Binary Classification for Satellite Image

[georg...@gmail.com]

Hello All,

                 I wanted to classify only two classes in landsat Imagery. For example Agriculture land and others. Anyone knows how to process it kindly help me to do it. Thanks

[Ujaval Gandhi]

This process is called 'supervised classification'. You need to build a 2-class classifier and classify the image. There is a lot of literature and documentation on how to do it. It is not a simple proces and will involve a bit of learning curve as well as manual effort to collect training samples.

In QGIS, you can use SCP plugin https://www.youtube.com/watch?v=ZPCpWl7NZlU

In Google Earth Engine, you can use one of the ML algorithms such as Random Forest to classify the image. I have a whole module on how to do it as open courseware https://courses.spatialthoughts.com/end-to-end-gee.html#module-3-supervised-classification

If yearly classification is ok, you can get ESA WorldCover data for 2020 or 2021 and just extract the Agriculture class. This will save you the effort of doing your own classification. I have a tutorial to work wth WorldCover data in QGIS https://www.qgistutorials.com/en/docs/3/calculating_raster_area.html

---

Ujaval Gandhi

Spatial Thoughts

www.spatialthoughts.com

On Tue, Jan 3, 2023 at 11:23 AM georg...@gmail.com <georg...@gmail.com> wrote:

Hello All,

                 I wanted to classify only two classes in landsat Imagery. For example Agriculture land and others. Anyone knows how to process it kindly help me to do it. Thanks

--
Datameet is a community of Data Science enthusiasts in India. Know more about us by visiting http://datameet.org
---
You received this message because you are subscribed to the Google Groups "datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datameet+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/datameet/80beb114-1b45-4d00-be16-7db4b5009f91n%40googlegroups.com.