Satellite Image Processing Project - Road Recognition and Extraction

I have performed Road-Network-Extraction using classical Image processing and few libraries.

PRE

I performed here various methods and steps and experimented a bit after reading various research paper. I have stored the part I liked, it includes loading the image and applying unsharp making via gaussian filter and then I performed canny edge detection.

Road-Network-Extraction-final-

Here is the final project which include steps as -

• Reading the image
• Image segmentation via K-means clustering
• Coverting image to grayscale image
• Thresholding using epsilon-neighbourhood
• Detect edges using laplacian-gradient method
• Overlay edge and original image

Example Usage

• Install required libraries

  • Opencv
  • Matplotlib
  • Numpy
  • Sklearn

• Clone Repo

  https://github.com/Akash-Ramjyothi/Satellite-Image-Road-Recognition.git

• Change working directory

  cd Road-Network-Extraction

• Run main file

  python3 Main.py

Sample Demo:

Input image

Segmentation using K-Means Algorithm

epsilon neighbourhood thresholding

Result: