Image Acquisition and Overlap

Aerial images

For details on image acquisition for Aerial images, please see the Aerial Nadir Image page (also recommended for Oblique imagery).

General Requirements

Every point on the surface should be observed by at least 3 images (5 are recommended). Furthermore, images should be taken at a similar image scale and thus similar distance to the object.

Image capture

Depending on the type of camera and the type of object/scene, the following acquisition patterns result from given requirements:

  • Object:          circular acquisition around the object 
  • Flat surface:  regular grid acquisition
  • Corridor:        panoramic acquisitions in corridor direction

Through the need of observing every point by at least 5 (recommended) images, sufficient image overlap should be ensured.
For further remarks and recommendations on image capturing, please refer to this paper.

In order to achieve good image quality - e.g. selecting optimal exposure settings, please refer to this page here.


Within SURE, the following  Scenario Presets are available:

  1. Default             - for all irregular and unknown image distributions
  2. Aerial Oblique  - for regular airborne maltese cross configuration acquisitions (regular strips)
  3. Aerial Nadir      - for regular airborne grid acquisition in Nadir configuration (regular strips)

For the latter two scenarios, the regular structure is exploited in order to identify a minimum set of stereo models. 

Please find further details on regular Aerial Nadir Image Overlap on the subsequent article.

The Image Overlap is necessary to derive depth information from images based on the parallax effect of different views (the stereoscopic effect). Thus, at least two images are necessary to recover depth information using stereo. For each stereo pair, the corresponding pixels between the images are estimated using Dense Image Matching. Within SURE, this is carried out for many stereo models combined with a point cloud generation across multiple stereo models (Multi-Stereo Triangulation). The final surface quality is thus strongly influenced by the image overlap and image similarity, which are the base for depth extraction. For further reading, please refer to this paper about the SURE method.