Hi, I have two different LandSat 7 ETM+ images that I want to detect the overall change for them. I follow the technique to classify each image and perform image differencing. my question is: do I need to layer stack all the bands and create multispectral images before classification, or I can do the change detection from one band only, and which Band is the best?
No need to stack all bands. You can stack only required bands. and secondly what is your purpose of classification (vegetation or built up area). If you are going to focus on one class then you need to required bands according to purpose for (true or false colour combination) visualization.
As Aqil said, no need to stack all bands and it depends on the used software for analysis. The issue of best bands depends on the purpose of the classification.
Nahed Osama you may want to use the "ENVI" software. 1 to 5 and 7 have the same spatial resolution of 30 m each.
Hi Nahed Osama , layer stacking is generally not suggested for change detection as it occupies more storage. If your objective is change detection identify the required bands and compute indices (vegetation, built up, water bodies, etc.) To get better insights on different band combinations, visit https://www.usgs.gov/faqs/what-are-best-landsat-spectral-bands-use-my-research?qt-news_science_products=0#qt-news_science_products
Hope it helps, all the best !
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification