Road network detection - 🔗 link

Get the road network Graph data from a screenshot of a map.

The Road network detection algorithm needs "clean" Maps Screenshots without landmarks, buildings, any labels or markers. You can use screenshots of google map from https://mapstyle.withgoogle.com/ with the following settings:

See example output in /z-output.

Usage

$ python3 main.py <path-to-image>
# python3 main.py ./z-input/kirba.png       # for example

Requirements

Used python v3.10.6

# requirements.txt

cv-algorithms==1.0.4
networkx==3.0
numpy==1.21.5
opencv-python==4.7.0.68

Implementation

See https://docs.google.com/document/d/15D3AVszg0GRrYaToa6aiCSQ3aPUiu3zRsPIFWeYtZm0/edit#heading=h.oislpifv04z4

You can see the Breadth First "Edge Detection" algorithm in action in a slideshow here

We have shown an usage of the obtained graph data in main.py. We are using the graph data to find the shortest path between two points on the map. Find the output image in /z-output/dijkstra.png.

References

See https://docs.google.com/document/d/15D3AVszg0GRrYaToa6aiCSQ3aPUiu3zRsPIFWeYtZm0/edit#heading=h.g74h65komzxz

Future Work

See https://docs.google.com/document/d/15D3AVszg0GRrYaToa6aiCSQ3aPUiu3zRsPIFWeYtZm0/edit#heading=h.6z64lx9akiy5