For RS and GIS applications such as yours, ERDAS IMAGINE, ENVI, ARCMAP, IDRISI may be of great help. They are dedicated remote sensing and GIS tools that provide platforms for a great variety of applications such as mapping, and change analysis. Since your study is mapping roofing materials, all of the above mentioned softwares provide good platforms for your work. There are thousands of videos available online for learning and understanding supervised classification. Data preparation is equally important as well. Refer "Introductory digital image processing" by Jenson for more detailed information on both classification and data preparation.

For simple use (uniforme roofs not complicated color-difference in the image) :

- You need to create your simple (for each roof-type), using the tool Region of interest in ENVI software for exemple

https://www.harrisgeospatial.com/docs/RegionOfInterestTool.html#ROIPixel

- If you need a more deep application (but who needs more knowledge in remote sensing, spectral reflectance of roofs...), you can use the Weka tool (Data mining):

https://www.cs.waikato.ac.nz/ml/weka/mooc/dataminingwithweka/

Good luck

Dear Walid

Artificial objects detection is a quite simple application from satellite images, but also an interesting one to understand the basics of remote sensing data processing. As mentioned before,

Many softwares can be used in your case.

Personally, I suggest ENVI 5.x.

The following document can help you:

https://www.harrisgeospatial.com/docs/ClassificationTutorial.html

Also, a lot of videos are available in YouTube.

Good luck.

Note that many open source softwares are also available!

Rachid

Hi Harold,

in case you do not have a license for the commercial softwares mentioned above, Python is great for supervised image classification. I usually use the gdal package (conda install gdal) to read the satellite images (.tif files) and scikit-learn to train the classifier (support vector machines, neural net, random forest...etc). You will have to create a training set, a set of labelled pixels (roof materials in your case) that you will use to build the classifier. Also you will need to select features, which in the case of satellite imagery can be the intensity of the pixels in the different spectral bands. Make sure that you split your training set so you have a subset for training and a subset for testing the accuracy of the classifier (or better use cross-validation).

You can visit my GitHub repo https://github.com/kvos/CoastSat where sand is classified on satellite images.

All the best,

Kilian

I agree that Python is a good option. Besides, with QGIS and grass you can do image classification. You may want to take a look on the semi-automatic plugin (https://fromgistors.blogspot.com/p/semi-automatic-classification-plugin.html). Also, Google Earth Engine provides a nice platform to do image classification with minimum code. You have good examples (https://developers.google.com/earth-engine/tutorials) and youtube videos https://www.youtube.com/watch?v=WjKhPyiSgb8

Sucess

Hello

building roofs have different colors so we can make more then class for them in the beginning classification then we can merge them to get more accuracy.

Wow! Thank you so much for all the help, sirs and ma'am. I now have a guide as to how i am going to work on my problem. I greatly appreciate all your help as well as for providing me with references and personal Github repo.

My sincerest gratitude. :)

Hello,

there are different roof materials, so if you want to make supervised classification for that you need to select many training areas for each class of them. really it is difficult, I tried that few days ago

Dear Muthanna,

How did you perform roof image annotation, Sir? Did you use drones?

What software are you using for supervised classification?

Thank you very much.