# How to calculate Impervious surface ratio

I have a task, that I need to compare the aerial photo of my town with an approximate five-level degree of Impervious Surface (0%, 25%, 50%, 75%, 100%) for the various districts of my town, e.g. residential areas with gardens, commercial areas, heavily sealed mixed-use quarters, etc.

For example:

In this picture, I can tell the proportion of the built-up property area and with the aerial photo from WMS. I can identify where the residential areas and commercial areas are. But to identify and calculate the Impervious Surface Ratio is still a challenge!

Does anyone have any suggestions, how do I calculate the Impervious Surface Ratio and identify the Impervious Surface Ratio for each speciality Area? (for example: residential areas with gardens, commercial areas, heavily sealed mixed-use quarters...)

• reading @KeaganAllan answer seem that using WMS aerial photo is not optimal. Have you check that there are no other source that you can use ? depending on your area of interest data that will let you get what you need may already exist
– J.R
Jun 29, 2022 at 15:40

Doing it off a WMS Aerial Photo will be a manual process.

An option to consider might be to use a satellite image. Sentinel-2 is relatively fine resolution and free. You could run a few processes on the image to try and identify vegetation / man-made surfaces.

You could then look at the ratio of these per district and see how they compare - bearing in mind the resolution of the image may be coarser than the size of the polygon you are investigating.

Some options from the satellite imagery side would be:

1. Use NDVI to identify areas of high vegetation and no vegetation
2. Create a rough unsupervised classification and manually group classes together that are obviously vegetation or hard-surfaces.

Sentinel-2 NDVI

Unsupervised Classification QGIS

Sentinel Hub

These are suggestions for further investigation. I am sure there might be fancier methods out there.

A quick Google Search brought up a few articles on the subject:

Research Article 1

ArcGIS Tutorial / Lesson

PDF Aricle using Landsat

• thank you for your Suggestion! i found out, in that WMS i have a Colorinfrarot Photo, which show me the vegetation / man-made surfaces. Just one problem, that was just a Photo and doesn't have value of each Area, so i need to classify this Photo! I just don't have much skill in Python, but i may find a way to solve it!
– AnAn
Jan 18, 2022 at 23:53
• I guess it depends on how the WMS is setup, you may not be able to get a "machine readable" dataset for use in Python...or any other process. You may be able to use some sort of Computer Vision (Python OpenCV, PIL etc.) to do this...but running that off a WMS...may not be possible. Jan 19, 2022 at 1:34