roadNetworkExtraction

Road Network Extraction from Aerial Images using Image Processing and Clustering

The whole project is done using java in intelliJ.

It is divided into three stages

1. Image pre-processing
2. Image Clustering
3. Image post -processing

In Image pre-processing the Image is first converted to grayscale Image and adaptive thresholding is applied so that the clustering becomes faster.

And Clustering is applied after thresholding of the Image. Each pixel is labelled accordingly by comparing its pixel value with its 8-neighbors.

In post-processing, the elongated-road clusters is extracted by applying height-weight ratio and its area density. Then thinning is applied to the road-clusters to get the skeleton image of the road and curve fitting is to be done to join the broken road clusters .