GISdevelopment.net --> Application --> Archaeology
Spatial information technologies to aid Archaeological Site Mapping
Dr Alan Forghani
Managing Director, Australian Spatial Technologies Pty Ltd
48 Katherine Avenue, Amaroo 2619 Act Australia
alan_forghani@hotmail.com
Denise Gaughwin
Forest Practices Board, Forestry Tasmania
PO Box 180, Kings Meadows, Tas 7249, Australia
deniseg@fpb.tas.gov.au
Historic sites have commonly been identified, located, and mapped by conventional ground surveys with the use of low to medium altitude aerial photography. Archaeologists and spatial information specialists are using digital remotely sensed image analysis plus ancillary spatial data to improve the efficiency of this process. Application of remote sensing imagery and GIS provide a more flexible and efficient way of identification, recording, and mapping historic sites and their association such as linear features (e.g. roads, vegetation alignments, drainage networks etc). In this work an integrated digital RS and GIS approach were used for identification and mapping a historic road network in the upper Forth River, North West Tasmania. Our particular interest was to reconstruct a section of one of these roads built in the 1820s by a private agricultural company. To identify and map part of the network, a GIS database, remotely sensed imagery, and field investigation were used.
Spatial information technologies to aid Archaeological Site Mapping
Dr Alan Forghani
Managing Director, Australian Spatial Technologies Pty Ltd
48 Katherine Avenue, Amaroo 2619 Act Australia
alan_forghani@hotmail.com
Denise Gaughwin
Forest Practices Board, Forestry Tasmania
PO Box 180, Kings Meadows, Tas 7249, Australia
deniseg@fpb.tas.gov.au
The application of an integrated GIS and remote sensing images to road finding and updating has been widely discussed. Key references are Van Cleynenbreugel et al (1990), Domenikiotis et al (1995), and Forghani (1999). The use of spatial technology into archaeological application have been discussed in the contemporary literature (e.g. Waara-Aswapatic, 1995, Forghani and Gaughwin, 2000). From early 1800 the British colonists of Van Diemen Land now known as Tasmania, created an infrastructure of roads and tracks. Parts of roads made by Aborigines were used by the early settlers, although due to very different economic systems of the settlers and the Aborigines there was very little overlap in access requirements (Rollins, 1988). Much of the knowledge of the location of these early alignments has been lost or over written, as there has been little later permanent settlement in the area under investigation. Timber harvesting activities may have reused the road or have snigged over the alignment altering its structure (Forghani and Gaughwin, 2000).
This paper considers sections of one of these roads built by the agricultural company, the Van Diemen™s Land Company (VDL). Spatial technology facilitated to reconstruct the missing sections of VDL Investigation. Timber harvesting activities may have reused the road or have snigged over the alignment altering its structure (Forghani and Gaughwin, 2000). This paper considers sections of one of these roads built by the agricultural company, the Van Diemen™s Land Company (VDL). Spatial technology facilitated to reconstruct the missing sections of VDL.
Fig. 1: Topographic map of the study area (approximately 1:100,000)
Extensive forest clearing for the purpose of timber production has been undertaken. Fig. 2 illustrates the vegetation spp. The rapid transition from native grassland clearing used for pastoral purposes can be seen when Figures 2 and 3 are compared. Fig. 4 demonstrates road construction for the purpose of forest logging. Roads and tracks in the study area have varying width, varying from 1 to 8 m width. Distinguishable major road classes include: (a) Highways are located in the northern and western of the area with a width of 8 to 5 metres, (b) Rural main roads extended in two directions, starting from north to south east to south west with a width between 4 to 6 metres, and (c) Minor roads and tracks located mostly in the forest areas with a width of 1.5 to metres.
Fig. 2: Native Eucalyptus within the area
Project Implementation
The following steps were involved in the completion of this work using standard software; ERDAS IMAGINE, ARC/INFO, and ArcView:
Fig. 3: Land used for pasture.
Estes et al (1987) pointed out that development and application of a GIS needs consideration of the study area, existing reports and maps, and the requirement for field investigation and data collection. It brings to attention of the importance of database development for this work. In order to develop a GIS database for this project, a number of tasks have been conducted such as:
Data subseting
A sub-region of the image was chosen for image rectification, which contained the missing VDL segments, and these need to be examined in more depth.
Fig. 4: Construction of an access road in a forest area.
Image Correction
To rectify digital images an afine interpolation of a triangulation and rubber sheeting using ARC/INFO GIS software were employed. Afine interpolation of a triangulation tries to overlay images with local geometric distortions using a network of relative displacement (links). It employs an interpolation algorithm, and rubber sheeting applies a triangulation technique to divide the adjustment region into transformation segments. This method provided 5-8 m accuracy, which was required for this work. The corrected image with digital road network overlaid is shown in Fig. 6. The corrected images were used to locate and map the route of the VDL Company roads.
Fig. 5: Tracks in the forests
Image Classification
Supervised classification approach was applied to delineate roads network employing ERDAS IMAGINE software to generate land use of the study area using a maximum likelihood classifier (MLC) technique. The prime concern is to map out roads, then generated land use classes are merged until two classes were produced namely as roads, and non-roads (background) which comprised classes that do not have any association with roads. This map is used in the process of road location.
Data Editing and Field Verification
As the roads coverage provided by DELM was based on 1989 map compilation, some parts of the VDL roads network exists in the topography map. Some parts of the VDL network did not show up those data since much of the knowledge of the location of the early alignment has been lost or overwritten as there has been little later permanent settlement in the area. Fig. 6 also provides missing VDL road segments which are presented in red colour, and the road segments that did not appear on the available digital data were digitised and edited into the VDL road coverage. This information has been compiled from field verification, local knowledge, and ancillary maps in order to reconstruct and update the VDL network.
Ancillary data and maps include topographic map sheets of Lea (4239), Pencil Pine (4039), Liena (4040), County Chart (Lincoln 1), Forestry Commission Forest Type (Daisy Dell), Archives of Tasmania VDL Co maps and Sorell Survey maps 1877 used to extract DVL segments and were input onto a GIS. The County Chart and Forestry Commission Forest Type were useful in identifying some sections of the route. But the older maps did not provide significant data other than giving an historical perspective.
Screen and tablet digitising from geocoded imagery and the topographic map respectively using ARC/INFO software generated the roads and tracks of the study area. Then maps were joined by an edge-matching technique. Field data collection has been undertaken by an archaeologist, a GIS operator, and foresters using a number of maps to verify local knowledge about the existence of VDL roads network. In addition, GPS (GPS Explorer) were used to verify both local knowledge, and the information from the maps. It was difficult to locate and identify some parts of VDL network. Much of the knowledge of the location of the early alignment has been lost or over-written, as there has been little later permanent settlement in the area. In addition, timber harvesting activities and logging practices in the 20th century may have reused the road or have snigged over the alignment altering its structure. For example, in the Liena area it was found that there are many tracks that either run parallel to each other or intersect one another. Even local knowledge could not assist in identification of which of these was the actual VDL segment. Thus, the only solution was to specify an approximate track (line) to represent the VDL route (Fig. 7).
Fig. 6: Shows road networks red and yellow were overlaid onto the georeferenced image. The yellow lines are the existing roads on the topographic map. The red lines are the compiled VDL road networks.
Remarks
Utilised GIS/RS integrated approach provided a scientifically justifiable and generally applicable approach to variable historic site identification, recording, and mapping. The output can be potentially integrated into a planning framework to support resource and environmental planning, management, and decision making. Despite the fact that this report has been able to provide the best possible estimate of the location of the road, we recommend that further ground truthing will be required to verify this conclusion.
Fig. 7: Diagrammatic representation of the reconstructed of the VDL route.
References
This paper considers sections of one of these roads built by the agricultural company, the Van Diemen™s Land Company (VDL). Spatial technology facilitated to reconstruct the missing sections of VDL Investigation. Timber harvesting activities may have reused the road or have snigged over the alignment altering its structure (Forghani and Gaughwin, 2000). This paper considers sections of one of these roads built by the agricultural company, the Van Diemen™s Land Company (VDL). Spatial technology facilitated to reconstruct the missing sections of VDL.
Fig. 1: Topographic map of the study area (approximately 1:100,000)
Extensive forest clearing for the purpose of timber production has been undertaken. Fig. 2 illustrates the vegetation spp. The rapid transition from native grassland clearing used for pastoral purposes can be seen when Figures 2 and 3 are compared. Fig. 4 demonstrates road construction for the purpose of forest logging. Roads and tracks in the study area have varying width, varying from 1 to 8 m width. Distinguishable major road classes include: (a) Highways are located in the northern and western of the area with a width of 8 to 5 metres, (b) Rural main roads extended in two directions, starting from north to south east to south west with a width between 4 to 6 metres, and (c) Minor roads and tracks located mostly in the forest areas with a width of 1.5 to metres.
Fig. 2: Native Eucalyptus within the area
Project Implementation
The following steps were involved in the completion of this work using standard software; ERDAS IMAGINE, ARC/INFO, and ArcView:
- Acquiring topographic maps at 1:25,000 scale, digital roads and land use, and other ancillary historic maps and reports.
- Acquiring colour and B/W aerial photography at 1:42,000 captured in 1994
- Scanning of the images with 600 DPI and stored in a 25 MB TIFF format
- Importing data into ARC/INFO software
- Rectification and registration of the image data
- Development of GIS Database
- Field data collection via local knowledge and GPS
- Image classification to extract the recent land use map
- Reconstruction of missing sections of the network
Fig. 3: Land used for pasture.
Estes et al (1987) pointed out that development and application of a GIS needs consideration of the study area, existing reports and maps, and the requirement for field investigation and data collection. It brings to attention of the importance of database development for this work. In order to develop a GIS database for this project, a number of tasks have been conducted such as:
Data subseting
A sub-region of the image was chosen for image rectification, which contained the missing VDL segments, and these need to be examined in more depth.
Fig. 4: Construction of an access road in a forest area.
Image Correction
To rectify digital images an afine interpolation of a triangulation and rubber sheeting using ARC/INFO GIS software were employed. Afine interpolation of a triangulation tries to overlay images with local geometric distortions using a network of relative displacement (links). It employs an interpolation algorithm, and rubber sheeting applies a triangulation technique to divide the adjustment region into transformation segments. This method provided 5-8 m accuracy, which was required for this work. The corrected image with digital road network overlaid is shown in Fig. 6. The corrected images were used to locate and map the route of the VDL Company roads.
Fig. 5: Tracks in the forests
Image Classification
Supervised classification approach was applied to delineate roads network employing ERDAS IMAGINE software to generate land use of the study area using a maximum likelihood classifier (MLC) technique. The prime concern is to map out roads, then generated land use classes are merged until two classes were produced namely as roads, and non-roads (background) which comprised classes that do not have any association with roads. This map is used in the process of road location.
Data Editing and Field Verification
As the roads coverage provided by DELM was based on 1989 map compilation, some parts of the VDL roads network exists in the topography map. Some parts of the VDL network did not show up those data since much of the knowledge of the location of the early alignment has been lost or overwritten as there has been little later permanent settlement in the area. Fig. 6 also provides missing VDL road segments which are presented in red colour, and the road segments that did not appear on the available digital data were digitised and edited into the VDL road coverage. This information has been compiled from field verification, local knowledge, and ancillary maps in order to reconstruct and update the VDL network.
Ancillary data and maps include topographic map sheets of Lea (4239), Pencil Pine (4039), Liena (4040), County Chart (Lincoln 1), Forestry Commission Forest Type (Daisy Dell), Archives of Tasmania VDL Co maps and Sorell Survey maps 1877 used to extract DVL segments and were input onto a GIS. The County Chart and Forestry Commission Forest Type were useful in identifying some sections of the route. But the older maps did not provide significant data other than giving an historical perspective.
Screen and tablet digitising from geocoded imagery and the topographic map respectively using ARC/INFO software generated the roads and tracks of the study area. Then maps were joined by an edge-matching technique. Field data collection has been undertaken by an archaeologist, a GIS operator, and foresters using a number of maps to verify local knowledge about the existence of VDL roads network. In addition, GPS (GPS Explorer) were used to verify both local knowledge, and the information from the maps. It was difficult to locate and identify some parts of VDL network. Much of the knowledge of the location of the early alignment has been lost or over-written, as there has been little later permanent settlement in the area. In addition, timber harvesting activities and logging practices in the 20th century may have reused the road or have snigged over the alignment altering its structure. For example, in the Liena area it was found that there are many tracks that either run parallel to each other or intersect one another. Even local knowledge could not assist in identification of which of these was the actual VDL segment. Thus, the only solution was to specify an approximate track (line) to represent the VDL route (Fig. 7).
Fig. 6: Shows road networks red and yellow were overlaid onto the georeferenced image. The yellow lines are the existing roads on the topographic map. The red lines are the compiled VDL road networks.
Remarks
Utilised GIS/RS integrated approach provided a scientifically justifiable and generally applicable approach to variable historic site identification, recording, and mapping. The output can be potentially integrated into a planning framework to support resource and environmental planning, management, and decision making. Despite the fact that this report has been able to provide the best possible estimate of the location of the road, we recommend that further ground truthing will be required to verify this conclusion.
Fig. 7: Diagrammatic representation of the reconstructed of the VDL route.
References
- Domenikiotis, C., G. D. Lodwick, and G. L. Wright, 1995. Intelligent Interpretation of SPOT Data for Extraction of a Forest Road Network. Cartography, Vol. 24, No. 2, pp. 47-57.
- Estes, J. E., C. Kenneth, and G. A. Fletcher., 1987. Coordinating Hazardous Waste Management Activities Using Geographical Information Systems. International Journal of Geographical Information Systems, Vol. 1, No. 4, pp. 359-377.
- Forghani, A. and D. Gaughwin., 2000. Identification of a Road Network in an Archaeological Site Using an Integrated GIS and RS Technique. Proceedings of the Second International Conference on Geospatial Information Agriculture and Forestry Lake Buena Vista, FL, USA, 10-12 January 2000, pp. 1-6.
- Forghani, A., 1999. An Expert System Approach for Detection of Roads from Remote Sensing Data. Proceedings of the Joint Workshop of ISPRS Working Groups 1/1, 1/3, and IV/4: Sensors and Mapping from Space 1999, September 27-30, Honover, Germany, pp. 1-6.
- Gaughwin, D., and Forghani, A., 2000. Finding Historic Roads in an Archaeological Site Using a GIS/RS. Proceedings of the 10th Australasian Remote Sensing and Photogrammetry Conference, Adelaide, Australia 21-25 August 2000, pp. 1-9.
- Rollins, B. J., 1988. Henry Hellyer, 1790-1832 Van Dieman’s Land Company Surveyor in His Footsteps. The Australian Surveyor, Vol. 34, No. 2, pp. 110 141.
- Van Cleynenbregel, J., F. Fierens, P. Suetens, and A. Oosterlinck, 1990. Delineating Road Structures on Satellite Imagery by a GIS-Guided Technique. Photogrammetric Engineering and Remote Sensing, Vol. 56, No. 6, pp. 893-898.
- Wara-Aswapatic, P., 1995. An Archaeological Application of Synthetic Aperture Radar (SAR) in Thailand. Geocarto International, Vol. 10, No. 3, pp. 65-69.