Snow Detection And Removal From Video Image

Snow weather may significantly have negative impacts on transportation sector,especially in most parts of high latitude during winter period.This can affect people’s travel and transportation.In particular,the sudden snow disaster causes serious disruption to rail transport and has a great impact on transport safety.It is of great significance to remove snowflakes from the background of the monitoring camera image to improve vision to prevent potential accidents on railway transport,as well as the development of real-time monitoring and forecasting and warning technology of snow.Snow removal from video/image is a very challenging and complex chore.Normally the snowflakes affect only a very small region of an image,hence the confusion to determine which region should be considered and which one should not.In this research,a new technique has been implemented which effectively uses the LO gradient minimization approach to remove the snow pixels.The minimization technique can globally control how many non-zero gradients are resulted in the image.The method is independent on local features,but instead locates important edges globally.These salient edges are preserved,low amplitude and insignificant details are diminished.The snow pixels are removed in this way.Experimental results show that the proposed method is highly efficient snow removal algorithm as it removes snow effectively even under heavy snow conditions,while preserving the details of the image.Main work of this study is:(1)To detect snow from the video image using of sobel filtering method.(2)To remove snow from video image by using an operation called smoothing.The smoothing operation identifies the salient regions i.e.the highest-contrast edges by confining the number of nonzero gradients in a global manner.