METHODS

To produce 7.5-minute quadrangle sheets for field verification maps, plotting raster data with point, line, text overlays was a straight-forward incorporation of the Arcplot image command in the Arc Macro Language (AML) file that generates the map. For example:

/* ..... map parameters

mapextent <pathname>/<filename> /* map area

mapunits meters /* unit of measure

mapposition ll .38 .73 /* location on the page

mapscale 24000 /* scale

/*..... map data

image <pathname>/<filename>.tif /* imagery

arclines <pathname/<filename> /* lines

pointmarkers <pathname>/<filename> 208 /* points

pointtext <pathname>/<filename> <field> /* text

/*..... end

In addition to plotting well points and identification codes or soil association groups on the DRGs, these text, point and line data were plotted on DOQ and National Aerial Photography Program (NAPP) raster format imagery. After the map extent and scale are set, the image command is issued followed by commands for appropriate vector data. Sequence is important. It is more efficient to write an AML assuming an opaque plotting format because the same file may be produced many ways--on a plotter (opaque or transparent), as a slide (opaque), or on screen (opaque). If text, points, lines, and area fills are plotted first, followed by imagery (or other area fills), the initial data will be overwritten by subsequent data as it plots (note that, in an image, white paper areas are also composed of value coded cells).

An image may be converted to Arc/Info GRID format depending upon the expertise or intentions of the user (e.g., to use analytical tools, to remap colors, or to create a plot) of the user. The standard color map for DRG data is:

	Color	R	G	B
0	Black	0	0	0	text, miscellaneous lines
1	White	255	255	255	background
2	Blue	0	151	164	water bodies fill
3	Red	203	0	23	highways, land survey data
4	Brown	131	66	37	topographic lines
5	Green	201	234	157	wooded areas
6	Purple	137	51	128	overprint for disturbed ground
7	Yellow	255	234	0	non-standard (woodlands, reservations, etc.)
8	Light Blue	167	226	226	water boundaries, lines or text
9	Light Red	255	184	184	municipal areas w/landmarks
10	Light Purple	218	179	214	extension of urban area
11	Light Grey	209	209	209	provisional urban
12	Light Brown	207	164	142	complex surface area

Note that rather than coding the background color (white) as 0, it is 1 and foreground text (black) is zero. This can be inconvenient when plotting a GRID format image. It is a three-step process to interchange the values for black and white cells in the GRID format image use the con command. The input grid (in-grid) is grid1. Step 1 opens an output grid (grid2) in which all in-grid cells with a value equal to 1 are set to 999; all other cells retain their original values. In step 2, grid2 becomes the in-grid. All in-grid cells with a value equal to 0 are set to 1 in the output grid (grid3); all other cells retain their input values. Step 3 opens an output grid (grid4) in which all grid3 cells with a value of 999 are set to 0; all other cells retain their in-grid values.

out-grid       in-grid values

step 1: grid2 = con ( grid1 == 1, 999, grid1)

step 2: grid3 = con ( grid2 == 0, 1, grid2)

step 3: grid4 = con ( grid3 == 1, 999, grid3)

Although a plot that contains solid area fills and imagery must be sent in transparent mode, a transparent effect can be achieved on the monitor for slides or demonstrations by using the Arcplot image command:

image <pathname>/<filename> TRANSPARENT 0

One way to remap colors, in Arc/Info, is to copy the value attribute table (*.vat) to a lookup table (*.lup) using:

copyinfo <gridfile>.vat <gridfile>.lup

additem <gridfile>.lup <gridfile>.lup symbol 3 3 i

Add the field "symbol" to the *.lup and set all values to zero except for the selected feature (i.e., topographic lines (4) in the following example):

gridname.vat				gridname.lup
Rec#	Value	Count		Rec#	Value	Count	Symbol
1	0	807783		1	0	807783	0
2	1	30723108		2	1	30723108	0
3	2	140292		3	2	140292	0
4	3	312763		4	3	312763	0
5	4	379627		5	4	379627	10
6	5	1401261		6	5	1401261	0
7	6	326482		7	6	326482	0
8	8	690254		8	8	690254	0
9	9	779983		9	9	779983	0
10	12	753567		10	12	753567	0

As an alternative, a new colormap file may be created and called when the command to print the grid is issued. For example:

gridpaint<pathname>/<filename> # # # <pathname>/<colormap file>

Sample Colormap File
Original file				Selected-features file
R	G	B		Value		R	G	B
255	255	255		0		255	255	255
0	0	0		1		255	255	255
0	151	164		2		255	255	255
203	0	23		3		255	255	255
131	66	37		4		0	0	0
20	234	157		5		255	255	255
137	51	128		6		255	255	255
255	2234	0		7		255	255	255
167	226	226		8		255	255	255
255	184	184		9		255	255	255
218	179	214		10		255	255	255
209	209	209		11		255	255	255
207	164	142		12		255	255	255

Conversion of lines from raster to vector format may be accomplished either interactively using Arcedit or semi-automatically using Arcscan. It is up to the user and the demands of a specific project to select the most appropriate method. In choosing between ArcScan and Arcedit extraction, the following factors should be considered:

1. the quality of the raster data (there will be varying amounts of hand digitizing even with automatic generation),

2. user experience with the software (Arcscan presents new users with a steep learning curve because its tools are complex and somewhat cryptic),

3. time limitations (Arcscan may be faster than Arcedit).

4. acceptable cartographic quality of both lines and polygons,

5. accuracy (some generalization may occur with Arcscan),

6. the scale at which the data will be used,

Interactive Arcedit extraction requires the usual Arcedit commands plus the image command (i.e., image <filename>.tif) which displays the DRG or DOQ. Initially, it may take a little time for the system to format the data, then the draw command causes the image to reappear each time the edit screen is refreshed/rescaled. The Arcedit session proceeds with the image as a background. The command image < off | on > may be used to turn the image off in order to speed the drawing time during editing, coding, etc. Since USGS DRG (<filename>.tif) files are associated with a world (<filename>.tifw) file, the lines, points, and/or polygons generated are added in "real space" because the image is georeferenced. If using an unreferenced image, ground control points in the image, such as road intersections, can be associated with "known" reference points in the DRG to transform data extracted from scanned imagery.

To start the Arcscan process, the raster image is converted to grid format (in Arc or Arctools). From the Arctools window select edit tools, file and open grid. The image will draw in false-color. To correct this, or to color only the cells from which vectors are to be generated, select display, draw env grid, and then choose an appropriate color format (gray scale, colormap.file, etc.). Select draw and the altered image will appear as a screen display that facilitates vectorization of the selected cells. Now select file coverage new, enter a name for the file, select arcs and then OK because a coverage must exist to hold the generated lines.

Vectorization parameters may be established in order to tailor the generation process to suit the particular job requirements. The tracing environment menu provides access to all properties that determine line and edit environment characteristics, mouse control, graphics, session termination, etc. Once a set of parameters has been entered, they may be stored in a file (*.vtp) for future use. A few basic metric settings (in the UTM projection) and their definitions might be:

Parameter	Unit	Definition
width	6.5	maximum size of connected cells that constitute a raster line
value	1	value ( from the *.vat ) of the line being traced
gap	15	maximum distance between raster line segments
dash	25	acceptable spacing of dashes in the raster line to follow
hole	1.5	ignores gaps in the raster line
fan angle	35	primary search angle for continuing a vector across a gap
second angle	100	secondary search angle for continuing a vector across a gap
variance	100	response to line width variations

To trace a line, select a starting point on a raster cell by positioning the crosshairs in the edit window and press the left mouse button. The arrow that appears does not point along the "line" but simply indicates a general direction that the trace will follow. To change the direction of the arrow press the middle mouse button, press the right mouse button to proceed with tracing.

PROCESSES USED AND
PRELIMINARY RESULTS

Color-averaging during the rasterization of the Vincennes quadrangle map resulted in a considerable loss of brown cells--the topographic lines. In addition, there are a large number of supplemental contours on the floodplain of the Wabash River. As a result, topographic lines from the DRG were vectorized "on-the-fly" in Arcedit using the color image as a background. Vectorizing, editing, and coding of the lines were accomplished in a single step.

Soils boundaries were extracted from scanned Lawrence County, Illinois, and Knox County, Indiana map sheets using a density slice procedure in PCI. These raster data were converted to vector format using Arcscan and registered using GCPs recognizable on both the DRG and scanned soils imagery. After the data are coded, updated by NRCS personnel, and field verified, they will be incorporated in the national soils inventory being compiled by NRCS.

The Vincennes DRG data were converted to grid format and remapped to produce four base grids that contain various combinations of original color cells whose use depends on the requirements of the final map product. The black and white areas were interchanged so that land survey information could be included to screen dump slides. Wooded areas were removed because they interfered with other colored data included on the soils and parent materials maps. The Wabash river boundary was updated using 1994 imagery. Considerable alteration had occurred in channel contours since the quadrangle map was last updated.

Three sets of 1:12,000-scale quarter-quadrangles were produced for initial field reconnaissance. The base data for the map sets were: 1988 NAPP-1 data, 1994 NAPP-2, and classified color Infrared (CIR). These maps cost $6.90 each to plot on a Hewlett Packard 750C ink jet plotter. Images produced at the 1:24,000-scale by a commercial vendor cost approximately $300.00 per print. The detail and quality of the plots and the "bird's-eye view" of the landscape with topographic and land survey data superimposed have been very helpful to geologists working in the field.

The imagery, plus various combinations of wells, soil groups, parent materials, and other base data have been plotted for presentation maps, field maps, slides, etc. At the completion of the Vincennes mapping project some of these maps and coverages will be written to CD for archiving and future data transfer.

REFERENCES

GIS by ESRI, 1994, Arc/Info Version 7 Commands Reference Volumes, Environmental Systems Research Institute, Inc.

Greene, A.V., 1990, Illinois Geographic Information System: An index to automated statewide databases, Illinois State Water Survey Circular 175.