Development of an Improved Image Processing Based Visibility Model

An image processing based visibility model is presented in which model calculations are displayed as synthetic color photographs that have the appearance of a smog event. A theoretically based model for light scattering and absorption in a plane parallel atmosphere is used to determine sky color and to provide accurate prediction of skylight addition to the line of sight between the observer and objects in the field of view. Data for testing the visibility model are collected experimentally, including standard photographs of chosen vistas and measurements of the chemical composition and size distributions of the atmospheric aerosol. The model is verified by comparison against radiometer measurements as well as point by point comparison of actual photographs and synthetic images. It is shown that the image processing based visibility model performs well, producing representations of the sky color and objects in the field of view that are close to those seen in actual photographs.