| GISdevelopment.net ---> Technology ---> Remote Sensing |
Airborne Laser Scanning – Cost effective spatial data
David Turton
Manager – Aam Geoscan
Email: D.Turton@aamgeoscan.com.au
David Jonas
Production Manager, Aam Geoscan
Email: D.Jonas@aamgeoscan.com.au
AAM GEOSCAN, A Division of AAM Surveys Pty Ltd.
11 Wicklow Street, Kangaroo Point Qld. 4169 Australia
Tel: +61738911033, Fax +61738911050
Abstract
Airborne Laser Scanning (ALS), also referred to as Light Detection and Ranging (LiDAR), is a remote sensing technology which enables large areas to be surveyed cost effectively using a fixed or rotory wing platform.
The ALS instrument emits a stream of discrete points, up to 70,000 every second, which are reflected from objects the aircraft flies over.
Post Processing of the data provides co-ordinates with height accuracies ranging from 0.1 metres to 0.5 metres and horizontal accuracies ranging from 0.3 metres to 1.5 metres, depending on flying height, GPS geometry and other project variables. Points reflected from the ground are separated from those above the ground. A closely spaced array of co-ordinated points can be obtained if required with point spacing ranging from 0.2 to 4.0 metres.
This paper will briefly outline the technology before discussing two specific ALS applications in forestry and flood plain definition.
Because ALS generates a large volume of data special attention must be given to data presentation. Some presentation options will be outlined.
Planning an ALS project requires consideration of issues not encountered when planning data gathering using other methods. A selection of these issues will be discussed as will the causes of error in an ALS survey.
Technical Description of ALS
An ALS system consists of three components:
The system components are shown diagrammatically below.
Depending on the ALS manufacturer at least two recordings of each of the laser points emitted are stored by the system. These include the first return received and the last return received. The first return may be from the top of a tree while the last may represent the ground beneath the tree. Coordinates are computed for both the first and last returns.
First Return: Measures the distance to the first object encountered … in this case, the tree foliage
Last Return: Measures the distance to the last object … in this case, the ground
By acquiring first and last return data simultaneously, it is possible to measure both the tree-heights and the topography of the ground beneath in a single pass
Information from the three ALS components is combined with ground base station GPS data to produce XYZ co-ordinates of the reflected points. These are then separated automatically into points reflected from the ground and those reflected from above ground features. The profile below shows ALS points reflected from power lines, nearby vegetation and the ground.
The return signal strength of each of the first and last returns is also recorded. This information can be used with a grayscale to create geo-referenced images. The image below shows ground and non ground returns.
Airborne Laser Scanning is an active remote sensing technology. The system emits laser signals and as such can be operated at any time during the day or night.
Applications
ALS in Forestry
Because of its ability pass between tree branches to record both ground and non ground features, ALS is particularly suited to forestry applications. Applications include the acquisition of data to compute average tree heights, the use of terrain data to plan the location of roads to be used in timber harvesting and the determination of drainage locations for the design of retention corridors. The diagrams below show contours derived from ground data, which clearly indicate drainage locations, together with a profile showing ground and tree canopy.
Flood Plain Definition
Because of the ability of Airborne Laser Scanning to observe large terrain areas accurately and quickly, flood plain mapping is a frequent application.
As an example, an area 30 km by 10 km can be captured, using a fixed wing survey platform, in about four hours. The vertical accuracy of the data would be around 15 cm and the average point spacing would be 1.5 metres. About 153,000,000 data points would be recorded, providing a detailed description of the terrain and objects on it.
Issues of Data Presentation
ALS surveys generate large amounts of data. Because each data point is randomly located, not all points are required in order to define the terrain and objects on or above it. In order to remove those points which are not required thinning techniques have been developed using nearest neighbour algorithms to compare adjacent points and remove any that do not add to the definition of the required object. Depending on the parameter used this can reduce the size of the data set significantly: a 40% reduction in file size is not uncommon.
Other presentation options include the triangulation of the data followed by contour interpolation. Profiles, cross sections, flythrough and hypsometric plots are other options. This latter option is particularly useful where the terrain is generally flat and thus the interpretation of contours is rather difficult.
Hypsometric plots assign colours to height ranges. The results greatly increase the ease with which the terrain change can be interpreted. In the example below a height range of 0.50metres has been used.
Project Planning
Issues which need to be considered when planning an ALS project include:
Sources of error in ALS Surveys
The principal sources of error in ALS surveys can be attributed to:
Absolute accuracy is important where ALS co-ordinated data has to be integrated with existing data. Relative accuracy will often be more significant in more localised survey projects, such as the determination of tree heights.
The relative accuracy of ALS points is generally in the range of 3 cm-5 cm.
The absolute accuracy is less, dependant on the four components above, viz
Indicative Costings
Costs of an ALS survey depend very much on where the survey is, (the cost to mobilise to the site), the time taken to acquire the data and the extent of data processing required.
An area 10 km by 4 km could be surveyed in one sortie and would cost approximately AUS$40,000.00, including basic processing, to achieve a height accuracy of 0.15m on clear ground. This cost would exclude the cost to mobilise to the site and also costs for a single ground base station, ground test points and any advanced data processing.
By comparison, to map the same area by photogrammetry to the same accuracy would require some 80 frames of photography and 40 ground control points!
Conclusion
Airborne Laser Scanning is a cost effective method of acquiring spatial data. In common with other acquisition methods it has specific positive aspects but also some negative ones! A knowledge of surveying issues is essential if client expectations are to be met consistently. ALS is not a “black box” technology!
The problem to be solved must be properly understood before an appropriate acquisition solution can be designed. Airborne Laser Scanning may be the most appropriate solution when some or all of the following circumstances prevail.
David Turton
Manager – Aam Geoscan
Email: D.Turton@aamgeoscan.com.au
David Jonas
Production Manager, Aam Geoscan
Email: D.Jonas@aamgeoscan.com.au
AAM GEOSCAN, A Division of AAM Surveys Pty Ltd.
11 Wicklow Street, Kangaroo Point Qld. 4169 Australia
Tel: +61738911033, Fax +61738911050
Abstract
Airborne Laser Scanning (ALS), also referred to as Light Detection and Ranging (LiDAR), is a remote sensing technology which enables large areas to be surveyed cost effectively using a fixed or rotory wing platform.
The ALS instrument emits a stream of discrete points, up to 70,000 every second, which are reflected from objects the aircraft flies over.
Post Processing of the data provides co-ordinates with height accuracies ranging from 0.1 metres to 0.5 metres and horizontal accuracies ranging from 0.3 metres to 1.5 metres, depending on flying height, GPS geometry and other project variables. Points reflected from the ground are separated from those above the ground. A closely spaced array of co-ordinated points can be obtained if required with point spacing ranging from 0.2 to 4.0 metres.
This paper will briefly outline the technology before discussing two specific ALS applications in forestry and flood plain definition.
Because ALS generates a large volume of data special attention must be given to data presentation. Some presentation options will be outlined.
Planning an ALS project requires consideration of issues not encountered when planning data gathering using other methods. A selection of these issues will be discussed as will the causes of error in an ALS survey.
Technical Description of ALS
An ALS system consists of three components:
- A Global Positioning System (GPS) to position the moving aircraft in space.
- An Inertial Measurement Unit (IMU) to record the aircraft attitude and acceleration.
- A laser light source which directs a stream of discrete laser points towards the ground at approximately ninety degrees to the line of flight. The time taken for each of these discrete points to return to the aircraft is accurately recorded.
The system components are shown diagrammatically below.
Depending on the ALS manufacturer at least two recordings of each of the laser points emitted are stored by the system. These include the first return received and the last return received. The first return may be from the top of a tree while the last may represent the ground beneath the tree. Coordinates are computed for both the first and last returns.
First Return: Measures the distance to the first object encountered … in this case, the tree foliage
Last Return: Measures the distance to the last object … in this case, the ground
By acquiring first and last return data simultaneously, it is possible to measure both the tree-heights and the topography of the ground beneath in a single pass
Information from the three ALS components is combined with ground base station GPS data to produce XYZ co-ordinates of the reflected points. These are then separated automatically into points reflected from the ground and those reflected from above ground features. The profile below shows ALS points reflected from power lines, nearby vegetation and the ground.
The return signal strength of each of the first and last returns is also recorded. This information can be used with a grayscale to create geo-referenced images. The image below shows ground and non ground returns.
Airborne Laser Scanning is an active remote sensing technology. The system emits laser signals and as such can be operated at any time during the day or night.
Applications
ALS in Forestry
Because of its ability pass between tree branches to record both ground and non ground features, ALS is particularly suited to forestry applications. Applications include the acquisition of data to compute average tree heights, the use of terrain data to plan the location of roads to be used in timber harvesting and the determination of drainage locations for the design of retention corridors. The diagrams below show contours derived from ground data, which clearly indicate drainage locations, together with a profile showing ground and tree canopy.
Flood Plain Definition
Because of the ability of Airborne Laser Scanning to observe large terrain areas accurately and quickly, flood plain mapping is a frequent application.
As an example, an area 30 km by 10 km can be captured, using a fixed wing survey platform, in about four hours. The vertical accuracy of the data would be around 15 cm and the average point spacing would be 1.5 metres. About 153,000,000 data points would be recorded, providing a detailed description of the terrain and objects on it.
Issues of Data Presentation
ALS surveys generate large amounts of data. Because each data point is randomly located, not all points are required in order to define the terrain and objects on or above it. In order to remove those points which are not required thinning techniques have been developed using nearest neighbour algorithms to compare adjacent points and remove any that do not add to the definition of the required object. Depending on the parameter used this can reduce the size of the data set significantly: a 40% reduction in file size is not uncommon.
Other presentation options include the triangulation of the data followed by contour interpolation. Profiles, cross sections, flythrough and hypsometric plots are other options. This latter option is particularly useful where the terrain is generally flat and thus the interpretation of contours is rather difficult.
Hypsometric plots assign colours to height ranges. The results greatly increase the ease with which the terrain change can be interpreted. In the example below a height range of 0.50metres has been used.
Project Planning
Issues which need to be considered when planning an ALS project include:
- What problem are you trying to solve?
Be clear about the main reason for the survey. Planning for a project primarily required to determine tree heights will differ from that of a project primarily designed to define the ground beneath trees.
- What is the general ground shape in the survey area?
Rugged terrain with changing gradients will need points close together than generally flat terrain.
- Any problem with dense vegetation cover?
Narrow swathes may be needed to increase ALS penetration in forested areas.
- Does the survey area contain hilly terrain?
Flight lines must be adjusted to accommodate changing ground heights.
- Will air turbulence be excessive?
Humid summer weather, particularly in the tropics, can cause severe air turbulence at low altitudes which effects both the coverage and accuracy of the survey.
- Is the survey area close to other air operation areas such as an airport?
Delays will result if the work is conducted at times of heavy air traffic movements such as may occur around an airport.
- Is there a suitable existing GPS base station in or near the survey area?
If no suitable permanent station is available a ground survey team will need to establish and run a GPS station for the duration of the survey.
- How will the ALS data be validated?
As with any remote data acquisition technique, including photogrammetry, ALS surveys may contain a systematic error. The existence of such an error can be determined and then removed by comparing the ALS data with some higher accuracy general points within the area covered by ALS.
- Is the survey to take place in a remote area?
Surveys in remote areas require additional planning to identify an appropriate base airport, arrange overnight accommodation and perhaps arrange an aircraft fuel supply.
Sources of error in ALS Surveys
The principal sources of error in ALS surveys can be attributed to:
- GPS issues
- IMU issues
- Timing issues
- Data Classification Errors
Absolute accuracy is important where ALS co-ordinated data has to be integrated with existing data. Relative accuracy will often be more significant in more localised survey projects, such as the determination of tree heights.
The relative accuracy of ALS points is generally in the range of 3 cm-5 cm.
The absolute accuracy is less, dependant on the four components above, viz
- GPS; the accuracy with which the moving aerial platform can be positioned using the Global Positioning System satellite configuration.
- IMU: the accuracy with which the dislevellment and acceleration of the aircraft can be measured using the Inertial Measurement Unit.
- Timing; the accuracy with which the time taken to transmit and receive a single ALS beam can be measured.
- Data classification; how well ground and non ground ALS returns can be separated.
Indicative Costings
Costs of an ALS survey depend very much on where the survey is, (the cost to mobilise to the site), the time taken to acquire the data and the extent of data processing required.
An area 10 km by 4 km could be surveyed in one sortie and would cost approximately AUS$40,000.00, including basic processing, to achieve a height accuracy of 0.15m on clear ground. This cost would exclude the cost to mobilise to the site and also costs for a single ground base station, ground test points and any advanced data processing.
By comparison, to map the same area by photogrammetry to the same accuracy would require some 80 frames of photography and 40 ground control points!
Conclusion
Airborne Laser Scanning is a cost effective method of acquiring spatial data. In common with other acquisition methods it has specific positive aspects but also some negative ones! A knowledge of surveying issues is essential if client expectations are to be met consistently. ALS is not a “black box” technology!
The problem to be solved must be properly understood before an appropriate acquisition solution can be designed. Airborne Laser Scanning may be the most appropriate solution when some or all of the following circumstances prevail.
- Site access is difficult.
- Ground and non ground spatial data is required.
- The survey area is large or remote.
- Closely spaced co-ordinated points are required to depict the surface.
- Data is required quickly.
- Weather conditions or time of year make aerial photography difficult.