Cloud removal
A detailed overview of the problem solving tool:
Cloud detection on satellite images
What is that?
Eliminating clouds on space data increases the amount of useful information and improves the quality of decryption.
When clouds are extracted, a mask is formed, which allows to increase the efficiency of fast imaging and speed up the image acquisition process. The quality of image processing with distant clouds is improved in comparison with processing of an image where clouds are present.
For example, this becomes evident in texture analysis and keypoint selection (clouds can seriously degrade the quality of matching).
How the cloud
detection tool can help
Cloud extraction can help with:
- stitching routes - to create homogeneous routes;
- thematic processing - to accurately analyze the data;
- geo-referencing the image - for more accurate automatic search of reference points;
- mapping - to exclude clouds in classification.
Example of cloud removal
on a satellite image
In the IMC PC, the cloudiness selection is automatic, the operator only needs to set the values:
- lower brightness threshold: possible values range from 0 to 100, where 100 is the maximum brightness value above which the area is considered as a possible cloud;
- upper gray point deviation threshold: possible values range from 0 to 100, where 0 is the gray point value at which R, G, B are equal;
- minimum cloud area, m2: areas smaller than the given area will not be considered as cloud.
After applying buffering to the Clouds vector layer, the operator should select cloud shadows, as the area under them is uninformative.
