With the spirit of reproducible research, this repository contains all the codes required to produce the results in the manuscript: S. Dev, Y. H. Lee, S. Winkler, Systematic Study of Color Spaces and Components for the Segmentation of Sky/Cloud Images, Proc. IEEE International Conference on Image Processing (ICIP), Oct. 2014.
Please cite the above paper if you intend to use whole/part of the code. This code is only for academic and research purposes.
The author version of this manuscript is manuscript.PDF.
All codes are written in MATLAB. Thanks to Florian Savoy for editing the original version of the codes with better readability and higher efficiency.
The dataset used in this manuscript is HYTA dataset from Li et. al, A Hybrid Thresholding Algorithm for Cloud Detection on Ground-Based Color Images, Journal of Atmospheric and Oceanic Technology, 2011. Please contact the respective authors for the dataset. Please save the dataset images and ground-truth maps in a new folder ./HYTA+GT.
bimod_degree_new.mCalculates the Pearson's Bimodality Index of any vector.extractChannels.mExtracts all the related color channels used in this paper for a given image.pca_analysis_new.mPerforms Principal Component Analysis (PCA) on the color channels.rectangleGrid.mGenerates the semi- rectangle grid used in the paper.
In addition to all the related codes, we have also shared the pre-computed results generated from HYTA dataset. The pre-computed files (in case HYTA dataset is not present) are contained in the folder ./preComputed.
The program main.m is the main script, that reproduces all the results. It uses different helper scripts stored in the folder helperScripts. It also reproduces the figures and tables in this associated paper.