Research on nighttime road visibility monitoring based on video images

Road traffic accidents have become a serious social problem, with a significant proportion of accidents caused by insufficient visibility on roads at night. Therefore, nighttime road visibility detection based on video images has become one of the difficulties and a key issue in domestic and international research.

This study analyzes the importance of nighttime road visibility monitoring, introduces the structure, working principle, and monitoring method of a video image nighttime visibility monitoring system, and proposes a nighttime road visibility monitoring method based on video images. Based on the characteristics of nighttime images, an improved dark channel prior method was adopted to calculate the nighttime road visibility. This method mainly includes eight steps: video image acquisition, image grayscale processing, calculation of image average variance, image average gradient, drawing grayscale histograms, image enhancement based on the calculated values, calculation of transmittance, and calculation of visibility.

The experimental results show that the proposed night road visibility monitoring method based on video images can effectively realize real-time monitoring of night road visibility, effectively overcome the inherent defects of traditional methods, and the constructed night visibility monitoring framework can realize high-precision visibility calculation, and has broad application prospects.

Through adaptive threshold and adaptive filtering technology, the improved dark channel algorithm has shown competitive advantages in both image quality index and practical application effect, especially in noise suppression and edge preservation. However, under extreme illumination conditions, the algorithm still has room for improvement in the processing of the strong light source region, and the dark channel prior may lead to bias in the transmission estimation.