Guided Filter Based Colour Image Rain Streaks Removal Using $$\varvec{L}_{0}$$ Gradient Minimization Method

In recent time, removal of rain from videos and images is becoming a popular research interest in computer vision world. Outdoor visibility degrades rapidly due to rain and it gives direct affect in object detection, surveillance systems and many other image processing applications. Different research works on image rain removal have already been proposed which perform their operations based on dictionary learning and sparse coding, different image decomposition techniques, probabilistic model, entropy maximization and many more. In this paper, a novel rain removal technique is proposed where high frequency rain component of a rainy image is guided by low frequency non-rain component to remove the rain streaks by guided filter. As the guided filter is not able to remove satisfactory number of rain streaks, therefore \(L_{0}\) gradient minimization technique is applied on the image obtained after guided filtering. Overall sharpness of the output image is enhanced by unsharp masking (USM) operation. Finally contrast limited adaptive histogram equalization (CLAHE) technique generates the contrast enhance desired output. Qualitative and quantitative comparisons are performed to show the superiority of the proposed technique.