input:linear scene referred data: the hazy landscape image
output:the image with haze removed
depth:the estimated optical thickness of the image
