On the processing of aerial LiDAR data for supporting enhancement, interpretation and mapping of archaeological features

1,261 views
1,086 views

Published on

On the processing of aerial LiDAR data for supporting enhancement, interpretation and mapping of archaeological features
Rosa Lasaponara - Institute of Methodologies for Environmental Analysis, National Research Council, Italy
Nicola Masini- Archaeological and monumental heritage institute, National Research Council, Italy

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,261
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
63
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

On the processing of aerial LiDAR data for supporting enhancement, interpretation and mapping of archaeological features

  1. 1. On the processing of aerial LiDAR data for supporting enhancement, interpretation and mapping of archaeological features Rosa Lasaponara 1, Nicola Masini2 2 GisSearch Lab. - CNR-IBAM (Institute of Archaeological and Architectural Heritage), Potenza, Italy 1 ARGON Lab. - CNR-IMAA (Institute of Methodologies for Environmental Analysis ), Potenza, Italy
  2. 2. <ul><li>OUTLINE </li></ul><ul><li>Archaeological feature detection: Potential and limits of optical remote sensing </li></ul><ul><li>LiDAR technology </li></ul><ul><li>Data processing </li></ul><ul><li>Post processing </li></ul><ul><li>Study case </li></ul><ul><li>Results and discussion </li></ul>
  3. 3. Magnetic Method GPR Electrical Method LiDAR Aerial prospection Electromagnetic Method Satellite Remote Sensing Ground Truth Ground Remote Sensing How detect Archaeological Features?
  4. 4. <ul><li>Changes of soil constituents </li></ul><ul><li>Variations of moisture content </li></ul><ul><li>Nutrient deficiencies </li></ul><ul><li>Differences in the plant growth </li></ul><ul><li>Crop marks </li></ul><ul><ul><li>differences of height or color in crops which are under stress due to lack of water or deficiencies in other nutrients </li></ul></ul><ul><li>Soil marks : traces of archaeological features visible on bare ground sites </li></ul>Neolithic settlement in Apulia (Souther Italy) <ul><li>Optical remote sensing </li></ul><ul><li>Traditional aerial archaeology </li></ul><ul><li>Multispectral satellite imagery </li></ul><ul><li>Hyperspectral airborne data </li></ul>
  5. 5. <ul><li>Changes of soil constituents </li></ul><ul><li>Variations of moisture content </li></ul><ul><li>Nutrient deficiencies </li></ul><ul><li>Differences in the plant growth </li></ul><ul><li>Crop marks </li></ul><ul><ul><li>differences of height or color in crops which are under stress due to lack of water or deficiencies in other nutrients </li></ul></ul><ul><li>Soil marks : traces of archaeological features visible on bare ground sites </li></ul>Neolithic settlement in Apulia (Souther Italy) <ul><li>Optical remote sensing </li></ul><ul><li>Traditional aerial archaeology </li></ul><ul><li>Multispectral satellite imagery </li></ul><ul><li>Hyperspectral airborne data </li></ul>
  6. 6. AIRBORNE/SATELLITE OPTICAL DATA : LIMITS i) the impossibility of surveying archaeological features of areas covered by dense vegetation; ii) the difficulty in detecting archaeological features related to microrelief ( shadow-marks ) also in case of bare ground surfaces; Limits of optical remotely sensed data How overcome these limits LiDAR Shadow marks : micro-topographic relief variations that can be made visible by shadowing in low sunlight angle conditions.
  7. 7. <ul><li>It provides direct range measurements mapped into 3D point clouds between a laser scanner and earth’s topography. </li></ul><ul><li>The laser scanner, mounted to an aeroplane or helicopter, emits near infrared pulses, at a frequency rate of 30.000 to 100.000 pulses per second, into different directions along the flight path towards the terrain surface. </li></ul><ul><li>Each pulse could be reflected one or more times from objects (ground surface, vegetation, buildings, etc.), whose position is determined by computing </li></ul><ul><ul><li>the time delay between emission and each received echo, </li></ul></ul><ul><ul><li>the angle of the emitted laser beam, </li></ul></ul><ul><ul><li>the position of the scanner (determined using differential global positioning system and an inertial measurement unit). </li></ul></ul>Airborne Laser Scanner (ASL) or LiDAR (Light Detection And Ranging)
  8. 8. <ul><li>conventional scanners or discrete echo scanners delivers only the first and last echo, thus losing many other reflections </li></ul><ul><li>full-waveform (FW) scanners detects the entire echo waveform for each emitted laser beam* . </li></ul>* FW offers improved capabilities especially in areas with complex morphology and/or dense vegetation cover Scanners <ul><li>Scanner employed by CNR </li></ul><ul><li>Scanner Riegel LMS –Q560, full waveform (unlimited echoes) </li></ul><ul><li>Density from 19 pts/m 2 to 150 pt/m 2 </li></ul><ul><li>Scanning frequency = from 70.000 Hz (at 1500 m flight altitude) to 200.000 Hz (at 500 m flight altitude) </li></ul><ul><li>Number of recorded echoes : unlimited; </li></ul><ul><li>Wave length of laser = 1550nm </li></ul><ul><li>Pulse =3.5 ns </li></ul><ul><li>Max flight altitude = 3000 m </li></ul>
  9. 9. Airborne laser scanning (ALS) sensors can penetrate vegetation canopies allowing the underlying terrain elevation to be accurately modeled The most exciting characteristics : Vegetation Filtering!!
  10. 10. <ul><li>Irsi o Yrsum o castrum Ursum close to the Northeastern border of Basilicata with Apulia </li></ul><ul><li>strategic location : the confluence of the Bradano and Basentello riversa </li></ul><ul><li>long human frequentation: iron age, roman period, byzantine age </li></ul><ul><li>12th cent : first documentary source </li></ul><ul><li>1123 : Yrsum depends on Episcopate of the near town of Montepeloso </li></ul><ul><li>1133: Yrsim depends on the monastery of the french order of Chase Die </li></ul><ul><li>13°-14 th cent : inhabitants (550 in 1277 and 5090 in 1320) </li></ul><ul><li>1288 : description of a part of the village from a documentary source (a church, a square, some houses, grain storage) </li></ul><ul><li>1370 : The village is lootedd,the monastery is destroyed. </li></ul><ul><li>Abandonmnent of the village </li></ul>Study case: medieval village of Yrsum
  11. 11. Study case: medieval village of Yrsum Archaeological features detected by optical (aerial/satellite) remote sensing 1) Ditch and cropmarks related to buried walls of the castle (A) 2) Microrelief related to some buried buildings of the medieval village (B) Issues to be addressed : 1) More detailed map of archaeological features 2) Reconstruction of forma urbis 3) Identification of building phases of the medieval village 4) Geomorphological pattern
  12. 12. WORKFLOW DATA PROCESSING <ul><li>Initial setup and data calibration; </li></ul><ul><li>Filtering cloud points; </li></ul><ul><li>Classify cloud points; </li></ul><ul><li>Creating delivery products (DEM, DTM) </li></ul><ul><li>Post processing (shaded DTMs) </li></ul><ul><li>Archaeological interpretation </li></ul>The identification of archaeological features for both bare and densely vegetated areas, needs a DTM with a high accuracy. For this aim, it is crucial to carry out the classification of terrain and off terrain objects by applying adequate filtering methods
  13. 13. DATA FILTERING The identification of archaeological features for both bare and densely vegetated areas, needs a DTM with a high accuracy. For this aim, it is crucial to carry out the classification of terrain and off terrain objects by applying adequate filtering methods (in detail, see Sithole and Vosselman 2004). Filtering methods available slope-based, block-minimum, surface-based clustering / segmentation Assessment of Filtering methods <ul><li>on test sites characterized </li></ul><ul><li>outliers , low or high (such as birds, low-flying aircraft, or errors in the laser range-finder); </li></ul><ul><li>spatial and morphological object complexity ( such as very large or small objects, complex shape ) which typically characterizes a urban setting; </li></ul><ul><li>attached objects spanning the gaps between bare-Earth surfaces ( bridges, natural/artificial ramps, building on slopes etc ..); </li></ul><ul><li>low vegetation on slopes ; </li></ul><ul><li>geomorphologic discontinuities due to steep slopes and sharp ridges. </li></ul>Best results in separating points on a ground surface from other points are obtained by the surface-based methods which assume as discriminant function a parametric surface with a corresponding buffer which defines a region in 3D space where ground points are expected to reside. (Axelsson 2000; Briese & Pfeifer 2001; Elmqvist 2001; Sohn & Dohman 2002; Wack & Wimmer 2002; Sithole & Vosselman, 2004) ( Sithole & Vosselman, 2004 )
  14. 14. CLASSIFICATION <ul><li>Rationale basis : </li></ul><ul><li>Starting from a coarse TIN surface obtained from reference points which are neighborhood minima. </li></ul><ul><li>Densification : new points are added in an iterative way if they meet certain geometric threshold values based </li></ul><ul><ul><ul><li>which prescribe possible deviations from the average topographic surface and builds a triangulated model. </li></ul></ul></ul><ul><li>In every iteration points (from the point cloud) are added to the TIN if they are below data derived thresholds. </li></ul><ul><li>The iterative process ends when no more points are below the threshold. </li></ul>representation of main parameters to construct TIN * Routine of TerraScan (Soininen, 2005) Triangulation Irregular Network (TIN) densification method by Axelsson (2000) <ul><li>threshold parameters </li></ul><ul><li>Iteration angle </li></ul><ul><li>Iteration distance </li></ul><ul><li>Terrain angle </li></ul><ul><li>Max building size </li></ul>
  15. 15. CLASSIFICATION Digital Terrain Model (DTMs) are obtained by the discrimination of on-terrain from off-terrain points (Classification) by using the diverse laser measurements : (i) height; (ii) intensity; (iii) echo width <ul><li>The elimination of outliers points is performed through classification of : </li></ul><ul><li>&quot;low points“ </li></ul><ul><ul><li>single points or groups of points with an height lower than 0,5 m compared the other points within a ray of 5 m </li></ul></ul><ul><li>“ air points. </li></ul><ul><ul><li>points present in the air (i.e. birds, etc..). </li></ul></ul><ul><li>isolated points </li></ul><ul><ul><li>points present in the air not classified as airpoints. </li></ul></ul>
  16. 16. POST PROCESSING Shading procedures to emphasize archaeological features <ul><li>Vizualization of elevation data as shaded relief, by lighting the DTM by an hypothetical light source </li></ul><ul><li>Selection of the direction parameters ( zenith angle z and azimuth angle  ) : on the base of the difference in height and orientation of the microrelief of possible archaeological interest </li></ul>z=60°     <ul><li>Single shading is not the most effective method to visualize and detect microrelief ( If features and/or objects are parallel to the azimuth angle, will not rise a shade ) </li></ul><ul><li>The right approach : observing and comparing DTM scenes shaded by using different angles of lighting. </li></ul><ul><li>In addition the different shaded DTM could processed by using the Principal Components Analysis </li></ul>
  17. 17. Zenith=45°/Azimuth=0° Zenith=45°/Azimuth=90° Zenith=45°/Azimuth=180° Zenith=45°/Azimuth=270° Principal Component Analysis PCA1 PCA3 PCA4 Hill shaded DTM AIRBORNE LASER SCANNING POST PROCESSING: Principal Component Analysis
  18. 18. POST PROCESSING: Convexity and Slope maps Convexity Slope DTM DTM with archaeological interpretation a a' Profile a-a’
  19. 19. MEDIEVAL SITE OF YRSUM
  20. 20. Slope map Both of maps put clearly in evidence the microrelief referable to the layout of buried buildings of Irsi medieval village Profile convexity map
  21. 21. Mapping of archaeological features and reconstruction of “forma urbis” “
  22. 22. Archaeological interpretation Paleo landlside Urban sector extra moenia Square ( platea )? Castle/Motte Niche of detachment B A C Landslide foot Landslide slope
  23. 23. Virtual reconstruction: scenario I (Foundation) Castle Bailey (basse court)
  24. 24. Virtual reconstruction: scenario II (urban expansion) B A Castle Castle Bailey (basse court)
  25. 25. Under the vegetation …..the Historical Landscape! The small dimensions of fielddivisions suggest an intensive farming in this area, likely related to vineward, vegetable gardens and fruit trees which supplied people living in Yrsum Ancient field divisions
  26. 26. <ul><li>CONCLUSIONS </li></ul><ul><li>Archaeological prospection needs an integrated approach of different remote sensing methods </li></ul><ul><li>LiDAR overcomes some limits of optical imagery in </li></ul><ul><ul><li>surveying archaeological structures and features covered by vegetation </li></ul></ul><ul><ul><li>Detecting archaeological features related to microrelief </li></ul></ul><ul><li>c) Post processing (PCA of hill shading procedure, convexity maps etc..) is crucial to fully exploit the potential of LiDAR in archaeology and enhance archaeological features </li></ul><ul><li>d) LiDAR provides archaeological information and digital topographic model for historical virtual reconstruction </li></ul>

×