition of intensities means that the digital information defining a pixel is no longer a single bit, but rather a number which specifies an intensity level or color for the pixel (see Fig. 1) The very first computer maps to be produced were based on the raster technique, namely alphanumeric printouts. Printer graphics are built up from symbols, where each symbol represents a picture element. Grey scale pictures are produced by selective overprinting. Because of the poor resolution, its graphics limitations are obvious. See Fig. 3 below. Why raster graphics The raster technique is attractive for several reasons. The full spectrum of color is easily obtained, while a vector display is limited to the number of pens available for a pen plotter or 3 or 4 colors of the expensive beampenetration graphics terminal (CRT) technology. Complete areas can easily be filled in with a raster technique by finding the pixels which are inside the boundary of an area and turn them "on." In vector graphics, shading must be simulated by letting the "pen" cross-hatch the area. The denser the lines are placed, the darker the area will appear (see Fig. 1). Many of the uses of traditionally line-drawing graphics can be performed on raster graphics devices. The extra capabilities of shaded areas and color, however, can enhance the resultant picture considerably. The major application areas of raster graphics form a continuous spectrum. This article will focus on a few application areas, where the raster graphics is superior to the vector graphics. 2. THE RASTER TECHNIQUE FOR PRESENTING GEODATA The geodata map is part of a broad class of maps, whose purpose is to communicate geographical concepts such as the distribution of densities, relative magnitudes, gradients, spatial relationships and movements. The main task is to represent on a two-dimensional diagram in a third dimension, which, in general, represents some statistical quantity. Choropleth map A widely used method is the shaded-zone or choropleth (to use cartographers language) map. An ex- 373 374 M. JERN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 0 0 0 0 0 0 0 ~ 0 0 0 0 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 2 2 0 0 0 C 0 0 0 0 0 0 0 2 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 } ~ 5 ~ 0 0 0 0 0 oooooo~5~oooo0 o 0 0 0 0 0 2 1 1 5 ~ 3 0 0 0 0 0 0 ~ 0 0 2 ~ 3 ~ 5 0 0 0 0 0 1 1 1 1 1 1 1 ~ 0 ~ 0 0 0 0 0 1 1 1 1 1 1 } 0 0 0 ~ 0 oooo1111550ooo5o o o o o l o o ~ 5 3 o o o o o o o o o o o o o o ~ o o o o o o Fig. 2. Black-and-white dot picture utilizing intensity levels. ample of such a map is shown in Fig. 5, which indicates by country the world population density. A suitable color scale is chosen to represent the statistical values. The map's primary objective is to symbolize the magnitudes as they occur within the boundaries of a general enumeration district--countries in the case of Fig. 5. The automated production of such maps is becoming commonplace, but does not really represent a unique capability of the computer; as such, maps have been manually drafted for years. A choropleth map generally portrays only one set of statistics at once, so it may require a series of maps to summarize all the data available. This is precisely where computer mapping outperforms manual methods; the human effort required to produce each new map is minimal if the procedures are automated. Figure 5, incidentally, is a good example of an ink jet hard copy map. It was produced by a mapping system called GEOPAK, distributed by UNIRAS in Copenhagen. Figure 3, shown previously, is another example of a choropleth map produced by printer technique. The resolution is poor, but plots can be produced quickly and inexpensively. A grey scale can be produced by selective overprinting. In Fig. l, a choropleth map is produced by vector technique. Shading is performed by the time consuming cross-hatching technique. 2-D contour m a p The contour map is perhaps the most common example of a geophysical map. This type of map is gen- erally done by measuring a variable at specific locations and then interpolating over the region. Each such sample point is simply a pair of coordinates plus an associated value. Examples of this technique are common in the earth sciences, where core or well samples are obtained for a region at certain test stations, as are weather and oceanographic observations. The usual way of displaying a contoured surface is by means of a plane map with contour lines and numeric annotations. The annotation technique is a very difficult task and must be performed automatically with a high level of computer program intelligence. Conventional vector contouring techniques used today do not meet the requirements of many applications. Modern exploration work, for example, uses large amounts of information of geological, geophysical and geochemical origin. To facilitate the interpretation of a map containing a large amount of information, a colored display is of great advantage. Color enhances the readability of the information and helps to avoid confusion. It is not always easy to distinguish the "highs" from the "lows" in a vector contour map. Hand coloring the maps can help to some extent, but this process is time consuming. With raster graphics, shadings can be plotted between the contour lines, which offers the geoscientist and the explorer a vision of an additional third dimension. Such colored maps can be plotted rapidly on a color CRT and therefore be used in an interactive environment. High resolution hardcopies can easily be produced by ink jet or electrostatical plotters. The colored results are striking in .° ,.P,.w...r,,.rp OOh 000000 ,,*,.,.,,p,,. ,,...p*,,,,,, O00000000U OO00000000000 )O00000000OO0 O00000000OO OOOOOOO , ~00 ,,,, ..,, ,., ,, 00000 O00OO00000 000~000000 ~*o~--~-~*--* ~**~-~÷~ --*~----~ ---~* •e ~--~-~-** OOO0000000 00000000000 0000000000 00000 O0 O00O 00000 U0000 0 ~.e.~.+**+ ~*~*ee**** eeeeeeeeee ++~+ '',,,' ,,,,,,,,, XXXXxXXXX XXXXxXXXX XXXXXXXXX XXXXXXXXX XXXXXXXXXX ..,,,,, ...Pp,,, *÷~*~-*+ ~ * ~ * * ~ * ~ ÷ ,,,,,,,, *~ee~*~*÷ ~ ~ XXXXXXXXXx .t,,,*, .,°, ~ ÷ *~+~ ~ + • X) XXXX ~.~,,*,,., 0000 *~÷*e~*eee XXXX XXXx ~,,,,,,,,, 0000 *eeeeeeeeee ~,ee,*e~ep 0000 eeeeeeeeee 00000 * e~ ~ XXXXX *,,,,,,*~, Fig. 3. Graphics produced by printers. ee X) ) The raster graphics approach in mapping 375 WORLD POPULATION DENSITY with s a l e s *,~'~" £. distribution .. "-,%-,.,~!~!iiii!iii!iiiiiiiiP:' Fig. 1. Line drawing with cross-hatching. SALES per DISTRICT, PRODUCT and YEAR Districts . . . . . . . . . . "') ) t 0 m (~ Product ~ Truoks BIB B u n s n Car-, GgaPHtGSO~TWA~ 1981 1982 luu~ ,~u- Year Fig. 4. Color example of an image generated by the raster technique using the Tektronix 4691 ink jet plotter. 376 M. JERH * Sales distribution * n,w l a N l a m l , n m n m m. n 2,~'7 ~e,z,i CARS ss.m mow UNJlni 3 TRUCKS pICKUPS Fi& 5, World population density map produced with ink jet technique. UNIRAS COLOR CONTOURING i -a-~ - - a . i DUB -33 - -~7 m ~2 n mm 4 ~ - -,LS - ~ 1 - -~La IBIm ~ m ~ - - U n -LS - - L a mm amLow SYSTEM - -Lt g -4ks Fig. 7. Example of contour map produced by an ink jet plotter. UNkma ~ ~ r ~ The raster graphics approach in mapping 3-D SHADED CONTOUR MAP PRODUCED BY UNIRAS ,tlllO¥| 3.0tl -3,m 2 3 1 - :Lll4 1:2:: mE am mm mm am mR zu - z.4T z~ z.uz ztl I.. i'm L7I 2~S 2.41 2.J2 2,tl :1.01 1.86 1.70 am 1.3o I..54 tu 1.=1J 1.24 1,08 I~ o.rr - o,IrJ - 0.77 Iml e.4e • 0,02 o.31 - O,46 ill o.m i 82t.ow o..11 o.1.5 Fig. 9. A 3-D shaded raster image. Fig. 10. Combined 2-D and 3-D map. 377 80.0 IJOYlC 40,0 30.0 18.0 13.0- 6O,O 40.0 O0,0 I0,0 %0 10,O 6.0 7.0 UN| +.o " C 4 ~ I : I - I ~ , ~ T W ~ m o , o3.0w 3,0 ....+ Fig. I I. This 4-D map shows the deposit of uranium combined with the topographic data. VEGETATION Fig. 12. A vegetation map showing the integration between raster and vector technique. UNJNHi! GI~ G~TWA~ The raster graphics approach in mapping 384.5 i il)lJ ',l~t , ~ i 384.3 !J/} /,TJ~.."P I 384.1 / - - ~ ,,,;,D~ I L;., 382.9- ~ ~,~ t ..... Jl'J, ~ ]-,, ; -" -.- ~- f _-2 iJ ~--.._-"-i~-"--r--J-4--. .... ~ I~ 1 I ~J I i, ?I ' , i " ./, +"----" ' I I IL_~I ", I' s , ")". \ . - ~ ')t #) ]/,,<m~ : p, :. .~. . . _, L " . , - - f ~ ",r . ~ ,/ ! '~-----f; ...... "~------~----,__~_~s----~ ~ , ~ ' ~ ' ,,,' 'x /.r.'/l/~I/ ' , , ,,[ I/Y/~ ,<.~"-~. 1 ( ~,,~:, ~-i---/') )~ ~ ' /1 / !]J));l,"lJ'k ,"),, ...j ,.-f~-m~- [.-~.,---77,7U,.__, o, " !I ,ll / ,i ' . ~ . - < ~ I s- - t - i - -' " , , " l) i , <,o ~t~-~:"'z~--~L,~ ~ ~--~, r, ) , LI'. I, ,I ~ ~ . , < i ! , ?~ .~ 381.7-f ~ J ~ -,~<=-~'__S _~-/-~.>. 381.5 ,,i]/ / '/ i," / , ~?"~";i~'~. ;;,;73"~¢"-r---/-T~i ] x , ~ , , ~ . 382.3-_~,-]./" 38~.i 7',. ~ 11 / : //~,/<- ' .~, -i "o,.j,' llillitll, -- ,-.,, 382.5. " "",C,~; \]\ /, ,'/X~W'5-<-,,'.V. _-~ ~ I ~~"~-~-°',J 03 383.3 \i \k ' "-,I'~ \ , 379 kI ~ i ' "l ' ! L.J : i ' X k ! t " / f [ -~ ~-- 7 ~-~ ~ " ?,I ~ t l ~ '/ i! ,"1 73.5 73.7 73.9 74.1 74.3 74.5 74.7 74.9 75. I 75.3 75.5 75.7 75.9 76.1 76.3 76.5 .... :, X-AXIS Fig. 6. Example of a contour map produced by "vector" technique. appearance and are easier to interpret than vector plotted maps, because shading and raster patterns eliminate confusion in viewing elevations and contours. The best result is possibly achieved when combining the shaded contour map with numeric annotated contour lines. Figure 7 shows an example of a shaded contour map with overlayed annotated contour lines. 3-D contour map A more interesting portrayal of a surface can be ohmined by mapping it in three dimensions rather than as a contour map. This is an area that has been pioneered by developments in computer mapping. The portrayal is usually accomplished by constructing a series of parallel profiles which are sufficiently closely spaced to appear as a surface. Vector based surface perspective programs use this form of representation. It is, however, almost impossible to read precise heights from mesh perspectives vector oriented pictures, which only gives you an overall view of the terrain. Figure 8 below is an example of a well-designed, vector-based 3-D surface with hidden lines removed. The raster technique also allows shading to be extended to these 3-D perspectives which require no concentration at all to interpret. By adding color and with the hidden surfaces removed, the visual impact and the amount of information transferred by a 3-D display are considerable. Combined 2-D and 3-D maps The contour plots has the great virtue of using only two dimensions to provide a 3-D presentation. Unfortunately, the contour map does not always give the viewer a good qualitative picture of the data, particu- Fig. 8. A 3-D surface produced by the vector technique. 380 M. JERN larly in complex situations. Often in data interpretation, it is the "feel" that is the important factor for the analysis. From the contour plot, it is difficult to visualize the sizes o f the peaks and lows, and get the true nature of the surface. A colored mesh drawn in 3-D perspective is among the representations providing a better sense of the nature of data. Why not get the best of both worlds by combining both displays on a single plot as shown in Fig. 10. The viewer can j u m p between these presentations, visually and mentally, to gain a deeper understanding of the data. Another advantage of this combined plot is that one type of presentation often suggests features not as evident in the other, and vice versa. For example, the heights of the peaks do not show up clearly on the 2-D contour map, but on the other hand the peaks may hide some important information behind them. Introducing the fourth dimension By using color to illustrate another variable, a "fourth" dimension can be added to these 3-D mapping applications. X- and Y-direction, height and color result in a 4-D map. A simple example of such an application would be the construction of a topographic surface with the addition of surface geology in color. Similarly, soil surveys, land-use maps, hydrological variables and so on can be viewed in conjunction with the topographic relief of the area. The use of color and a three-dimensional view together is demonstrated in Fig. 11. Two data sets are combined in one picture. The height of the 3-D projection shows the topographic relief and the color shading indicates the concentration of uranium in the area. 3. INTEGRATING RASTER AND VECTOR SYSTEMS For over a decade, the highly complementary technologies of vector-based computer graphics and raster-based image processing have been developing in isolation of each other. An interesting application where both raster and vector capabilities are integrated is a digital mapping system. An important application for vector-based systems has been the production and maintenance of cartographic maps and databases. A complementary application for raster-based systems has been the display and analysis of raster imagery from remote sensing satellites such as LANDSAT. Those two technologies are being brought together through the requirements of the users. Cartographers would like to use the wide variety of digital raster data as a cost-effective source of data for map creation. Remote sensing users have an increasing need to produce cartographic quality maps and to maintain resource related data in a geographically related database. Those complementing requirements emphasize the need for a digital mapping system providing both raster and vector capabilities. Figure 12 includes the rasterization of a vector map and the overlaying of this map on an image of the same area. Land use data received from LANDSAT is plotted together with digitized lakes, creeks and coastlines. Another very important feature of the raster technique is demonstrated here. In many applications the same background picture is used for visualizing a variety of data. The vegetation land use map exemplifies the ability to display data on previously stored backgrounds. The basic map of the land remains the same from plot to plot. Instead of redrawing the basic map for each new variable, it is far more efficient to generate the basic map only once and simply overlay (logically merge) the variable input data to produce the different plots. It seems likely that integrated vector/raster image processing systems will become more common in the future. Other applications will develop in addition to mapping and resource analyses. 4. HARD COPY DEVICES When the first color graphic terminals were introduced, the only way to get a hard copy o f a CRT image was to take a snapshot. Until recently the resolution from inexpensive hard copiers has been too poor to attract users of computer graphics. However, as image quality is improving, users become more and more aware of the advantages of hard copiers. Several techniques exist for creating a hard copy of computer generated graphics: • • • • • • Pen Plotters Ink Jet Printers Impact Printers Photographic color video copiers Xerographic copiers Electrostatical plotters are the most common hard copy devices available today. Each of these has its capabilities, advantages and disadvantages. The need for color output is expressed by users from all application areas, such as CAD/CAM, business graphics, mapping and medical applications. To many users hard copy devices are more important than color terminal displays. Most users acknowledge the terminaps role for previewing graphics, but what they really want are copies for presentations and the communication of ideas, as well as for filling and archiving. The desire to bind a paper copy as part of a document, mail it or save it for future use in a file has created a demand for the small format hard copy devices. The need for color as a tool plays an important role in graphics. More information can be shown all at once with color than with only black-and-white. Another request is for copies with better quality than the graphics appearing on a screen and to have colors which match those specified for the screen. The first color raster hard copy In 1971, Professor Hertz and Jern (author) developed one of the first color raster-based graphics systems in the world at the University of Lund, Sweden. The system, based on an ink jet color plotter and a graphics software package, received considerable attention. The ink jet technique involves spraying a continuous stream The rastergraphicsapproach in mapping of electrically charged ink through a high-voltage field at a sheet of paper. Switching the electrical field on and off determines whether the ink jet droplets reach the paper, or are repelled electrically and sucked up. At a rate of 1 million drops per second and 250 dots/ inch resolution the image quality approaches near photographic detail. The author provided the software expertise in this project by writing a graphics software package, completely based on the raster technique, which was used to control the plotter. By the end of 1972 a prototype had been developed and was in operation with the software. In 1974, more than 20 software systems were installed in Scandinavia that produced raster data for the ink jet plotter. This was the same license. Applicon Inc., U.S., acquired the rights to market the Hertz/ Jern technology all over the word. Until 1980 more than 250 systems were installed throughout the world. Ink jet--a hot topic The ink jet plotters that can make high quality copiers of color graphics have become a hot topic in the last year. After many years of bad reputation the reliability and resolution are now improving with every new product being presented. The ink jet technique offers a clear advantage over most available hard copiers. A broad spectrum of ink jet plotters are being manufactured today: Tektronix 4691, 4692 and 4695, ACT2, Benson Colorscan are only a few examples of these high resolution color plotters. 6. HOW TO PRODUCE A HARD COPY In many cases the approach to generate hard copies directly off a video screen causes problems. A color terminal is a "low resolution" device with a limited number of picture elements, while many of the latest hard copiers are "high resolution" devices. For example, 35 m m slides were made for a conference presentation by a film recorder attached to an IBM 3279 terminal. On the screen, the pictures looked like they would make great slides. But when the slides were projected the promise turned sour. Comments overheard after the presentation indicated that people felt that anyone caring so little about the quality of their presentation materials could not have anything very important to say. The importance of picture ftles The chart seen on the terminal must be converted into a high-quality presentation by directing the image directly to the appropriate hard copy device. This leads to the need for some storage mechanism whereby pictures can be saved. The user designs the picture on a CRT and when the picture is satisfactory, he wants to make a hard copy on, for example, a color ink jet plotter. This is achieved by creating a picture file (metafde) at the same 381 time we are drawing on the CRT. The picture file can later be copied to the ink jet plotter. No extra computer resources are required for double drawing. This technique also allows you to use a number of picture files together so several different graphs could be combined on a single chart. Host rasterization, an example of a device intelligent function Let's take the Tektronix 4695 ink jet plotter as an example of what we mean by the phrase "truly device intelligent." The 4695 is a "dumb" device. Only data in binary raster format is accepted as input. The resolution is 980 × 1600 picture elements ( 120 dpi). Four colors cyan, magenta, yellow and black can be used in two intensities "on" or "off" for a total of eight color combinations. The Tektronics 4105 color screen with a resolution of 480 X 360 is used to design the picture. If the image is to be taken directly from the 4105 color screen and copied to the 4695 ink jet printer, each screen pixel will have to be represented by 3 × 3 printer pixels in order to retain scale. This, however, results in the "jagged edge" phenomenon. Another problem is to copy color intensity from the screen to a device which only can be used in two intensities "on" or "off." The software system UNIRAS offers a solution, where the plotted result is not limited to the resolution of the terminal, but to the resolution of the hard copy device. Thousands of colors can be plotted by varying the density of filled pixels within an area. Why is device independence and device intelligence important? Hard copy devices are sufficiently different so that a graph that looks good on one device may not look good on another. A device independent and device intelligent system is one that works with and adapts intelligently to all graphics output devices, vector and raster. Note the stringency here: not "many," but "all" graphics devices. If, for example, a graph specification asks for a line thickness, the plot should have the requested line thickness. If the device can do it, then fine; if not, the software will have to emulate this function. Software suppliers claim device independence and device intelligence UNIRAS offers the only truly device independent and device intelligent software system. Both vector and raster based output devices are fully supported. Fill area, color shading, line thickness, hidden surface removal, segments, 3-D manipulations, etc. are examples of higber level graphics which are emulated by soRware. N o device function will be referred to without fullsoftware emulation on devices where it is not supported. Anything less does not deserve the name "device intelligence."