Almost everyone in the field of Remote Sensing and GIS would have felt thezcpinch of stripes in Landsat 7 image one or other time in their career. In 2003 Landsat 7 has developed a technical snag which called SLC (Scan Line Correlator) off, which has resulted in stripes or gaps throughout the image. Using these images for Scientific Analysis were virtually impossible.
NASA has developed a few techniques to destripe these images. Here is a method using ERDAS imagine to Gapfill Landsat 7 image as suggested by NASA for scientific analysis. Though this problem can never be addressed perfectly as the problem is with the sensor, this is probably the best available solution.
Here you will need a historical images. Basically means to say, you will have an image with stripes for which you will considering for gapfilling, and another image (either without gaps or with gaps) from which the data will be extracted and filled in the gaps of the first image.
A simpler method for gapfilling using Focal Analysis method in ERDAS imagine is also available in my channel, however that method is best suited for display purpose and not for scientific analysis.
Do watch video completely as I have also briefly explained as to how to select the images for carrying out this work. If not understood completely, you might end up selecting wrong images resulting in accurate and unsuitable result.
Following is the video.
.
Notes and Tips:
- While selecting historical images for gap filling, select images which are temporally and seasonally close to each other.
- Do remember to layer stack all the bands once you have completed the gapfilling process for individual bands
- If you are looking for an easy process for display purpose, you can use focal analysis in ERDAS Imagine. Look for the other post for the video tutorial on this.
- Do try to select cloud free data to the extent possible.
No comments:
Post a Comment