1 of 12

Module 6a: Mapping & Data Processing

Lesson 1: Photogrammetry

AIFS Drone Curriculum Package

2 of 12

At the end of this presentation, students will be able to:

  • Describe the principles of photogrammetry
  • Review mission planning terminologies: ground sample distance (GSD), ground control, overlap
  • Understand the photogrammetry workflow to obtain drone data products

​

​

​

3 of 12

Introduction to Photogrammetry

  • Photogrammetry
    • The science of making measurements from photographs
    • Pioneered in 1893 by C.B. Adams who used two balloons and two cameras to create overlapping photographs to measure ground features
  • Stereophotogrammetry (modern data analysis)
    • Estimating 3D coordinates of points on an object by using measurements made in two or more overlapping photographic images

4 of 12

Parallax

  • Displacement in the position of an object caused by a shift in observation
  • Implementation: Track features seen in two images captured at different positions and obtain displacement

​

5 of 12

Modern Photogrammetry & Intro to SfM

  • If an object has been viewed from multiple viewpoints in many overlapping images, we can generate massive numbers of tie points (points of reference)
  • Through Structure from Motion (computer vision technology), we can construct the 3D shape of an object using these tie points

6 of 12

Structure for Motion

  • Solves for 3D location of all features in overlapping images
  • Requires massive number of tie points to evaluate 3D location, requiring high image overlap

​

7 of 12

Role of Mission Planning

Photogrammetry relies on image overlap and tie points – crucial to acquire imagery with the right characteristics

  • Considerations:
    • Ground Sample Distance (GSD): Distance between 2 consecutive pixel centers measured on the ground
      • Depends on flying height and a camera’s focal length and pixel size
      • GSD = (H (flying height) * px (camera pixel dimensions)) / f (camera lens focal length)
    • Image Overlap: By capturing the same features in multiple images, we can generate massive number of tie points
      • Sidelap: overlap across flight path
      • Frontlap: overlap along flight path

​

​

8 of 12

Data Products Obtained from Overlapping Images

  • Points Cloud
  • Textured Mesh
  • DTM (Digital Terrain Model)
  • DSM (Digital Surface Model)
  • Ortho-mosaic
  • NDVI and other vegetation indexes

9 of 12

Steps of Photogrammetry

Image Collection

Camera information

Tie Point Matching

Orthorectification

Color balance & seam matching

Orthomosaic

Vegetation Layers

Point densification

Point Cloud

Textured Mesh

Classify: Ground / Above Ground

DSM

DTM

10 of 12

Point Cloud & Textured Mesh

  • Point Cloud: x,y,z position and position of each point
  • Mesh: representation of the shape of the model with vertices, edges, and faces
  • Textured Mesh: texture and color of images on mesh

​

Textured Mesh

© Google AR/VR

© Westphalian University

11 of 12

DSM & DTM

  • DTM (Digital Terrain Model): Bare-earth, topographic model
  • DSM (Digital Surface Model): Top of surfaces like buildings and vegetation

© Plex Earth

12 of 12

Ortho-mosaic & Vegetation Indexes

  • Orthorectification: Removal of distortions from DSM (camera tilt, viewpoint) produces a measurable topographic model
  • Orthorectification + color balancing help create ortho-mosaics and vegetation indexes:
    • NDVI (normal difference vegetation index): analyze targets for green health
    • VARI (visible atmospheric resistant index): emphasizing vegetation by removing light and atmospheric distortions

​

© Atom Aviation