Determination Of The Distance Between The Front Vehicle And The Vehicle Based On Video Recognition In Foggy Environmen

Road traffic safety has always been a hot research topic in the field of transportation.Severe weather conditions,particularly foggy weather,are more likely to induce traffic accidents as they reduce visibility and affect drivers’ ability to judge vehicle distances,leading to rear-end collisions.Effective detection of vehicle distances in foggy weather can provide valuable assistance to drivers in making safe driving decisions,reducing the occurrence of accidents,and improving road traffic safety.This article analyzes the applicability of various detectors in foggy weather vehicle distance detection.Due to the lack of precision of laser and ultrasonic detectors in foggy weather,the high cost of radar,and the low accuracy of infrared distance measurement over long distances,this article focuses on the research of vehicle distance detection under foggy conditions based on computer vision technology.Firstly,this article analyzes the formation mechanism of fog,the influence of atmospheric scattering on image quality,and the degradation mechanism of foggy images.Then,the principle of dark channel prior dehazing is analyzed,and it is verified that the traditional dark channel prior algorithm can effectively remove fog.However,there are also drawbacks,specifically that the brightness of the image after dehazing is darker than the original image,and the image details and contrast are reduced.Then,the deficiencies in the improvement of the traditional dark channel prior defogging algorithm are analyzed;it is pointed out that the reason for the dimming of the brightness and the deterioration of the details is that the transmittance and the estimation of the ambient light intensity are too small.On this basis,this paper proposes an improved dark channel prior defogging algorithm that introduces brightness and saturation.First calculate the saturation and brightness relationship between the dark primary color and the fog-free image;then introduce the brightness and saturation into the dark primary color;select the pixels with the lowest transmittance of 0.1% in the image to establish a pixel set,and calculate the ambient light intensity.Thereby improving image brightness and enriching image details.Through experimental verification,the algorithm can be used to deal with foggy traffic scenes,and can effectively help improve the driver’s visual effect.On the basis of effective defogging,the research on vehicle distance detection is carried out.Firstly,a multi-task convolutional neural network license plate detection algorithm was constructed,and combined with the traffic scene,the structure of network input,output,and convolution kernel was designed,and a number of learning strategies,error functions,and redundant frame suppression methods were formulated.measures to improve accuracy.Experiments were conducted with Chinese urban parking datasets to verify the effectiveness of the algorithm and improve the accuracy and real-time performance of license plate location.This paper proposes a vehicle distance detection algorithm based on coordinate conversion,further improving the accuracy of distance detection by detecting the license plate of the preceding vehicle.A hardware experimental system is also constructed to verify the accuracy of distance measurement of the preceding vehicle in foggy environments through experimental results and data analysis.The research of this paper has important practical value and theoretical significance for assisting drivers to improve driving safety in foggy and adverse weather conditions.