Rain drop detection and removal using K-Means clustering
Authors:
- 1 December 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 7, 1-5
Abstract
A novel hybrid algorithm has been proposed in this paper to detect raindrops, remove and then restore the image back ground in a single image. This new hybrid algorithm is framed based on K-Means clustering and Median filtering for the fast retrieval of Rain droplets from the single image. K-Means clustering algorithm is an efficient algorithm for image clustering. The algorithm proposed in this paper has different approach from other established numerical schemes. The hybrid algorithm is framed in order to identify the rain droplets using clustering and shape modeling of raindrops. The proposed system is fast compared with alternate droplet identification schemes. The process allows rapid evaluation of the contour of the raindrops and convergence to its final resultant with very little iteration. The experiments demonstrate the efficiency and accuracy of the method.Keywords
This publication has 5 references indexed in Scilit:
- Video-based raindrop detection for improved image registration2009
- Raindrop detection on car windshields using geometric-photometric environment construction and intensity-based correlation2009
- Single image dehazingACM Transactions on Graphics, 2008
- Visibility in bad weather from a single image2008
- Detection and removal of rain from videos2004