ďobo-CADzone, true men don't kill coyotes and use AutoCAD…

V tejto časti seriálu CAD a archeológia sa pokúsim natieniť základné princípy kreslenia a modelovania v dokumentačnej praxi.

In this part of CAD in archaeology I’ll try to sketch a basic principles of CAD modeling in documentation praxis.

1. Those principles are application independent.
2. They’re result of my humble opinion not a some codified standard.
3. Big motivation and good resources were also :

• articles from Harrisona Eiteljorg II on CSA server, and his CAD manual CSA CAD Guide for Archaeology and Architectural History
• ebook Archaeological Computing (Eiteljorg, Limp), for free download
• items from English Heritage Guidance (Presentation of Historic Building Survey in CAD 1, 2)
• server Archaeology Data Service, mainly its projekt CAD: A Guide to Good Practice.

Model not the drawing

• While the drawing is only graphical substitution of documented reality, CAD model is its realistic representation. The drawing shows the reality through graphical symbols but CAD enables to depict the points, lines, surfaces and mass like real objects.
• Drawing is (in most cases) only one presentation of the possible situations – typically is the terrain documentation presented by the set of thematic drawings. CAD is one closed corpus of all graphic data.
• CAD model offers all real metric information (length, area, volume) natively, no additional calculation (through scale) is needed.
• CAD model is structured (see below) and could be queried by its atributes
• CAD model is compatible and interchangeable with allied fields of specialisation (geodesy, cadastre planning, architecture, GIS, BIM)

CAD model in two simultaneous views with distinct projection (fig.01) and selection only some data for print version in grayscale colorspace (fig. 02).

The difference is obvious in comparison with wide-spread use of Corel Draw during documentation processing (in Czech and Slovak rep.). For Corel based documentation the visual evaluation of drawing is typical, no additional analyse is potential. Nevertheless the ideal manner how to acquire the full potential of data is a CAD model (seen and evaluated from different points of view by archaeologist) and only then filtered to specific visualisation for printing.

2D, 2.5D and 3D

Regarding native 3D CAD engine and the fact that raw terrain data from total stations are usually georeferenced, standard documentation attitude is 3D modeling. Reduction to 2D is in any case trivial⁴ but not contrary. Saying 3D doesn’t mean layouts with additional info (nivelation markers with height etc.). There is a obvious strong tendency to model in "false 3D" or 2.5D respectively as a consequence of original 2D planes composed (3D rotation and aligning) into 3D. This procedure gives typical little incosistencies on the fringes of slightly irregular shapes like a walls and others⁵.

Obr.03-04. Inconsistencies on wall edge – originaly 2D models composed together to "2.5D".

Visual styles on modeled entities

Modeled entity would be on basic wireframe composition (lines, points), two dimensional surfaces (face, mesh) and a solid (3D solid). It holds generally that if you comply the conditions of minimal acceptable entities (see below) there is posibility to transform among compositions in both directions – wireframe to surface and to solids and vice-versa. This modeling case study is typical for that statement.

It is very important to define the projection type for the whole model and keep it stedily during modeling. Wireframe is very quick for building and has lowest memory requirements but can not be rendered and doesn`t enable some kind of analysis (e.g. collision analyse). Modeling on solid level again is very rare, for "nice visualisation" the surface modeling seems to be sufficient.

Archaeoentities and rules of their modeling

While the major documented reality should be modeled by procedures of geodesy or architecture (excavation polygons, DMR surfaces etc.) some pure archaeological entities (stratigraphic units, SU onward) have no drafting rules defined so far. Basically they are divided into three types – deposite, negative and construction⁶. Identification of SU = recognition of it’s surface hence documentation of SU = it’s surface/surfaces documentation. The minimal quota for their documentation in 3D on condition their future surface and solid conversion is following⁷:

• Deposit – its upper surface outline, by closed polyline or spline
• Negative – its upper outline (closed polyline) and one profile line at least
• Construction - surfaces of its interfaces by contour lines

The definition of minimal documentation quota for SU – red lines. To fulfill (at least schematic) subsequent 3D modeling you need the upper outline of deposite (fig.08), upper outline and at least one profile for negative (fig. 09) and all edges of vertical construction entity (fig. 10).

The surface definition is a matter of terrain observation in any case. An usual ailment is disability to atomize the research situation into simple units (events) and subsequent fault definition of their original/authentic surfaces. This is typical before all for negatives that break the deposition. Figures 11-14 serve like an few examples.

	Stone block paving on groundplan (fig. 11) is defined by one type of outline – but with exception of northern side is its original shape "bitten" by later negatives that are not defined.. Fig 13 shows the righ definition of pavement – "thick" line defining authentic outline only on one side.

Fig. 13-14. Identical situation in profile view – there is no negative defined that breaks the construction on the sketch.

For excavation situation documented by the cuts exclusively the rules are quite identical:

• deposit is identified by the line of its upper surface – in the case of topological analyse the deposit is "all below its surface till another surface". Drawn by opened polyline/spline.
• negative is traced according to its real running (opened polyline/spline)
• construction is a region delimited by its borders/surfaces (closed polyline)

An example of SU drafting on cuts documentation. Negatives are drawn with red lines, construction with blue one. Deposites have its own color each other for better recognition. Figs 15 and 16 only introduce the situation, fig. 17 defines the rules mentioned above. Fig. 18 is the use of simple topological rule "the volume of deposite is delimited by its surface and everything below it until next surface". Please notice that SU 7000 (loess) / original surface is not preserved (negatives countersunk into it) and therefore not traced.

Deposites drafting according to H. Eiteljorg (bottom surface and side limits) leads to two problem at least – 1. it requires to track the terrain surface additionally and 2.in fact there is no possibility to define the lowermost deposite (very often geological subsoil) because you can not reach its lower surface.

Fig. 19. Borrowed from Eiteljorg – Limp publication , y. 2006, p. 197, figs. 42, 43. This kind of deposite tracing is not acceptable for me.

CAD model structuring

Probably the most important question. It requires to define the hierarchy of CAD model for searching, filtering and viewing purposes (H. Eiteljorg uses a pregnant term "segmentation"). Almost all resources mentioned at the beginning of this article envisage this problem as a pure CAD-layers problem (Layer Naming Convention). This is a sinister legacy of using pure AutoCAD and not a profession branches of Autodesk applications – those enabling asociation with object data, external databases, and creating "GIS features".

For example on excavation at Ústí nad Labem – FORUM (Czech rep.) was CAD model structured subsequently – the basic superior Map Project (Autodesk Map 2008) that manages underlying source drawings (divided acording to excavation sectors I-III). Archaeoentities are in all drawings are structured into Layers according to type and there is a database link for every SU (external database where SU number is unique therefore primary key). Visualisation in master Map project run over queries to slave drawings (allowed queries to database, topology, object data etc.) and finally there is possibility to tun off/on layers with objects (see also this description of whole process and integration of Harris matrix into system – Czech language only).

Fig. 20. Screenshot from CAD data structures, Ústí nad Labem – FORUM excavation. Left is the hierarchy of underlying drawings and database links, rigt above the list of recalled Layers and right below the dialog window for creating queries. Autodesk Map 3D 2008.

Conclusion

1. CAD model is consistent set of vector and other data from excavation project, not necesarilly one file in physical sense.
2. Is created in 3D, in real (1:1) scale, is gereferenced in defined projection.
3. Drafted data are structuerd according defined key that enables data querying and filtering.
4. Stratigraphic units are of 3 types and each of them has its own rules of data collection and modeling. Generally speaking you must define the surface of unit.

1. 1 2.5D nie je ten správny termín sensu stricto viď http://en.wikipedia.org/wiki/2.5D , ale pre uchopenie myšlienky asi najvhodnejší [↩]
2. 2 V intenciách Sira E. C. Harrisa. Nemá význam na tomto mieste rozoberať jemné odtienky “stykových plôch”, ktoré sú často považované za samostatnú jednotku stratigrafie [↩]
3. 3 je samozrejme možné ísť nadštandard a zvýšiť množstvo dat zbieraných v teréne – napr. za účelom realistickejšieho modelovania povrchov depozitov, vystihnutia zlomových hrán a negatívov a pod. Je irelevantné, akým spôsobom sú tieto data získané – kreslením na milimetrový papier, trackingom za pomoci TS, fotogrammetricky… [↩]
4. 4 e.g.in AutoCAD by one command "Flatten" [↩]
5. 5 2.5D is not appropriate term sensu stricto – see http://en.wikipedia.org/wiki/2.5D but for general understanding. [↩]
6. 6 in rough accordance with Sir E. Harris [↩]
7. 7 of course there is possibility to go overstandard and increase the terrain data gain – e.g. for more realistic look of irregular surfaces, to precisely describe the edges of negative etc. Its completely irrelevant how those data were collected – by drawing on paper in scale, by total station tracking or by photogrammetry… [↩]