Restoration


See how to execute an Image Restoration in SPRING.

The Restoration is a radiometric correction technique where the purpose is to correct the distortions inserted by the optical sensor in the generation process of digital images.

One can say that the digital image is a blurred scene copy, given that the details in the scene are smoothed because of the sensor limitations.

The image restoring idea is to reduce this blurring effect, and thus get an enhanced image.

The correction is performed by a linear filter. The restoration filter weights are obtained from the sensor characteristics, and not in an empirical way as it is performed by the traditional enhancing filters. In this case, the filter is specific for each type of sensor and spectral band.

This processing type is recommended to be performed over the original image without any other processing type, such as enhancing and filtering, that change the image radiometric characteristics. Notice also that it is not possible to process a resampled image, once the image radiometric and spatial characteristics were changed.

The access to the Restoration function, in SPRING main menu, is only possible when there are images in a project without cartographic projection, thus guarantees that the same will not be resampled.

NOTE: After processing the desired bands, it can export the same (RAW or TIFF format), convert to GRIB in the " Impima" module (with a new resolution generated by the restoration) and lately execute the registration.

The figures below show two TM-5 images (original with 30 meters and restored with 10 meters).

restau01.gif - 88436 Bytes  restau02.gif - 133324 Bytes

See also:
Other Image Processing techniques.
How to execute a Restoration.