DTM Chapter 2

DTM Chapter 2:
Method for DTM
Generation from optical and SAR stereo-imagery, digitization of

cartographic source, Hydrographic surveys
Theoretical Background for sampling:
• Sampling for different parts of view:
 Statistical based sapling
- The whole surface is taken as a population thus samples of the surface are taken
 Geometry based sampling
- The surface is divided into certain geometrical figures
 Feature based sampling
- Prominent features of the surface are taken
• Sampling strategies for data acquisition:

 Selective sampling
• Fewer points represent true nature of terrain
 Sampling with one dimension fixed
• Contouring and profiling
 Sampling with two dimension fixed
• Progressive sampling, regular grids, applied in photogrammetry
 Composite sampling
• Combination of selective + photogrammetric sample
Theoretical Background for sampling:
• Attribute of sampled source data
 Distribution of sampled source data
 Density of sampled source data
 Accuracy of sampled source data
• Point :
 Feature specific
 Random
 Surface specific
• Sampling a terrain:
 Feature specific point - true nature of the terrain
 Other points - maintain density
• In photogrammetry - sampling in performed in a stereomodel, which needs a trained
professional , which is selective sampling
Data sources for Digital Terrain Modelling
1. Terrain surface
2. Aerial photographs
3. Satellite imageries
4. Lidar point clouds
5. Cartographic outputs
Data acquisition techniques for DTM
1. DEM from analytical photogrammetry (stereo-plotting)
2. DEM from digital photogrammetry (image matching)
3. DEM from Interferometric SAR
4. DEM from stereo-radargrammetry
5. DEM from stereo-satellite images aka satellite photogrammetry
6. DEM from Airborne Laser Scanning (LiDAR)
7. DEM from UAV photogrammetry
8. DEM from field survey and gps survey
9. DEM from cartographic source
DEM from Digital
Photogrammetry(Image
matching)
Digital Photogrammetry and DEM Generation
Digital photogrammetry
• Measurements in photograph in a digital environment for acquiring
the geo information.
• Photographs may be taken with either analogue or digital camera.
• The photographs taken with analogue camera are converted in digital
system by scanning.
• High quality professional scanners are used for scanning.
Digital photogrammetric workstations (DPW)
• Photographs are processed by DPW.
• DPW is a set of software/hardware used for processing the
photographs and extraction of different types of geodata.
• One example LPS.
• Different photogrammetric processes (orientation, AT and bundle
block adjustment) can be carried out and several products (DTM,
orthophoto) can be generated.
Orientation
• Interior orientation – Camera Calibration data as input.
• Exterior orientation
• Relative orientation
• Absolute orientation
Aerial triangulation
• Aerial triangulation is the process of measuring points in images or
stereo models such that the orientations of these photographs can be
reconstructed
• Measurement of corresponding points in overlapping images
• Extension (densification) of control points in images
• Measurement of ground control points in images
• Estimation of orientation parameters of all images (block adjustment)
Independent model and bundle block
Mathematical model: independent
Mathematical model: bundle block
DTM generation from Digital
Photogrammetry
Left image Right image
Orientation (interior, Exterior, AT)
Stereo model
Image matching (matching technique, parameters)
DSM
Editing/Filtering
Final DEM
Image Matching
• Process of finding matching points in the corresponding stereo image

pairs.
• The conjugate point in the corresponding image is determined by
comparing the similarity measure, e.g. brightness
• Similarity is measured by correlation coefficient
• Correlation coefficient varies from 0 to 1
• 0 = no match at all
• 1 = 100 % matched
Image matching
Main steps in image matching
• Select a matching point in one image,
• Find its conjugate point in the other image
• Compute 3D position of the matched point in object space
Types of matching: area based matching
Area based matching
Difference
• Cross correlation
• Only considers the gray value intensity for comparing the pixels in template and search window
• Deals with translation only
• Do not deal with the rotational aspect of the pixel
• Do not regards the geometry
• Less accurate
• Least square matching
• considers the geometric transformation in the corresponding images
• The template looks for gray value change and translation as well as the rotational parameters and scale
change between the image patch
• based on the principle of least square
• determines the accurate estimation of affine parameters and corrects the perspective geometry of the
images
• More accurate
Feature based matching
• Distinct features like points, edges and patches are independently
extracted in all images
• They are matched with the corresponding feature entities of the
other image
• Feature attributes and consistency in location of corresponding
feature is compared
Feature based matching
• Least square feature based matching method identifies distinct points
as features to be matched in images
• these points are extracted in corresponding images
• initial correspondence is established by calculating correlation
coefficient
• affine transformation parameters are calculated for robust least
square adjustment
Steps in DTM generation
• Mass point generation by image matching
• 3D Ground coordinates are determined
• Interpolation between the mass points for dense points
• DSM
• Filtering process are applied for reducing the height point to the
ground
Parameters influencing the matching results
• Matching technique/algorithm
• Search window
• Target window
• Correlation coefficient
• DTM Cell size
• Interpolation methods
Constraints
• To make the search process faster and efficient, different constraints
may be applied for image matching
• Image pyramid
• Epipolar geometry
Image pyramid
Reducing complexity/increasing reliability

Matching in image pyramid
Epipolar geometry
Epipolar geometry
Epipolar geometry
Epipolar geometry
Problems in image matching
Activity:
• Reading: DTM Principles and Methodologies: Chapter 2 & 3
Image matching and it's principles in photogrammetry
Photogrammetry Mathematics
Bathymetry and LiDAR data in DTM Generation
Use of Satellite Imagery for DEM Extraction
• Assignment:
1. State different image matching methods and their types. Explain in
brief Correlation and Least Square method of image matching.
2. Summarize the article “Use of Satellite Imagery for DEM Extraction”.
3. Describe the workflow for the generation of DTM form ‘Hydrographic
Surveys’.

DTM Chapter 2

Uploaded by

DTM Chapter 2

Uploaded by

DTM Chapter 2:

Method for DTM

Generation from optical and SAR stereo-imagery, digitization of

• Sampling strategies for data acquisition:

Orientation (interior, Exterior, AT)

Image matching (matching technique, parameters)

• Process of finding matching points in the corresponding stereo image

Reducing complexity/increasing reliability

You might also like