Vehicle Detection Based on Improved Gaussian Mixture Model for Different Weather Conditions

Abstract

In the field of vehicle traffic management, the accurate detection of vehicles under varying weather conditions remains a critical challenge. This paper introduces a new method to detect vehicles in different weather conditions, such as regular daytime and nighttime, as well as rainy conditions. Current vehicle detection systems often struggle to work well in bad weather, which can be dangerous. The proposed method is an Improved Gaussian Mixture Model (Improved GMM) designed to adapt to various weather situations. The goal is to ensure accurate and reliable vehicle detection in all conditions, which is important for traffic management. The Improved GMM was tested using real datasets, enabling the simulation of real-world scenarios and the evaluation of its performance. Appropriate measures were selected to assess its effectiveness. Results indicate that the Improved GMM significantly outperforms the traditional GMM in vehicle detection, particularly in challenging weather conditions. Furthermore, the findings reveal that this method is not only effective in diverse weather conditions but also maintains a high level of computational efficiency. In summary, the research suggests that employing the Improved GMM for vehicle detection in different weather conditions can substantially enhance traffic management systems, ensuring reliability regardless of the weather.