COMPUTER AIDED RECORDING TOOLS AUTOMATE THE CREATION OF A SITE INFORMATION SYSTEM

David Whiting
Senior Conservation Architect
Public Works Government Services Canada
Real Property Services for Parks Canada
Calgary, Alberta, Canada K2P 3H8
e-mail: david_whiting@pch.gc.ca

Steve Nickerson
Nickerson & Associates Designed Building Systems Ltd
Ottawa, Ontario, Canada K1Y 3Z4
e-mail: steve@icomos.org

KEYWORDS
As Found Recording, Information Systems, G.I.S., Computer Aided Design

ABSTRACT

Much of the time of a restoration architect or heritage recorder is spent, on site, analysing the building and collecting photos, data and measurements which will be used later to create reports and drawings. The computer revolution has provided at least partial solutions to these tasks but, so far, has addressed only the individual components. Compiling a complete record of the structure is still a mostly manual procedure and the organization of the material for archiving and future monitoring of the site is seldom completed.

A suite of software tools, currently under development, tries to address these problems - first by helping to structure the on-site survey process so that more consistent data is collected, and then, by automatically creating a 2D or 3D CAD model from the resulting database. We do this on a laptop computer, so that the model can be compared with reality and faulty or missing measurements corrected before the recording team leaves the site. While the software supports virtually any source of geometric information, the point collection system what will be described uses an electronic theodolite with a visible laser EDM which has more than adequate precision for most architectural objects.

More significant than the automatic drafting is the fact that this combination of database and CAD drawing is linked, allowing queries of associated data by clicking on an object inside Autocad or the assembly of a drawing based on a database query of object characteristics. Other data collected on site or later including field notes, reports, sketches, drawings and photos can be accessed directly from the CAD model.

Two Canadian Heritage sites recorded in this way will be presented.


1. INTRODUCTION


The management of our cultural resources requires an understanding of their value, their history and condition and their heritage defining characteristics [1]. Access to all of the information on a cultural resource is a key to its protection and presentation. Typically, it has been difficult to access relevant information. A number of computer aided recording tools are available today that not only assist in the gathering and organizing of information but also enables the query and display of information required for specific purposes in the protection and presentation of built heritage resources. This paper presents one approach to creating an integrated site information system.


Much of the data gathered on a resource comes from various conservation disciplines spending time on and off site carrying out research, analysing buildings and their setting and collecting photos, data and measurements which will be used later to create reports and drawings. This work is carried out for a wide variety of purposes including heritage recording, research and analysis, maintenance and monitoring. These activities are typically carried out at different times, by different people, using different techniques with the records residing in different places, with differing levels of protection and accessibility. A basic concept of collecting, organizing and accessing information gathered over time on cultural resources is required.

The computer has made it easier to describe a cultural resource. The computer systems we have at our disposal today offer many amazing tools to sort, analyze, display, present and communicate our data. However, much less progress has been made at the other end of this process, that of data acquisition and its accurate input into the computer formats that will do the processing and prepare the reports and graphics.

As a result of the preoccupation with the presentation of a resource, a lot of information has found its way into a variety of proprietary CAD. wordproceesing and imaging formats where it is difficult to search and which may not even be accessible to the next generation of software. This data may also be available in field sketches and notebooks (even harder to search) but often its only computerized manifestation is in a CAD model painstakingly developed from a combination of survey and hand measurements and enhanced with a great deal of assumption and generalization into a "drawing" that looks right and presents the hypothesis in the best possible light.

What is needed is a data acquisition strategy that acknowledges that its destination is to be a database of all the available information which can be easily searched. The database not the drawing should be the heritage record.

A system currently under development (CART) addresses the shortcomings in creating and accessing data on a cultural resource. CART is an integrated set of tools and techniques used to collect measurements and data in the field and to integrate them into an overall site information system. It has been recently used as the basis for the recording of the Royal Suite at Rideau Hall in Ottawa and Skoki Lodge in Banff National Park.

2. SITE INFORMATION SYSTEM


Much of the time of a restoration architect/engineer or heritage recorder is spent, on site, analysing a historic building or site and collecting photos, data and measurements which will be used later to create reports and drawings that will become the basis for the protection and presentation of the resource. Recording and documentation activities are often time consuming and expensive. Unfortunately, the collection of information on site is just the beginning of a long process, often delegated to junior staff unfamiliar with the resource, of drafting the as-found situation. Errors are inevitable in this type of work but typically they only come to light (if at all) back in the office where confirming a measurement or note may entail a long trip back to the site. Heritage records prepared under these circumstances are often incomplete for future purposes and are seldom 'updated' to reflect changes in the building or site's status. The traditional heritage record is not an information system for the site.

Heritage recording is usually carried out at a few key stages in the site's research and development phases. It seldom reflects the total scope of knowledge of the resource. Compiling a complete record of the structure is still a mostly manual procedure and the organization of the material for archiving and future monitoring of the site is seldom completed. As a result, the record is often not effectively used by those making decisions about the care and development of the resource and is seldom the basis for the 'everyday' monitoring and ongoing care of the resource.

To be considered as the basis for a site information system, the heritage record must be a comprehensive and living document. It must reflect the full scope of all of the information collected on the resource at many points in time. It must reflect both the as-found condition of the resource as well as its current state. It must be convenient to those carrying out onsite investigations and analysis. It must reflect the interests of a wide variety of disciplines interested in its care. The heritage record must be easy to access and update without costly updates of drawings and reports. The heritage record should not be limited to information that can be conveniently conveyed in the form a drawing or written report.

A site information system should be based on a comprehensive database of all known information, including both graphic and textual. It should be permanent and accessible. It should incorporate a framework for easily adding new information and a program or tools (vector tools) to model or illustrate information based on specific queries.

Traditional tools of the architect/engineer and the heritage recorder which were 'drawing' and 'paper' based are no longer adequate to the task of documenting our built heritage. Developments in technology have benefitted the recording of buildings and sites in the recent past and will continue to do so into the future. For example, photography is now an indespensible tool for the recording and assessment of a building.

The computer systems we have at our disposal today offer many amazing tools to sort, analyze, display and present our data. Rather than communicating ideas on paper, recorders today use word processors, spreadsheets, databases, digital images and Computer Aided Drafting (CAD) to create a wide variety of information on our built heritage. In spite of these developments, the initial record is still being collected on paper in the form of field notes and the measurements manually converted to CAD formats for use by restoration architects. The wide range of information that exists on a building is not being effectively integrated into an accessible information system for the site. While the tools for presenting heritage information is advancing quickly, much less progress has been made on the process of data acquisition and its accurate input into the computer formats that will do the processing and prepare the reports and graphics. A wide range of tools and technology are available to the heritage recorder. Clearly what is needed are concepts for an information system that will respect the requirements of all of the conservation disciplines but still enable the preparation of CAD drawings and models as well as written text and photographic images on which subsequent conservation decisions will be based.

3. THE DATABASE AS THE HERITAGE RECORD

What is needed is a data acquisition strategy that acknowledges that the first destination of information on a heritage resource should be a database of all the available information which can be easily amended, updated and searched. We envisage a site information system that has, as its core, a database of all the information on a cultural resource.

The database would include references to field notes, photographs sketches and reports etc. It could also include scanned images of original drawings and field notes and sketches. More importantly, it would also include a record of the building or site in the form of spatial coordinates. Software tools would be available to help structure the on-site survey process so that more consistent data is collected. Using computer aided recording tools, 2D or 3D CAD models could be automatically created from the resulting database. More significant than the automatic drafting would be the linking of the database and CAD drawing, allowing queries of associated data by clicking on the object while inside a CAD program or the assembly of a drawing based on a database query of object characteristics.


This automatic drafting could save vast amounts of time and provide valuable tools for managing the resources of the site as processing can be done on a laptop computer before the resource has to be altered or recorders leave the site. The most significant feature would be that all of the field records (notes, sketches, photographs and measurements) could be integrated into an easily searchable structure to which the specialized databases describing object types could be related. Queries to any of these tables would produce a CAD model with each entity of the model linked to all the tables, text files and images associated with that object, all accessible with a mouse click. .

A specific set of Computer Aided Recording Tools (CART) is currently under development to structure the on-site recording process as outlined above and to automatically create 2D or 3D models from a query of the resultant database. Subsequent queries can be made of any of the objects in the models. The following sections outline the principles on which these tools have been developed and summarize the experience of using CART to record two Canadian heritage structures.

4. UNDERLYING PRINCIPLES - THE BASIC COMPONENTS OF CART

The Database

< ------ WHERE ------ >                   WHAT   WHICH

PT_X         PT_Y     PT_Z     P_NO     P_ID    O_NA    SZ_X    SZ_Y    SZ_Z    UV_1    COMMENTS

315.000     2.828     0.000     0.0         FS          UCS1                                                         foresight
0.000         0.000     0.000     1.0         W           W1         2.500                               4          start of wall 1 2.5 high
1.000         0.000     0.500     1.1         win         W1         0.100                               2         lower left window point
2.000         0.000     1.500     1.2         win         W1                                                 2         upper right window point
4.000         0.000     0.000     1.3         W           W1                                                 1         end of first wall segment
0.000         0.500     0.000     1.4         W           W1                                                 3         jogs 0.5 from recorder
6.000         0.000     0.000     1.5         W           W1                                                 3         continues in X (layer 3)
0.000       -0.500     0.000     1.6         W           W1                                                  3         jogs towards recorder
7.000         0.000     0.000     1.7         W           W1                                                 4         end of wall 1
0.000         2.828     0.000     1.8         0             W1                                                 4         origin of "W1" in ADZ
113.000     5.385     0.000     1.9         X            W1                                                 4         direction of "W1" in ADZ

Fig. 7. A CART Database

The underlying structure of the heritage record is the database. Five fields are mandatory though their exact nature is configurable. These fields tell the program WHERE a point is located, WHAT type of object it is and WHICH specific object is recorded. The first three fields define where the point is located in space. The examples use X-Y-Z coordinates but several other systems are supported. The next field identifies what is being recorded with a P_ID (point ID) field containing a code which tells the CAD program how to deal with the point. The fifth field notes which specific object is being recorded by means of a unique O_NA, (object name). The size fields (SZ_X, SZ_Y, SZ_Z) are not always required but can be used for enhancements such as giving a size to an insertion symbol or extruding a line (representing a wall for instance) to give it a height or thickness. There is also a field, P_NO for the point number which links the database to the field notes and survey data, and one for unstructured comments. User variable fields are optional and are not used by the CAD program except for things like layer naming. They would typically contain information categories such as condition, construction date, material description, etc.

Data Collection

The first step in creating a heritage record, of course is data collection. The coordinate information can come from virtually any source, from scaling a paper drawing, to extractions from existing CAD drawings to a state of the art total station. For situations where a large number of points need to be collected quickly, we have adopted a modified surveying system which uses a visible (and reflectorless) EDM coupled with an electronic theodolite. This combination allows fast collection of points (4-5 per minute in good conditions) and levels of accuracy more than adequate for the objects in question. Additional information on size, condition, materials, etc is captured by hand in a worksheet with the information later added to the database.

The order in which measurements are taken can vary. At an archaeological site, the measurements may came first, in a polar-cylindrical format (angle,distance,depth) followed by the point number, the object name and some optional keyword fields. When recording a building where a total station is utilized to record points, the geometric information (X,Y,Z coordinates) may be collected automatically so these fields may not appear on the recording sheets and the first entry will be the point number. In the case of recording the Royal Suite and Skoki Ski Lodge, each board, log or rafter required values for its length, width and thickness, and keywords and comments were used extensively.

At the end of a day in the field the measurements collected must be entered into the database for subsequent processing. Sometimes this means manually keying in this information on any available computer. In the case of Royal Suite and Skoki Ski Lodge, the point coordinates were already in the memory chip of the total station so these were downloaded directly to the computer and the annotations and hand measurements supplementing the survey data keyed in as described above.

In the future it is hoped that programming the total station to accept all the keywords and annotations that we have been recording manually will eliminate the data entry task. Where the computer can be kept out of the elements, such as for facility recording, data can be captured directly from the survey instrument, annotated on the computer and the model generated on the spot, before changing stations.

Point Identification

The automatic component of the system is controled by the Point ID. As each point is processed, this value tells the CAD program whether to use that point to draw a line or symbol, to help define a coordinating system or to store it until enough information has been processed to allow some more complex parametric operations to be performed. At present the codes triggering lines, 3D faces and symbol insertions can be configured by the user while more complex structures such as arcs and 3D solids are pre-defined.

Object Naming

The basic intelligence of the model is a function of the Object Name which groups all the points that make up a particular object. The database is processed by the interpretation software in groups defined by this field and when it changes everything drawn under the previous object name is blocked and labelled with that value. It is these labels that provide the connection between the generated model and the various database tables describing the objects, By default, each object is linked to both the recorder's database and to an object specific table determined by the Point ID. However, any table that has this Object Name in a key field could be accessed by clicking on the object.

With the CART system the structure becomes a list of objects, walls, doors, outlets, or whatever it is you are interested in recording. When recording, a system needs to be developed from the outset that gives a structure to the names assigned to each object so that time is not wasted making decisions and objects are not lost or duplicated due to incorrectly remembered names (Nickerson 1994, 1996).(3)

What can be measured is points and surfaces (defined by collections of points). Objects represented by a single point might be nails or observations, and surfaces might be walls or floors. More complex objects such as boards, logs, stairs, furniture etc. can be handled in the model by inserting symbols to represent them so they can generally be defined by two points, one for insertion point, one for rotation, perhaps supplemented with some dimensions. The bottom line is that what we are determining is points in space and their characteristics. An ordered list of these points, grouped according to the objects of which they are a component, can be used to automatically generate the model.

The information we are collecting is, for the most part, implicit in the individual objects we measure. The figure above indicates the different types of information, how each might be captured and the path this information follows between the field and the electronic record. From the beginning this information must be carefully organized and as easy as possible to locate. We do this by means of an object name with which we christen each object as it is encountered and which becomes the key for all measurements, notes, photographs and sketches that follow. It is this framework that the CART system uses to generate the model and link the automatically generated graphics to all the other information concerning the object.

Symbology

There are two types of symbols that may be used in the model. One is the unit or 1x1x1 symbol which may be based on a single record and used simply as a comment and a connection to the database that describes the object represented. By default they are inserted at a standard scale determined by a variable from the configurable settings though if values are entered for the X, Y & Z scale factors these will be used instead. Optionally, a second point can be used to define the rotation angle and the positioning of an object can be further enhanced with the addition of a third point which allows the definition of a complete coordinate system for the insertion of the object.

The library supplied with the software contains several of this type of symbol and if others are needed they can usually be created by copying one of the existing ones to a new filename. The other type of symbol is drawn to real world scale and inserted with either two or three defining points. The most common example of this type would be a building element such as a board or log but could also include a furniture library representing a real inventory such as might be needed for a facility drawing. These would need to be created prior to generating a drawing which specifies their insertion. It is possible to use unit symbols for furniture though the scale factors for each individual component would have to be entered, a much more time consuming exercise than the creation of a symbol library.

Automatic Drafting

Traditionally a rough sketch was made of the building and then a team with tape measures would dimension the sketch and a good recording team would also be making notes about conditions, materials, colours and the like. When everything was measured the recorders would move to the drafting board or computer and render these notes into measured drawings. Much of the time taken to do this work can now be eliminated. Recording teams can focus on the collection of relevant information without having to be concerned about producing a drawing.

To the user of CART the generation of the model is simply a matter of choosing the "generate" option from the menu. The database is sorted by point number then read by AutoCAD, each record adding a point to the drawing. Much of the time the record will trigger some AutoCAD function, a line or insertion for example, and in other cases, such as arcs and coordinate systems, several points are needed before the AutoCAD function can be processed. Whenever there is a change of object name all entities associated with the name just closed are blocked and reinserted on a layer determined by the database entries for that object.

Database Integration

Using CART, a recorder measures things and make notes, and other than some clarifying sketches, the graphic representation of this data is just a by-product of the original record. In most cases the CAD model created by this method need not even be saved as it is just as easy to generate it again as it is to retrieve it, and next time you may want to represent it differently. For instance, by changing some global parameters in the settings sheet, different layer groupings can be generated, and component drawings can be created of only those objects identified by a pre-defined or interactive SQL (Standard Query Language) query. This functionality can provide a way to keep large databases, and potentially huge drawings, manageable on normal computer equipment, even laptops.

As useful as the graphic component of this tool is, the potential of the associated database is probably much more significant. As well as the obvious connection to the database that produces the model and to the other tables that support the data type, the object name is also the key to gaining access to files of other types. A code is typed or selected from the menu (VV for Images, TT fro ascii text, WW for word processor etc.) And the user is prompted to select an object. If a file exists for the object selected, in the format specified, the necessary program is launched and the file presented for viewing or editing. If no file exists the user is asked if they want to create one. Used in this way the automatically generated model becomes a front end providing access to all of the files and database tables associated with each object without the user needing to know the specific name of the object. Digital images and text files are directly supported by the software in its current manifestation but any program and file type could easily be added.

Data Storage and Archiving

The first principle of any heritage record should be that the data should be as universally accessible as possible (Nickerson reference....1996-2). This means that, as far as possible, the ability to read the computer files should not be dependent on any proprietary software (including CART). Thus the archival resting place of the geometric and keyword information should not be an AutoCAD .DWG file but a .DBF database, a file type that is readable by almost every database or spreadsheet program available today (and as it happens, AutoCAD as well). Carried to its logical conclusion this policy would suggest that these data files be written out in a comma or tab delimited format and saved as ASCII text or even printed out. In this case researchers in the future could reconstruct the site manually based on the coordinates and other information, even if they had no computer facilities at all. Alternatively, they could write an interpreter for their current software just as we have done for AutoCAD.

The database, when stored in comma delimited format, occupies an absolute minimum of disk space and, in this form, it is even possible to e-mail it to another site as pure ascii text. There it can be generated and the resulting model analysed by the specialists who, without having to visit the site, might be able to make recommendations on where the recorders should concentrate their efforts or on elements that may need more detailed study.

5. THE ROYAL SUITE

Rideau Hall, in Ottawa is the residence of the Canadian head of state and the Royal Suite is the accommodation, within this complex, which is provided for heads of state visiting from other countries. The main bedroom of the suite is oval in plan and has a domed ceiling featuring several bands of plaster relief. It was the appearance of cracks in this ceiling that prompted the study in which the CART system was used.

The original mandate was to map the cracks in a way that would both quantify their extent and allow future monitoring of their progress. With the furniture and fixtures removed the theodolite/laser tool was set up in the centre of the space and almost one thousand points were collected in less than two days. At this point however the mandate was expanded to include the structure above the dome, a considerably less straightforward task.

The first step was to establish benchmarks in the attic space to allow the coordination of the two spaces. This was accomplished by removing the sprinkler heads which penetrate the ceiling thus allowing the laser to shine from the room below, through the resulting holes and illuminate reference points on the structure and the underside of the roofing material. These benchmarks formed the basis of the coordinate systems on which the recording of the structure was based.

These spaces vary from a maximum height of less than two metres down to zero and allowed little freedom in the choice of station location. Often the instrument was screwed directly to the joists or even the ceiling support structure below them. The only real difficulty that this caused was that the adjustment of laser parallax had to be performed somewhere else where it was possible to sight through the optics after which the laser/instrument assembly had to be carefully transported to the tribrach mounted on the structure.

Four days were required to capture the over five hundred points that were needed to model these areas. This effort was well rewarded however as it was immediately obvious what was causing the problem once the resulting model was sectioned and plotted. Renovations fifty years earlier had removed webs from two of the timber trusses and these had twisted out of column causing the attic floor to sag and apply forces to the plaster dome that it was incapable of withstanding.

Another interesting discovery was that as-found recorders were not the only ones disoriented by an oval room. When setting up the drawing we had assumed an axis that ran through the centre of the central medallion at the apex of the dome and bisected both the fireplace and the plaster fans on the ceiling. We did notice that the side window was not quite where one would expect it but it was only when the attic structure was coordinated with the room below that it became clear that the window was, in fact, on the transverse axis and that it was the fireplace and fans had been installed four degrees off axis by the original craftsmen. We suspect that if these elements had been recorded by any other method this four degree discrepancy would have been assumed to be a recording error and the drawing "fixed" to reflect the assumed geometry of the space rather than its reality.

6. SKOKI SKI LODGE

Skoki Ski Lodge in Banff National Park was constructed in 1931 as a destination for back country ski holidays. It was the first commercial facility of its kind in Canada. It was constructed of local logs and stone utilizing indigenous building techniques and is a fine example of the Rustic Architecture tradition of the period. It is a 1 ½ storey log structure accommodating guest sleeping rooms, a living room, dining room and kitchen. Few changes have been made to the building since the 1930's.

The building was originally constructed by hand using simple tools and techniques. The walls were constructed of logs laid up with simple saddle notches. Roof and floor framing was entirely of logs. Each log was unique in terms of its size and placement in the structure. Individual elements were seldom placed level or square. Over time, portions of the building had settled resulting in complex assemblies of natural materials that were not level or plumb. Many of the structural elements were deteriorated either in whole or in part. Modifications or additions had been made to the original structure in the late 1930's.

The building had been designated as a National Historic Site and a complete record of the building was required to serve as the basis for a detailed condition assessment and a long term conservation and maintenance plan. The building was recorded in 1995-96 utilizing CART over a 1 week period on site with an interdisciplinary team including archaeologists, conservation architects and heritage recorders. Each individual element of the building including each log, post and rafter as well as details of all of the doors, windows and other building components was given a unique object name and recorded utilizing a combination of total station points and detailed hand recording noting size, species, condition, date of construction or modification etc. A total of over 1,000 individual elements were recorded with all of the information summarized in a database record.

The first step in recording the building was to establish a unique name for each and every element. The location and orientation of each element was then recorded using a total station to capture x,y,z coordinates, point ID and unique object name. The records where later supplemented using information from field note recording sheets noting unique characteristics including size (log diameters were measured at each end), species, condition etc. The location of unique building components such as windows, doors, chimneys etc were also recorded utilizing the total station to establish a unique 3D coordinate system for each component. This permitted the later insertion of detailed 3D records of each of these components into the CART model. Photographic and hand recorded sketch information as well as notes on condition and history etc of each element were referenced in the database and made accessible when queried from the CART model of the structure.

The heritage record for this building now resides in a database of all of the information gathered on the structure. This record has been updated and amended over the last year to reflect new information. Utilizing CART, 'traditional' heritage records in the form of plans, sections and elevations have been generated and saved in a .dwg format to serve as the basis for heritage reports and construction documents. The database has also been queried to generate 3D models of parts of the building to be used as the basis for structural analysis. Unique models of the building have also been created to illustrate the location of elements based on various queries about object type (rafter, log, purlin, post, etc), object size, date of construction, species and condition etc. Unique 'views' can be generated whenever required to support analysis and saved in a variety of formats to illustrate reports. Models of a portion or the whole building generated by CART serve as a convenient means of accessing the large database of related material (photos, sketches etc) by means of querying unique elements in the model. As the building undergoes conservation work, details regarding the location, condition or history of each new or conserved element can be easily updated in the database and made accessible. Future queries of the database will be able to 'model' the various conservation interventions over time.

7. THE PHOTOGRAMMETRIC CONNECTION

To date photogrammetry has not been used as a source of coordinate information for the CART program though photogrammetric drawings have been post-processed to build a CART database after the fact. The procedure is to take a CAD drawing and select the entities that make up each "object". These are labelled with Extended Entity Data and this data, along with the coordinates and entity information (is it a line, face etc.). While this is useful it does not provide nearly as efficient or flexible a database as would be created if the whole process was premeditated.

All that is needed to make a photogrammetric CART is to provide an interface similar to the one we use for enhancing surveyinformation that would allow the operator to specify the What & Which of an object. Then, instead of "stupid" a CAD drawing a database would be created that would allow the generation of an intelligent model as described above. Photogrammetry would also be an excellent source of the details needed for presentation drawings as the intelligence of the object, captured with only two or three surveyed (or plotted) points, will be maintained through any amount of refinement that may be deemed necessary, any time in the future.

Of particular applicability would be the semi-automatic systems such as Dipad(&Streillein) or Paros(&Florenzano et al.) as the CART system is designed to supply the supplimentary (non-geometric) information that can be used by such tools to more completely process their photogrammetric data.