Enhancing Your Photogrammetry Skills:
Techniques for Successful Data Collection

Welcome back, fellow photogrammetry enthusiasts! In our previous blog post, we explored the transformative capabilities of photogrammetry, which turns ordinary images into extraordinary 3D models. Today, let’s delve into the exciting methods of data collection in photogrammetry.

Photogrammetry is a powerful technology that extracts valuable information from images, particularly in the field of geospatial data collection. It opens up a whole new dimension of possibilities. In this blog post, we will explore the methods and techniques professionals use to collect data using photogrammetry. From ground-based image capture to aerial platforms, we’ll discuss optimal shooting techniques, camera settings, and the importance of image overlap and coverage.

Join us on this fascinating journey through the world of geospatial data collection with photogrammetry. Together, we’ll uncover the power and versatility of these approaches and explore their real-world applications.

Don’t forget to subscribe and share your thoughts and experiences in the comments below. Let’s embark on this geospatial adventure together!


When it comes to collecting geospatial data through photogrammetry, there are several methods that utilize different technologies or tools, depending on the desired data outputs and project objectives.

  1. Aerial Photography:
    • Traditional Aerial Photography: Involves capturing images from aircraft equipped with high-resolution cameras.
    • Unmanned Aerial Vehicles (UAVs) or Drones: Small remotely operated aircraft equipped with cameras for capturing aerial images.
  2. Satellite Imagery:
    • Satellite-Based Photogrammetry: Utilizes imagery captured by satellites orbiting the Earth to extract geospatial data.
  3. Terrestrial Photogrammetry:
    • Ground-Based Photography: Involves capturing images from the ground using cameras or mobile devices.
    • Close-Range Photogrammetry: Utilizes calibrated cameras and targets placed on objects of interest at close proximity to capture detailed 3D data.
  4. LiDAR (Light Detection and Ranging):
    • Airborne LiDAR: Uses laser sensors mounted on aircraft to measure distances and generate precise elevation data.
    • Terrestrial LiDAR: Involves stationary or mobile laser scanners for capturing detailed 3D information of objects or environments.
  5. Structure from Motion (SfM):
    • SfM Photogrammetry: Uses overlapping images captured from multiple angles to reconstruct 3D geometry and extract measurements.
  6. Mobile Mapping Systems:
    • Vehicle-Based Mobile Mapping: Combines sensors like cameras, LiDAR, and GPS on a moving platform to capture data along road networks.
    • Backpack Mobile Mapping: Involves wearable systems to collect data in pedestrian areas or areas inaccessible by vehicles.
  7. Remote Sensing:
    • Hyperspectral Imaging: Captures images across numerous narrow bands of the electromagnetic spectrum, enabling detailed spectral analysis.
    • Thermal Imaging: Utilizes infrared cameras to measure temperature variations and identify thermal anomalies.
  8. Augmented Reality (AR) and Virtual Reality (VR):
    • AR/VR Photogrammetry: Integrates photogrammetry techniques with AR or VR technologies to create immersive virtual environments.
Common Techniques for Capturing Images

When capturing images for photogrammetry, the technique for image acquisition can vary depending on the specific application and the desired outcome. Here are some common techniques for capturing images on the ground or from the air:

Ground-Based Photogrammetry:
  1. Stationary Camera: Place the camera on a tripod or stable surface and capture images by manually triggering the camera or using a remote shutter release. This technique is suitable for capturing static objects or small areas.
  2. Walking or Handheld: Move around the object or scene while capturing images at regular intervals. This technique is useful for capturing larger scenes or objects with complex geometry.
Aerial Photogrammetry:
  1. Drone-Based Photogrammetry: Use unmanned aerial vehicles (UAVs) or drones equipped with cameras to capture images from an elevated perspective. Drones offer flexibility in terms of flight paths, altitude, and camera angles. They are commonly used for mapping large areas, aerial surveys, and 3D modeling of landscapes or buildings.
  2. Manned Aircraft: In some cases, manned aircraft or helicopters equipped with specialized cameras can be used for aerial photogrammetry. This approach is typically employed for larger-scale projects that require high-resolution imagery or cover extensive areas.

Info source from TOPS


When it comes to sourcing photogrammetry data, there are several avenues you can explore. Consider the following options:

  1. Capturing your own data: One option is to capture your own photogrammetry data by taking photographs of the subject or area you want to reconstruct. You can use a camera, drone, or other imaging devices to capture the necessary images. This approach gives you control over the data collection process and allows you to tailor it to your specific project.
  2. Online repositories: There are online repositories and platforms where you can find publicly available photogrammetry datasets. Websites like Sketchfab, OpenTopography, and NASA’s Earth Observing System Data and Information System (EOSDIS) provide access to various types of photogrammetry data, including 3D models, point clouds, and orthophotos. These datasets are often contributed by individuals, organizations, or research institutions for sharing and collaborative purposes.
  3. Commercial providers: Several companies specialize in collecting and providing photogrammetry data for specific industries or applications. These providers may offer aerial or satellite imagery, LiDAR data, or specialized photogrammetric datasets. Examples of commercial providers include DigitalGlobe, Airbus Defence and Space, and DroneDeploy.
  4. Collaborations and partnerships: Depending on your project, you may be able to collaborate with research institutions, universities, or government agencies that conduct photogrammetry research or have access to relevant datasets. Establishing partnerships or reaching out to experts in the field can potentially grant you access to valuable photogrammetry data.


What Data Is Needed?

The data required for photogrammetry depends on the specific project and the desired level of detail and accuracy. Generally, the more data you have, the better the results will be. Here are the key types of data needed for the photogrammetry process:

  1. Images: A set of high-quality images is crucial for photogrammetry. The number of images required depends on factors like subject complexity, desired detail level, and reconstruction method. For a basic 3D reconstruction, a minimum of 20-30 well-overlapping images is recommended. Additional images from different angles can enhance accuracy and robustness. Larger or complex scenes may require hundreds or even thousands of images.
  2. Image Metadata: Having metadata associated with each image, such as camera parameters (e.g., focal length, sensor size) and GPS coordinates, aids in aligning and registering the images accurately in a real-world coordinate system.
  3. Ground Control Points (GCPs): GCPs are reference markers with known coordinates placed within the scene. They enhance the accuracy of photogrammetric reconstruction by providing control and spatial reference. Typically, a minimum of three well-distributed GCPs is recommended, and more can be used for increased accuracy.
  4. Calibration Data: The availability of calibration data from the camera used for image capture is beneficial. It includes intrinsic camera parameters (e.g., lens distortion) and extrinsic parameters (camera position and orientation). Calibration can be done using specialized targets or known parameters provided by the camera manufacturer.
How Much Data Is Needed?

To ensure accurate processing, each point on your object should be visible in at least two photos. Generally, you’ll need around 20 to 250 photos to capture an object well. For a basic 3D model, a minimum of 20 to 30 well-overlapping images is recommended, providing a decent level of accuracy and detail. However, the required number of images can vary based on factors like subject complexity, desired detail level, and the software or algorithm used.

While it’s possible to create a basic 3D model with fewer than 20 images, having more images usually leads to better outcomes. Additional images offer more data points and better coverage, enhancing accuracy and robustness. Taking images from different angles and viewpoints further improves the overall quality and completeness of the 3D model.

Image (1) shows a better 3D model compared to image (2) due to higher number of overlapping images. Image (1) has less distortion and provides greater details of the model compared to image (2).

Orthographic image and oblique image: requirement for 3D Modeling

The number of images needed for 3D modeling depends on whether you’re using orthographic (top-down) or oblique (captured at an angle) imagery. Oblique images generally provide more information and coverage, enabling more accurate and detailed 3D reconstructions.

Flight plans for orthographic images (1), oblique images (2), and an optional set of oblique images at a higher altitude (3) (Skycatch)

Orthographic imagery, captured from a top-down perspective, lacks depth information. To compensate for this, a greater number of well-overlapping images is required, capturing the subject from different angles and viewpoints to reconstruct the missing depth information.

Oblique imagery, captured at an angle, naturally contains more depth information. Consequently, a smaller number of well-overlapping oblique images can be sufficient for creating a 3D model compared to orthographic imagery. However, using a higher number of oblique images still improves accuracy and detail in the final 3D reconstruction.

In summary, a minimum of 20 to 30 well-overlapping images is generally suitable for basic 3D reconstruction with both orthographic and oblique imagery. Oblique images require fewer images due to their richer depth of information. Nevertheless, using more images, regardless of the type, enhances the quality and accuracy of the resulting 3D model.


Sign up to receive email updates on new product announcement, case study, special promotion, sales and more.

Copyright © 2023           Designed and developed by OFO Tech