As of now, I am trying to explore the potential of using high/very high-resolution satellite imagery with machine learning (ML) to identify or map forest cover (e.g. rubber). As a newbie in this field, I have a few questions related to my work:
1. Is it possible to use ML for this application?
2. Is there any existing approach/algorithm to map/to count rubber? If yes, what are advantages and its limitations? How to improve it?
3. Is there any way to use the same training data for different years, (e.g. 2019, 2020, 2021, etc.)? (just want to avoid fieldwork).
I will be happy if you can give me your advice on this kind of mapping.
1. Google Scholar is your friend: https://scholar.google.com/
2. Type in the box "classify rubber using satellite imagery with machine learning"
3. Profit
9000+ results for the last two years, I would guess someone was successful.
So, 'yes'.
I think the ML such as spectral angle Mapper can be used to get reflectance attribute of Rubber and after getting SAM targeted reflectance file it can be used in different years. But accuracy will be increase if you use hyper spectral data if available for region such as AVIRIS NG or Hyperion.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps