Recognising daytime and nighttime driving images using bayes classifier

Variation in outdoor light is a very important consideration when developing an intelligent vehicle recognition system. Previous researchers have focused on the development of recognition systems intended exclusively for either daytime or nighttime use. This is because a constant parameter threshold value cannot be used in an algorithm for a system working at different times in the day or night. However, this study presents a system that can be used for intelligent vehicle recognition at any time. This system automatically recognises whether the image outside the vehicle is a daytime image or a nighttime image. The algorithm used in this system includes a detection module for candidate daytime sky regions, a horizontal position search module to calculate the vanishing point of a road and a recognition module to determine whether the image is a daytime or a nighttime image. The detection module and the search module are used to find a system-defined feature region, and the recognition module is used to recognise feature values using Bayes classifier. Experimental studies demonstrated a recognition rate of 96.22% and confirmed the feasibility of the system for highway images, city images and mountain road images.