IMAGE DE-RAINING CONDITIONAL APPROACH IN REMOVAL OF RAIN STREAK USING CONVOLUTION ALGORITHM

Due to the pollution and the various parameters can affect the atmospheric condition. The aerosols are present in the air over the surrounding environment. The particles can scatter in the light illumination. When the image taken in the atmospheric condition it is affected by the scattering particle and the haze or snow. To remove the haze or illumination in the background image is the difficult task by the hand craft. In this paper they propose the removal of the snow and the illumination by using the convolution algorithm. The experiment has been conducted for the different color spaces. The great intention in the color channel to control the luminance in YCrCb. The neural network can be used for the dehazing method. The CNN can not only achieve the end to end frame it also satisfies the image to image structure. This method can be more comfortable when compare to other method. The multiscale convolution network that can finds the haze automatically and helps in the identification of the texture. The accuracy of the reconstructed image is about 89%. The luminance of the image can be maintained and the contrast is adjusted to the quality image.