I need to know what is the scientifically best way for supervised clustering classification of landsat 8 as below?
- First mosaic multiple landsat scenes then classify
- First classify each of the landsat scenes then stitch the classification result.
N.B. I found images from different date have mismatched visual. May I expect a scientific reference?
Thank you in advanced!
very good question.
If the dates of images are same then use "First mosaic multiple landsat scenes then classify"
but
If dates are different then try to mosaic first ( always use AOI). If mosaic image are visually ok then do classification on it otherwise classification each scene and mosaic the classified output images.
Attached an article answering your question and including a reference as well.
The scientific approach would be to follow both procedures and compare the findings. And publish.
Article A comparison of fine resolution census and image-based natio...
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing