Performing image segmentation in QGIS?

Question

I have been trying to learn to work with QGIS (open-source GIS software) through some tutorials available online for classification of image for land cover. These tutorials used a dataset from Landsat 8. The tutorial used the semi-automatic plugin v 3.0

Now I want to use the same process for an image in .jpeg format that I downloaded from internet. Step one involves converting the raster bands from DN to Reflectance. For this a .txt file (a metadata file named .MTL) is required. How do I generate this .txt file for my image.

Answer 1

1. where should I begin?

Do you know what Image Classification is? If not here's an intro article ESRI wrote about for arcgis. You don' need arcgis to read it.

Read it, and in the end you'll understand what you should need. Keep in mind that image classification is about creating classes.

To do that should well defined classes beforehand (how many, where are they and etc), and ground truth areas in your dataset that speak of those classes. The classifier logic afterwards tries to find other areas that have similar characteristics using very smart statistical techniques.

2. what are the appropriate Python modules that should be used? P.s. I am learning to use Spectral Python. Is it good enough?

you don't need python yet. Go with off-the-self software first. When you think that they do what you want them to do, try to switch to python libraries.

Good off-the-self software is :

1. Arcgis
2. eCognition
3. Erdas Imagine
4. NB: There's a good variety of Open Source that that can classify images as well (Grass, R, SAGA, Orfeo Toolbox) but the learning curve is much more steeper that those above)

3. Do I need to get a software for the project or just coding in python will suffice? Which option is better? I have read about some good softwares like

see above

4. where can I get the required dataset and what type of the dataset will I be needing? What format of images are appropriate to be used?

Find papers in bibliographies that try to classify ...staff. Check what they are using. Do the same. The datasets specs change with you objectives. Finer datasets costs more. Study the aboves and discuss with your supervisor for what kind of results you want based on your budget.