The safety and effectiveness of transportation systems are greatly dependent on vehicle detection in tunnel environments. Due to poor lighting, vehicle occlusions, and the unpredictable nature of traffic flow, this work is difficult. The development of reliable and precise vehicle identification techniques that are especially suited for tunnel conditions has been the subject of extensive research in recent years. This study provides an overview of vehicle detection methods developed specifically for tunnel conditions. This study evaluates various methodologies, such as conventional computer vision techniques and deep learning-based algorithms and discusses their advantages and disadvantages. In this study, the proposed methodology used bar filter and grate filter-based technique for vehicle detection and shows better outcome than the traditional methods. In tunnel settings, this study discusses different datasets and evaluation metrics commonly used for benchmarking vehicle detection algorithms. Traditional methods for detecting vehicles in tunnels often rely on manually designed features like color, texture, and motion-based cues, combined with traditional machine learning classifiers. These methods can produce satisfactory results but struggle in difficult lighting conditions and heavy traffic scenarios. However, in recent years, deep learning-based approaches, particularly convolution neural networks (CNNs), have demonstrated impressive performance in vehicle detection tasks. These approaches utilize large-scale annotated datasets and can learn complex representations directly from pixel data. As a result, they can effectively handle challenging lighting conditions, occlusions, and various vehicle orientations. Various datasets have been created specifically for assessing the effectiveness of vehicle detection algorithms in tunnel settings. These datasets encompass a range of lighting conditions, tunnel structures, and traffic scenarios, offering realistic and diverse testing scenarios. Common evaluation metrics, such as detection accuracy, false positives/negatives, and computational efficiency, are used to gauge the performance of different algorithms. The progress made in vehicle detection within tunnel environments has significant implications for enhancing traffic management, safety, and autonomous driving systems. Precise detection of vehicles in tunnels enables improved traffic flow optimization, incident detection, and real-time decision-making for autonomous vehicles. In this study, a proposed method utilized bar and grate filters based on mathematical calculations to evaluate vehicle detection in tunnel environments. The study demonstrates superior outcomes for vehicle detection in tunnel conditions compared to previous traditional methods.