Version 06/11/ :12:00

Transcription

1 AN ADDENDUM TO THE METRIC SURVEY SPECIFICATIONS FOR ENGLISH HERITAGE - THE COLLECTION AND ARCHIVING OF POINT CLOUD DATA OBTAINED BY TERRESTRIAL LASER SCANNING OR OTHER METHODS Version 06/11/ :12:00

2 Preface The purpose of this document is, in light of recent developments in metric survey, to provide information to clients and contractors on the use of terrestrial laser scanning in the survey of cultural heritage for English Heritage. The objectives of the document are: To present a user s guide that describes the recommended practices for the collection of point clouds by terrestrial laser scanning. To present the standards which point cloud data collected by terrestrial laser scanners and other methods should meet. It is not intended to provide a beginners guide to laser scanning and assumes some previous knowledge about the technique. It should be used in conjunction with the Metric Survey Specification for English Heritage - ISBN Definitions for words highlighted in bold italics can be found in Part 3 of this addendum. Acknowledgement The original addendum was prepared for English Heritage by Dr. Jon Mills and Dr. David Barber, the School of Civil Engineering and Geosciences, University of Newcastle upon Tyne. It was supported by English Heritage s Archaeology Commission under grant number 3378 MAIN. The authors were aided by a steering committee that provided valuable guidance and knowledge. The members of this committee were Bill Blake, Clive Boardman, Paul Bryan, Tom Cromwell, Tony Davies, Graham Hunter and Tony Rodgers. The preparation of this document was also aided by a variety of individuals and groups. Thanks is given to all those who gave time and expertise within English Heritage and the wider survey and cultural heritage community. Revision History This document is a revised version of the original addendum produced as part of the Heritage3D project. Heritage3D was sponsored by English Heritage s Historic Environment Enabling Programme (project 3789 MAIN) and undertaken by the School of Civil Engineering and Geosciences at Newcastle University. This two year project developed and supported best practice in laser scanning for archaeology and architecture, and disseminated this best practice to users. Further details on the project can be found at the Heritage3D website For queries please contact Mr Paul Bryan The Metric Survey Team English Heritage 37 Tanner Row YORK YO1 6WP paul.bryan@english-heritage.org.uk Tel August

3 1 Part 1: Terrestrial Laser Scanning User s Guide Contents 1.1 Introduction 1.2 Data collection Point density and data voids Coordinate systems Targeting Registration Additional data Weather Access to high areas Overview and detail scans 1.3 Use of point cloud data Primitive fitting Meshing/surfacing/DTMs D/3D line drawings Orthophotographs Screen Shots Cross section/profile generation Animation Simple measurements 1.4 Bad practice Data voids Geometric artefacts cause by the scanning process 1.5 File formats and naming 1.6 Health and safety Dangers Precautions Other considerations 3

4 1.1 INTRODUCTION Terrestrial laser scanning Use of a ground based device that uses a laser to measure the three-dimensional coordinates of a given region of an objects surface automatically, in a systematic order at a high rate in (near) real time. This user guide is intended to provide information on the recommended use of terrestrial laser scanners for the provision of point clouds for use in architectural and archaeological applications. Point clouds are suitable for use on a wide variety of subjects including small objects, details of architectural design, building facades and whole sites. Point clouds do not, however, automatically lend themselves to the generation of survey products traditionally seen in cultural heritage applications. Instead they are often more suitable for collection of surface information or use in surveys to provide 3-D spatial orientation, or where other measurement techniques are unusable (for example where there is a lack of surface texture which makes the use of photogrammetry inappropriate). Point clouds should not be seen as a replacement to existing techniques and the appropriateness of the technique chosen should always be considered. 4

5 1.2 DATA COLLECTION The variety of laser scanning systems available and their differing design and operation will mean that the exact procedures used to perform a survey will differ from system to system. However, a number of observations can be made to ensure that appropriate and high quality data is collected and delivered to the client Point density and data voids Data voids should be minimised during the scanning process through the selection of appropriate scanning positions (overlapping where necessary) and minimising temporary obstructions to the scanner during operation (such as those caused by vehicles and pedestrians). Section provides further details on data voids. In order to minimise the creation of scanning artefacts within the scanning process, point densities should be equal in both scanning axes. Section provides further details on such problems. Point density during the scanning process depends upon the range to an object. It is not, therefore, currently possible to maintain a constant point density over an entire subject during the scanning process. It is most likely however that a particular subject will be of interest, such as a façade or building detail, and that a regular density of points is preferable. The point density specified by the client or selected by the contractor should be understood as the maximum value for the subject in question. Selection of an appropriate point density should be done in consultation between the client and contractor. It is possible to filter overlapping scan data in order to reduce the point density in the final registered point cloud and hence reduce file sizes and strain on software systems during processing. Generally, this is not recommended so as to ensure a full archive of the subject is collected. It is possible to provide a guide on the appropriateness of the point density of a cloud using the following equation 1 : m Q = 1 λ Where Q is the quality of the data, m is the point density on the object and the minimum feature size. The value Q therefore indicates to the client the level to which the object has been scanned. For example, a point density of 10 mm when recording the galleting in masonry jointing made up of flint whose smallest feature is 5 mm x 5 mm: 10 Q = 1 = 1 5 which would be considered an unacceptable fit. If a density of 5 mm is used 5 Q = 1 = 0 5 indicating that the density won t consistently detect the feature over the entire object. However, a point density of 2 mm would provide: 1 Provided by Eric Lange, Veraxis Mundi 5

6 2 Q = 1 = indicating a 60% confidence the object will be detectable Coordinate systems In the majority of surveys dealing with fixed objects, such as buildings and monuments, a site coordinate system will be available to which to reference the data. Where a previous coordinate system does not exist a new system may need to be established. In order to ease the processing of point cloud in CAD packages the Z-axis should be defined as the vertical axis. A number of methods, networks and equipment are available for providing survey control. These include techniques such as intersection, triangulation or radial (direction and distance) positioning Targeting Targets located in a defined coordinate system will usually be used to transform scan data to a common system: The design of targets can change between available laser scanning systems but generally should allow the positioning of the target by an appropriate survey method. The use of targeted points, positioned using a survey instrument, such as a total station, provides additional checks on the geometric quality of the scan data. Targets should not be positioned, or be so large, that they obscure important details of the subject. Targets mounted on the surface of the subject must be fixed with an adhesive that will allow removal without damage to the surface (see Section 1.5 of the MSSEH). The use of points of natural detail should generally be avoided, although where unavoidable the use of distinct features may be suitable providing the point density of the scan is sufficient to maintain the requirements of the registration process. However, the use of features at distinct corners or edges is unsuitable due to possible errors in measurement caused by the footprint of the measurement beam Registration Although a single scan may be sufficient to fully record certain scenes, multiple scans are likely to be required, especially when dealing with a large site or structure. It is likely to be necessary to transform the collected point clouds to the local site coordinate system. The process of registration can be performed using four methods: Data collection can be performed from a station with known coordinates and a known orientation. Targets can be used to transform the data onto common coordinate systems. Surface matching algorithms can be used to transform data onto common coordinate systems. The use of surface matching algorithms alone would normally involve the use of an arbitrary coordinate system. A combination of the previous methods may be used to improve the speed and ease of the registration process. Generally, each method will provide an indication on the quality of the registration process. Data collection from a known station and known orientation will be reliant upon the precision to which the control information is known. In the case of using targeted points the quality of the registration process is best indicated by be the residuals of the transformation process 6

7 and the estimated precisions for each transformation parameter. In the case of surface matching alone the quality of the registration would be indicated by the residuals of the surface matching algorithms, along with the estimated precisions of the transformation parameters. It is unlikely that surface matching algorithms alone would be suitable for metric survey applications and that some targeted points will be required to tie the collected data to a common coordinate system Additional data Additional image data should be used to provide an overview of the subject being scanned, in addition to providing imagery for narrative purposes. This imagery should be of a high resolution and clearly portray the subject in question Weather Scanning/image capture should not be performed in adverse weather conditions where the quality of the observed data may be affected. For example, scanning in heavy rain may lead to data voids due to falling raindrops or erroneous data points due to returns from airborne raindrops or erroneous range measurement due to refraction of the measurement beam Access to high areas Laser scanning relies on a line of sight to a target and, therefore, where a subject of a significant height is to be surveyed access to higher areas may be a problem. Also, as laser scanning is not an instantaneous measurement process methods used to gain access to higher areas must provide a platform stable enough to support the scanning instrument for the duration of the scan Overview and detail scans A combination of overview and detail scans provides both a clear description of the general area in question in addition to providing detailed information of selected parts of a site or object. It is likely that scanning surveys will combine a mixture of scans. 7

8 1.3 USE OF POINT CLOUD DATA Point cloud data is not necessarily suitable for the traditional products currently commissioned by English Heritage. The following list provides example products where point cloud data may be of use: Primitive fitting Stonework and architectural features are unlikely to be acceptably modelled using CAD primitives, although in some situations it may be acceptable to use such a modelling process Meshing/surfacing/DTMs Laser scanning is an excellent method for the collection of surface information. It is likely that meshing will therefore be one of the most appropriate modelling techniques. This may involve the use of algorithms to simplify the mesh in areas of low topography and hence reduce the strain on display and modelling systems. Meshing/surfacing is generally reserved for non-topographic surfaces. A digital terrain model (DTM) relates to terrain, rather than arbitrary surfaces (which may be on a building or object). Any DTMs should be generated with vegetation and (non-terrain) man made objects removed. Tynemouth Priory, West Door: meshed model 8

9 Clifford s Tower, York: Motte DTM and point cloud of the tower D/3D line drawings The extraction of vector information from point clouds is possible, albeit at present using largely manual processes. See Section 3 and Section 4 of the MSSEH for more details on the standards these products should meet Orthophotographs Orthoimages are normally generated using a surface model collected from stereo photography. Such surfaces models may previously have had errors due to depth discontinuities. For this reason point cloud data from laser scanning may be of utility in generating orthophotographs. Refer to Section 4 of the MSSEH for more details Screen Shots Screen shots are simple but effective ways of displaying the scan data, however, assumptions that may be made by non-professionals should be considered. For example the mixing of orthographic and perspective projections, or the colouring of scans using different palettes. 9

10 Tynemouth Priory, West Door point cloud (greyscale intensity information) Perspective view, May 2003 Tynemouth Priory, West Door point cloud (greyscale intensity information) Orthographic view, May

11 Tynemouth Priory, West Door point cloud detail (greyscale intensity information) Perspective view, May Cross section/profile generation Scan data may also be used to create cross sections or profiles. The height and location at which these are generated should be considered. Tynemouth Priory, West Door 1 m profile (above ground) Orthographic view, May Animation The use of animations to display point clouds and surface meshes allows the user to understand the 3-D nature of the data, in addition to rapidly gaining a 3-D understanding of a site or structure Simple measurements The point cloud may be used by the end user for simple interpretation processes such as the collection of basic dimensions of an object or scene. 11

12 1.4 BAD PRACTICE Data voids Data voids should be minimised during scanning. Data voids due to occlusions to the line of sight can normally only be eliminated by using multiple scans. Data voids due to temporary obstructions such as pedestrians or vehicles should be limited by appropriate positioning of the scanner or restriction of such obstructions. Data voids due to permanent obstructions (top and side views) 12

13 Points recorded on a temporary obstruction (in this case the side of a moving vehicle) A data void resulting from the temporary obstruction Geometric artefacts cause by the scanning process A number of situations can occur that result in scanning artefacts. Uneven point densities for example may lead to the lack of definition of horizontal or vertical features. Therefore, point densities should be kept equal in both axis of measurement. 13

14 AN ADDENDUM TO THE METRIC SURVEY SPECIFICATIONS FOR ENGLISH HERITAGE Use of an even point density giving even emphasis on horizontal and vertical mortar joints Use of an uneven point density resulting in less emphasis on vertical mortar joints Scan artefacts may also occur at distinct geometric edges on a surface, where due to the footprint of the laser erroneous points are recorded. Although this is minimised through the use of lasers with a small measurement footprint this effect may still occur. 14

15 Footprint 2 Footprint 1 Laser beam The foot print of the beam is reflected from two widely separated surfaces resulting in a point erroneously positioned between the surfaces (after Dias et al., ). Additional scan artefacts may also occur through poor framing of a scan may result in the limits of scan data being interpreted at actual geometric edges, even though they may simply have been created by poor positioning of the scanner. Example of the poor framing of scans The reflection of highly reflective materials (generally retro reflective) can also result in scan artefacts. Consideration should be made to the surface types in question before scanning takes place. 2 DIAS, P., SEQUEIRA, V., GONCALVES, J. G. M. and VAZ, F., Combining intensity and range images for 3D architectural modelling. International Symposium on Virtual and Augmented Architecture. Dublin:

16 The result of scanning a piece of flat high intensity material attached to a flat wall (viewed from above) The material as it was when scanned 16

17 1.5 FILE FORMATS AND NAMING The Archaeological Data Service (ADS) provides recommended practices for dealing with file formats and file naming. In particular: Do use only alphanumeric characters (a-z, 0-9), the hyphen (-) and the underscore (_). Both upper and lower case characters and numbers can be used in a filename but keep file names within your project consistent and ensure that supplied documentation accurately reflects the case of your filenames. Don t use spaces or full stops (.) within filenames. Full stops should only be present where the filename is separated from the file extension e.g..doc or.pdf. Spaces can usually be replaced with the underscore (_) character. Do use a consistent scheme and case when naming files. A descriptive filename helps explain the contents of the file, for example _trench_1.tif could be a digital photograph of trench 1 taken on 12/10/2004. A non-descriptive file name might be a unique id number allocated to an image within an accompanying image catalogue database. Non-descriptive filenames are acceptable but their content must be adequately described in accompanying metadata. Consistent use of case ensures that files can be reliably identified on case-sensitive operating systems such as UNIX where report.doc would be recognised as a different file to Report.doc. While no standard format for point clouds derived from laser scanning exists, the ASPRS LAS format can generally be used to facilitate easy transfer of point cloud data between users. This format has been used for the transfer of airborne datasets, but may also be used for terrestrial scan data. 17

18 1.6 HEALTH AND SAFETY The European Standard Safety of Laser Products Part 1: Equipment classification, requirements and users guide (IEC : 2001) provides information on lasers and describes precautions on the use of laser products. Users should refer directly to this document when preparing health and safety assessments. However, a brief summary is provided below Dangers Lasers used in survey applications may have risks associated with eye damage. The European Standard (IEC : 2001) provides seven classes of lasers: Class 1 lasers are safe under reasonably foreseeable conditions of operation, including the use of optical instruments for intrabeam viewing. Class 1M lasers are safe under reasonably foreseeable conditions of operation, but may be hazardous if optics are employed within the beam. Class 2 lasers normally evoke a blink reflex that protects the eye, this reaction is expected to provide adequate protection under reasonably foreseeable conditions, including the use of optical instruments for intrabeam viewing. Class 2M lasers normally evoke a blink reflex that protects the eye, this reaction is expected to provide adequate protection under reasonably foreseeable conditions. However, viewing of the output may be more hazardous if the user employs optics within the beam. Class 3R lasers are potentially hazardous where direct intrabeam viewing is involved, although the risk is lower than that for Class 3B lasers. Class 3B lasers are normally hazardous when direct intrabeam exposure occurs, although viewing diffuse reflections is normally safe. This class of laser is generally not suited for survey applications. Class 4 lasers will cause eye or skin damage if viewed directly. Lasers of this class are also capable of producing hazardous reflections. This class of laser is not suited for survey applications. Users of laser scanning systems should always be aware of the class of their instrument. In particular the user should ensure the correct classification system is being used (e.g. IEC : 2001 and not BS EN : 1994 or other standard which differ slightly in classification) Precautions The European Standard IEC EN : 2001 provides a number of safety precautions that should be observed during use of laser scanning surveys. For lasers up to Class 3R (those normally used in survey applications) and where applicable to laser scanning for metric survey these precautions are briefly outlined below. However, for a full description the user is referred directly to the European Standard. Generally: Care should be taken to prevent the unintentional specular reflection of radiation. Open laser beam paths should be located above or below eye level where practical. Only persons who have received training to an appropriate level should be placed in control of laser systems. The training, which may be given by the manufacturer or supplier of the system, the laser safety officer or an approved external organisation should include, but is not limited to: familiarization of operating procedures; the proper use of hazard control procedures, warning signs etc; the need for personal protection; accident reporting procedures and bioeffects of the laser upon the eye and skin. 18

19 Particular care should be taken through the use of magnifiers or telescopes around laser devices that may pose a risk when intrabeam viewing is used. The instrument should only be used in accordance with the manufacturer s instructions. For lasers that emit energy outside the wavelength range of 400 nm to 700 nm special considerations are often required. For example: Where using a Class 3R laser a laser safety officer should be appointed. Beam paths should be as short as possible and avoid crossing walkways and access routes. Particular precautions and procedures are outlined in the IEC standard for Class 1M, Class 2M and Class 3R laser products used in surveying, alignment and levelling. Those with relevance to laser scanning are: Only qualified and trained persons should be assigned to install, adjust and operate the laser equipment. Areas where these lasers are used should be posted with an appropriate laser warning sign. Precautions should be taken to ensure that persons do not look into the beam (prolonged intrabeam viewing can be hazardous). Direct viewing of the beam through optical instruments (theodolites, etc.) may also be hazardous. Precautions should be taken to ensure that the laser beam is not unintentionally directed at mirror-like (specular) surfaces. When not in use the laser should be stored in a location where unauthorized personnel cannot gain access Other considerations In addition to the risks associated with lasers, users should be aware that due to the size and weight of some systems there is a risk of injury to visitors, especially children, if systems are left unaccompanied. The effect of laser scanning on features such as lichens and delicate fabrics is not well understood. Consideration should be given to the use of lasers in the vicinity of such features. 19

20 2 Part 2: A standard specification for the collection of point cloud data by terrestrial laser scanning or other methods Contents 2.1 Introduction 2.2 Provision of point cloud data Pre-survey deliverables Certification requirements Point density and measurement precision Overview and detail scans Overlapping scans Data voids High level coverage Methods used and required accuracy of control Registration procedures Targeting/control points Intensity/colour Supporting imagery Delivery of survey material 2.3 Health and safety 2.4 Storage and archive of point cloud data Data format File naming convention Scan metadata Project metadata Registration information Control information File sizes Media Retention of survey documentation 2.5 Example metadata Scan position metadata Project metadata 20

21 2.1 INTRODUCTION The specification described here defines the standard which point clouds must meet if they are to be accepted by English Heritage. It does not define the standard to which products derived from point clouds must attain. The specification should be consulted in combination with the user s guide contained in Part 1 of this addendum and the MSSEH. 21

22 2.2 PROVISION OF POINT CLOUD DATA Pre-survey deliverables Prior to survey a method statement as defined in Section of the MSSEH is required. In the case of terrestrial laser scanning the method statement will also include: Technical specifications about the scanning system, or systems, to be used. The proposed point density. A description outlining the location and extent of potential data voids and a proposed method for data collection in these areas Certification requirements Laser scanning systems used must be accompanied with: A certificate confirming the system is in good working order; Or details of tests, performed in the last 12 months, which show the scanner to be achieving the required precision and accuracy. Exact requirements for certification or tests should be discussed with English Heritage and described in the method statement before work begins Point density and measurement precision The accuracy and point density required will be stated in the project brief. This will either: Be defined based on a scale of survey: SCALE EFFECTIVE POINT DENSITY PRECISION OF MEASUREMENT TYPICAL USE 3 1: mm +/- 2.0 mm Small details/objects (up to 5 m x 5 m) 1: mm +/- 4.0 mm Larger details/objects (up to 10 m x 10 m) 1: mm +/ mm Small structures (up to 20 m x 30 m) 1: mm +/ mm Large structures (up to 40 m x 60 m) Or based a minimum feature size defined in the project brief. For example: The minimum size of feature required to be discernable in the point cloud is 10mm in depth and 10mm in width and height. Users should see Section on the use of a Q value to quantify the level to which the minimum feature size has been achieved. The footprint diameter of the measurement beam must not be greater than double the effective point density. All reference to point density will be provided as the average 3-D distance between points at a defined range. 3 Typical use is an indicator only the project brief will specify the exact requirements for point density and precision of measurement. 22

23 2.2.4 Overview and detail scans The number and location of overview and detail scans required will be specified in the project brief were required Overlapping scans Where it is acceptable to filter areas of overlapping scan data, to reduce the point density in the final registered point cloud and hence reduce file sizes and improve software performance/data handling during processing, this will be noted in the project brief Data voids The number of data voids must be minimised during the survey. The project brief will outline the requirements for handling and acceptance of data voids in point cloud data High level coverage Methods used to achieve high level coverage must be described in the method statement and outlined in the final survey report Methods used and required accuracy of control The methods and networks used for providing survey control are discretionary. However, details of the method and equipment proposed must be included in the method statement. Where a previously defined survey co-ordinate system exists: The necessary information will be supplied in the project brief to allow the reoccupation of previously installed points. This will include a full listing of 3-D coordinates and witness diagrams. Where a previous survey co-ordinate system does not exist: A new system may be established as described in Section of the MSSEH. Individual survey control points are to be provided to a geometric precision/accuracy of twice the geometric precision/accuracy required by individual measurements Registration procedures The residuals of the registration process must be shown to be equal to or better than the geometric precision required by the end deliverable. Where registration is done solely via a resection calculation: Each scan must contain a minimum of 4 appropriately distributed XYZ control points/targets. The residuals of the registration process and the geometric precision of the estimated parameters should be noted in the survey report. Where registration is performed using surface matching techniques: The data must include at least n + 3 appropriately distributed XYZ control points/targets, where n is the number of scans made. The residuals of the registration process and the precisions of the estimated parameters should be noted in the survey report. The geometric accuracy of the fit should be noted in the survey report. Where registration is done using a known station position and orientation: The data must include at least 3 appropriately distributed XYZ control points. 23

24 The residuals of the registration process and the precisions of the estimated parameters should be noted in the survey report. Irregular features in the scan data caused by cracks or features on the subject that could be misinterpreted as errors in the registration must be augmented with illustrative photography and noted in the final survey report Targeting/control points Targets must not be positioned, or be so large, that they obscure important details of the subject. In addition targets mounted on the surface of the subject must be fixed with an adhesive that will allow removal without damage to the surface. A description of the targets to be used must be given in the method statement and the location and naming of targets is to be clearly given on the site sketches that accompany the survey report. The use of natural detail points should be avoided, but where necessary the use of distinct features is acceptable providing the point density of the scan is sufficient to maintain the registration requirements in Section of this document. The use of features at distinct corners or edges is not permitted. Where natural detail points are to be used this must be noted in the method statement Intensity/colour Intensity/colour information will be recorded on a per point basis at each scan position where the instrumentation allows this and such information has been specified in the project brief Supporting imagery Additional image data to show the location of the scanner and the subject being scanned is required for narrative purposes. This imagery will be of a high resolution and clearly portray the subject in question. It will be delivered in a manner that agrees with the current MSSEH Delivery of survey material Certain standard deliverables are required for every survey performed. See Section for a full description of the appropriate media, formats and required metadata. The standard deliverables in digital form, unless stated in the project brief, are: Project metadata. Raw scan data (for archive). Scan metadata Control information. Registration information for all raw scans to the site coordinate system. Registered scan data (for archive). A survey report, in PDF format, is also required containing: Witness/illustrative diagrams outlining the position of scanning stations and control points. Details of the traverse/control network used, a list of the three-dimensional positions of all control points and residuals for the computed XYZ control. The precision of any parameters derived in the registration process for each scan along with the residuals of the registration. A summary outlining the completeness of the point cloud and all known data voids. Any site sketches/additional field notes made during field work. 24

25 Two copies of all data and reports are required. 25

26 2.3 HEALTH AND SAFETY Readers are referred to IEC 60825:1 (2001) for the full precautions on the user of lasers. However, explicitly: Only an appropriately trained individual may operate a laser scanner on site. Signs warning visitors that lasers are in use must also be displayed. Systems that use Class 3B or Class 4 lasers are not acceptable for use on English Heritage sites. 26

27 2.4 STORAGE AND ARCHIVE OF POINT CLOUD DATA Data format To assist in the future management of scan data all data is required to be delivered in a prespecified format with emphasis on the transferability of data between software systems. The raw scan data (the data collected by the scanner before pre-processing) should normally be delivered in the ASPRS LAS format. Contactors must ensure all of the standard header information is included. Until the LAS format is revised to include registration information relating to terrestrial laser scanning, contractors should where available, use the Heritage3D LAS implementation which includes records for registration information. Where this isn t possible LAS V1.1 should be used and registration information supplied as separate information File naming convention Filenames should be logical and be located in an appropriate directory structure. The contractor must provide a readme file to document the contents, structure and naming of the delivered data. Only use alphanumeric characters (a-z, 0-9), the hyphen (-) and the underscore (_) in naming files. Both upper and lower case characters and numbers can be used in a filename. Don t use spaces or full stops (.) within filenames. Full stops should only be present where the filename is separated from the file extension e.g..doc or.pdf. Spaces can usually be replaced with the underscore (_) character. Contractor's should use a consistent scheme and case when naming files. A descriptive filename should be used that helps explain the contents of the file, for example _trench_1.tif could be a digital photograph of trench 1 taken on 12/10/2004. A nondescriptive file name might be a unique id number allocated to an image within an accompanying image catalogue database. Nondescriptive filenames are acceptable but their content must be adequately described in accompanying metadata Scan metadata Metadata (information relating to the captured information) is required with all raw scan data and scanning projects. Metadata should be provided in both hardcopy and digital form. It must include: File name of the raw data Date of capture Scanning system used (with manufacturers serial number) Company name Monument name Monument number (if known) Survey number (if known) Scan number (unique scan number for this survey) Total number of points Point density on the object (with reference range) A record of the weather conditions during scanning (external scanning only) 27

28 2.4.4 Project metadata A single project metadata file is required with the project. This must include the following: Filename(s) of the raw data used in the registration Data of capture (month and year) Scanning system(s) used (with manufacturers serial number(s)) Company name Monument name Monument number (if known) Survey number (if known) Number of individual scans Scan numbers of all scans Total number of points Filename of the control data Description of registration method (e.g. All scans registered to local site grid using targeted points. ) An index plan showing the data collected with individual scan points named Weather during survey (external scans only) Any scanner specific information Registration information The following information should be supplied as registration information: Translations in the X, Y and Z axes necessary to transform the scan origin to the scan position. Rotations around the X, Y and Z axes. This should be carried out in the order X, Y and Z Control information The following information should be supplied as control information: Point ID, X, Y, Z, DX, DY, DZ, comment (optional) File sizes Individual file sizes are to be limited to the capacity of a standard single (CD 640 MB). The compression of files is acceptable using standard compression software such as Winzip. If capacity allows, multiple scans can be placed on a single CD Media Unless otherwise stated, all data is to be provided on Gold CD-ROMs, DVDs, Hard Disk or, by prior arrangement, electronic transfer. Any text referencing is to be provided on a suitable label applied to the top surface of the CD-ROM/DVD. On no account must any text be written directly onto the surface of the CD-ROM/DVD Retention of survey documentation On request the Contractor shall make available to English Heritage all materials used for the compilation of the required survey. This information must be retained on file by the contractors for a minimum of six years. 28

29 This will include: field notes and/or diagrams generated whilst on site; the raw and processed data used for the final computation of coordinate and level values; and a working digital copy of the Metric Survey data that forms each survey (this is to include formatted 2-D and raw 3- D data files). The precise digital format and file type of this archive is that specified in the project brief. If during this period the contractor wishes to change the format of this data archive, they are to seek English Heritage s permission. 29

30 2.5 EXAMPLE METADATA Scan position metadata PARAMETER EXAMPLE File name of the raw data: TYP03L02.xyz Scanning system used (with serial number): Cyrax 2500 #055 Monument name: Tynemouth Priory Survey number (if known): NA Total number of points: Date of capture: 29/05/2003 Company name: University of Newcastle Monument number (if known): NA Scan number (unique scan number for this 2 survey): Point spacing on the object: m (@ 30 m) Weather during survey: Sunny and calm 30

31 2.5.2 Project metadata PARAMETER EXAMPLE Filename(s) of the raw data used in the registration TYP03L01.txt (1 in index plan) TYP03L02.txt (2 in index plan) TYP03L03.txt (3 in index plan) TYP03L04.txt (4 in index plan) TYP03L05.txt (5 in index plan) TYP03L06.txt (6 in index plan) Data of capture (month and year) May 2003 Scanning system(s) used (with serial number(s)) Cyrax 2500 #055 Company name University of Newcastle Monument name Tynemouth Priory Monument number (if known) NA Survey number (if known) NA Number of individual scans 6 Scan numbers of all scans 1-6 Total number of points Description of registration method All scans registered to local site grid using targeted points and resection calculation Filename of control data TYP03H.xyz Weather during survey Sunny and calm An index plan showing the data collected with See below individual scan points named 4,5,6 2,3 1 31

32 Part 3: Glossary Data voids Sections within the point cloud, more than twice the point density of the scan in size, which contain no data despite surface information on the object itself. Data void Scan points Detail scans Scans using a high point density that give a detailed representation of a feature (see also overview scans): Geometric accuracy The closeness of a measurement to its true value. This is commonly described by the RMS error. Where ε is the error (the difference between the measurement and the true value) and n is the number of measurements the RMS error can be defined as: 32

33 σ = ± n i= 1 n ε 2 i Geometric precision The distribution of a set of measurements about the average value. This is commonly described by the standard deviation. If n is the number of measurements and v is the residual (i.e. the difference between a measured quantity and the most probable value for that quantity) the precision of a sample set of data can be represented by the standard deviation: S = ± n i= 1 v 2 i n 1 All reference to the standard deviation of a quantity should be accompanied by the probable error value e.g. +/- 3 mm (67% probable error.) Minimum feature size The size of the smallest geometric feature required for measurement. MSSEH MSSEH The current Metric Survey Specification for English Heritage (ISBN ). Overview scans Scans using a low point density intended to give an overview of the site/area being surveyed (see also detail scan): 33

34 Peripheral data Additional scan data collected during the scanning process not explicitly defined in the project brief. Point cloud A collection of XYZ coordinates in a common coordinate system that portrays to the viewer an understanding of the spatial distribution of a subject or site. It may also include additional information such as an intensity or RGB value. Generally a point cloud contains a relatively large number of coordinates in comparison with the volume the cloud occupies, rather than a few widely distributed points. Point density The average distance between XYZ coordinates in a point cloud. This is commonly represented in two forms, spatial and angular. The average 3-D distance between points at a specified distance (assumed to be taken from/to the centre of the footprint): The angular step of the two axes: 30 mm x 30 mm (at 50 m) 0.25 x 0.25 The average 3-D distance between points on the subject is likely to be clearer to a nontechnical user. The quoted point density for any scan should always be seen as the minimum density of points for data on the subject, not including peripheral data. Raw scan data Any unedited data collected by a laser scanner during one scan. This includes any peripheral data captured during the scanning survey. Registration The process of transforming point clouds onto a common coordinate system. Scan The noun Scan refers to a single session of data collection at a single scan origin and scan orientation. The verb Scan is the act of using a terrestrial laser scanner. Scan orientation The approximate direction in which a scan is performed if the system does not provide a 360 degrees field of view. Scan origin The origin of the arbitrary coordinate system in which scans are performed (normally set as 0, 0, 0). When the scan origin is transformed into the site coordinate system it is known as the scan position. Scan position 34

35 The location, in a known coordinate system, from which a single scan is performed. If the system does not perform a 360 degree scan, several scans may be taken from the same scan position, but with different scan orientations. Scanning artefacts Irregularities within a scan that are a result of the scanning process rather than features on subject itself see Section 1.4. These may be geometric or radiometric in nature. System resolution The smallest discernable unit of measurement of the laser scanning system. Terrestrial laser scanner Any ground based device that uses a laser to measure the three-dimensional coordinates of a given region of an objects surface automatically, in a systematic order at a high rate in (near) real time. This includes scanners used to fixed objects, such as buildings and monuments, and scanners used to scan small objects and artefacts (sometimes referred to as close range laser scanning). 35