Fog Visibility Based on Neural Network Algorithm Research

摘要

In this paper, by confirming the elements of meteorological factors of fog formation, we can find the relationship between fog visibility and the four factors including temperature, relative humidity, air pressure, and wind speed. With different foggy weather conditions, the images of object rendering have diverse characteristics. The visibility prediction value can be obtained from the following steps that Seven factors extracted; image saturation, the minimum gray value, the contrast method of image calculation by using MATLAB as the input layer of the artificial neural network algorithm, visibility as the output layer, training, and simulation study. Through testing, compared with predicted visibility and actual visibility, the discrepancy is only 7.56%. This method improves visibility prediction accuracy, which is of great significance to study the expressway management in foggy weather.