In this exercise we'll examine some of the possibilities of using SketchUp (any CAD program could do something similar, if with some more trouble) to create a basic 3D model from readily available information--in this case a site plan and an accompanying image for reference.
- Download the two image files that we'll use as basic reference material: tikal_partialsiteplan.jpg and tikal_templeII.jpg.
- With a blank SketchUp window open, choose File-->Import, and navigate to the reference files; import tikal_partialsiteplan.jpg. The second reference image (tikal_templeII.jpg) you'll find useful to have open simultaneously in an image viewer application in a different window, for reference.
- When you import siteplan.jpg, SketchUp will want to align it with two of the axes. Treat the y axis (green) as north-south and the x axis (red) as east-west, and make sure the image is rotated so that north is 'up' if you are looking straight down (in plan view) at the image (View-->Standard Views-->Top will give you this perspective, or use the 'Orbit' tool to approximate it yourself). Note that if the image were not drawn with north parallel to the lateral edges of the image we would have to rotate the image (e.g. in ImageJ) and then import it into SketchUp, which aggressively snaps objects into the cardinal directions of its coordinate system (you might be able to get around this by repositioning the axes after importing the image).
- We'll now digitize the site plan just as we would in Didger or ArcGIS, turning some of the information present in raster form into vectors in the coordinate system native to the software in which we're digitizing.
This introduces an added complication, however: as Eiteljorg points out about CAD in general, we are modeling, not drawing. In this case, what that means is that we are drawing lines in three dimensions rather than two. In other words, as we look down at our imported .jpg in plan view (from the top), we're really looking through an infinite number of planes stacked above and below the plane of that image. If we want to draw a line, then--even presuming the simplest case, which is that we want a line parallel to the plane of our image--how do we indicate which of these planes we want our line to exist in?
- The easy answer is by snapping, and you'll notice that SketchUp will give you several options once you've selected the drawing tool and have your cursor hovering around the image. "On Face in Image" is one, which selects the plane in which we've placed our .jpg. As you begin drawing you'll see several others: SketchUp likes to work in rectilinear fashion (because it assumes that you do), and will alert you when your cursor is positioned to draw perpendicular to the line segment you've just drawn, parallel to one of the axes, etc. (this 'inference engine', in SketchUp-speak, makes life much easier but can be frustrating if it's inferring differently than you're thinking). If you want to draw less rigidly rectilinear features, use the 'Freehand' tool rather than the 'Polygon' or 'Line' tools.
- Digitize Temple II, joining your lines into polygons. At this point, you're effectively digitizing in two dimensions, just as you might in Didger--with the significant exception that SketchUp operates only in an internal Cartesian coordinate system.
- However, the polygons you digitize create faces that can be extruded. Use the tikal_templeII.jpg (as a guide for relative heights, etc.) and see what kind of model you can create, using in particular the 'Push/Pull' tool. (This, incidentally, is where SketchUp really differentiates itself from true CAD software, which would never put up with the sort of imprecise drawing you're doing when you manually extrude surfaces this way). Where do you encounter problems, either of difficulty in digitizing something as you'd like or of having to make decisions based on too-little information? For what uses might the model that results be appropriate?
It's also possible to import precise point locations into SketchUp (as .shp or .dxf files, for instance), and use those as a basis for drawing. You might do this with total station data, for instance. In that case it becomes vital to have adequate metadata that you know which points to connect in which order, as this can dramatically alter the solids in question.
- Import SKETCHUP_EXERCISE.DXF into SketchUp, and open sketchup_exercise.txt (which includes the metadata for the points, such as it is) in Notepad. Note that SketchUp wants nothing to do with a .txt file, and while it will import a .dxf, the points are not accompanied by any accompanying information (CAD programs generally look down upon point files, preferring lines, polygons, and solids).
- You can break the .dxf file into its component parts by right-clicking and choosing 'Explode'. It's then very easy to 'connect the dots'--but which dots? If you zoom in far enough, you'll notice some vertical distinction in the points as well. See if you can figure out how to connect the dots to produce a coherent image.
- What are some of the difficulties this suggests about working with SketchUp, and how might they be circumvented? At the same time, what possibilities does this suggest?
For an interesting example of what's possible with SketchUp in an archaeological context, see http://www.gisarch.com/category/sketchup/.
A similar hybrid (not created with SketchUp, though it could have been):