detects horizon line in images of landscapes
This code works to detect the horizon line in images of landscapes which can be used by drones. To run code, run python horizon_detect.py --scale <scale_value> <image_name> Example: python horizon_detect.py --scale 50 “image3.png”
Image with line will be saved to current directory.
The value of scale chosen is a tradeoff between speed and accuracy. Higher the value of scale chosen, higher will be the accuracy in detecting the horizon. However, this would mean more processing time, which would not work well with real time video.
I have performed the algorithm by down-sampling the image to a much smaller scale and then extrapolating the line obtained onto the original image size. The algorithm works in such a way that it finds a straight line to maximize an optimization function, such that all the similar sky pixels are grouped above the line and the ground pixels are grouped below the line. Further details can be found in the comments of the code.