TRIANET

Remote Sensing

Home
News
Methods
Analysis
Regions
Nature - Culture - Society
Conatct
Links


Image Interpretation

Image interpretation deals with the extraction of reliable information from displayed images using the human eye-brain mechanism. The image can be displayed on TV screen or on hard copy printed matter. The reliability of the extracted information is tested using ground truth samples or ground surveys. Images for interpretation, are processed to increase the contrast between the information to be extracted and its background. The person which does the interpretation must be well trained and have enough experience.

The following rules are suggested for interpretation:

  1. Make a good quality of the displayed image with increased contrast
  2. Identify objects from their shape
  3. Identify objects from their relative size and pattern
  4. Identify objects from their tone. Different tones are created from different surface textures and usually soft texture results to a light tone while hard texture results to a dark tone. Errors may occur because of the effect of color and because of the effect of relation of the light source to the camera (specular reflection).
  5. Identify objects from their shadow. Errors may occur because of cloud shadows, or, because it is hard to distinguish the object from its shadow.
  6. Identify objects from their relation to the surroundings.

 

4.1 Interpretation of vertical aerial photographs

Vertical aerial photography is taken from the airplane with camera axis vertical. Photographs are taken in adjacent line strips, which form a block to cover an area. The aerial photographs along a strip have over 60% overlap and this allows stereoviewing through a stereoscope.

hatz05.gif (4849 Byte)

 

Stereo images obey to the following rules:

  1. Camera axes must be parallel to each other and perpendicular to the line O1 – O2 which is called base B.
  2. There must be an overlap E. Only images in the overlaped area can be viewed in stereo.
  3. The object must be static if one camera in two positions is used, or, the object could be dynamic if two cameras synchronized to obtain simultaneous exposures from fixed positions are used.
  4. Better geometric quality of stereo image is obtained if the ratio B/H is close to one, if it is near zero there is no stereo photography.

 

Assuming that H is the altitude of the airplane over the MSL and h is the elevation of the ground above MSL, then the scale S of the aerial photograph at the level of the ground is given by the formula:

hatz06.gif (1060 Byte)

Where f is the principal distance of the camera (it is the focal length for aerial cameras).

Considering a ground point P which provides images p’ on the left photograph and p" on the right photograph, then the elevation h of point P can be computed from the formula:

hatz07.gif (1095 Byte)

Where p is called parallax and it is the distance between p’ and p" if we lay both photographs over one camera frame and match their corresponding fiducial marks.

The formulas given above help to compute horizontal and vertical coordinates of points.

Stereo aerial photographs are viewed using a stereoscope which helps identification of objects in the three dimensional space. The preparation of aerial photographs for stereo viewing is done as follows:

hatz08.gif (4374 Byte)

 

  1. Join the opposite fiducial marks as shown above, to determine the principal points PPL , PPR for the left and right photograph. The fiducial marks establish also a coordinate system x, y in each photograph. It is suggested the x – axis to be oriented along the flight line of the strip and the shadows to fall towards the observer.
  2. Identify the image of the PPR on the left photograph and determine its conjugate CPPR. Identify the image of the PPL on the right photograph and determine its conjugate CPPL.
  3. Draw the flight line (PPL – CPPR ) on the left photograph and the flight line (PPR – CPPL ) on the right photograph.
  4. Tie the left photograph on the table using a masking tape.
  5. Place the right photograph at a distance 3 – 5 cm from the left photograph and move it so that the flight line (PPL – CPPR ) - (PPR – CPPL ) to form a unique straight line then tie the right photograph using a masking tape.
  6. View the oriented stereopair of photographs by aligning the stereoscope so that the eye base to be parallel to the flight line.

 

Experienced personnel who are trained for this purpose perform the interpretation of aerial photographs. Part of the training consists of detail study of complete library of stereograms of known subjects. Such stereograms are stereophotographs of a well-known object and they are obtained either by cutting pieces from aerial photographs or by taking a stereopair of photographs directly from the object by a hand held camera. Stereograms usually have dimensions 10x15 cm so that to be viewed directly by a pocket stereoscope and they are numbered and organized to form a library.

 

4.2 Interpretation of digital images

Digital images are processed by the software so that to create a good quality screen display, for interpretation purposes. Such processes involve:

  1. Geometric corrections to match the image with a map projection system.
  2. radiometric enhancement (Histogram equalization, Histogram stretching)
  3. Filtering. Filtering can be used for noise removal or to strengthen certain features of the image such as edge enhancement.
  4. Display the image using pseudocolors from a properly selected palette to increase the contrast between the information to be identified and its background.

1. Satellite Introduction to the Remote Sensing System
2. Image formation and viewing systems
3. Digital images and processing systems
4. Image interpretation
5. Image classification
6. References

previoushomenext

dot_clr.gif (46 Byte) Methods User Guide GIS - Geographic Information System Remote Sensing DTM - Digital Terrian Model
 

© 1999 TRIANET, Program of the European Union Socrates-Comenius
Last update on 04.05.1999 by Markus Zapke-Gründemann