Python implementation of shadow detection and correction algorithm proposed in academia
This repository contains functions to detect shadow - covered areas in aerial/satellite imagery and to correct for the brightness of the image in the shadow - covered areas as proposed in the paper 'Near Real - Time Shadow Detection and Removal in Aerial Motion Imagery Application' by Silva G.F., Carneiro G.B., Doth R., Amaral L.A., de Azevedo D.F.G. (2017).
In addition, the multilevel Otsu thresholding method is not used in view of the computational complexity and time cost incurred. As such, it is replaced by the K - Means clustering algorithm, which is another method to determine the multiple thresholds in a global context, which would give a close approximation to that of the multilevel Otsu thresholding method.
Requirements:
- cv2
- gc
- numpy
- pandas
- rasterio
- scikit - image
- scikit - learn
Test Image (courtesy of INRIA Aerial Image Labeling Dataset):
Shadow Mask:
Shadow Compensated Image:
Close Up View of one portion of Test Image:
Close Up View of corresponding Shadow Compensated Image:
