Image-based urban road surface adhesion coefficient acquisition method

Abstract

The invention provides an image-based urban road surface adhesion coefficient acquisition method. The method comprises the steps of firstly establishing a road surface image information base, then establishing a road surface image data set, establishing and training a road surface image area extraction network, then establishing and training a road surface type recognition network, and finally acquiring road surface adhesion coefficient information. According to the method, road surface adhesion coefficient information can be provided for development of an intelligent driving assistance system and an unmanned driving system; according to the method, the adhesion coefficient is obtained by obtaining a front road image, and the road surface adhesion information can be obtained in advance; according to the method, a serial method based on the image road surface area extraction network and the road surface recognition network is designed, and then the road surface recognition network is structurally simplified, so that the adhesion information of the front road surface can be quickly acquired in real time.

Description

technical field [0001] The invention belongs to the technical field of smart cars, and relates to a method for obtaining road surface adhesion coefficient, and more specifically, relates to an image-based method for obtaining road surface adhesion coefficient of urban roads. Background technique [0002] With the development of automobile intelligence, users have higher and higher performance requirements for vehicle-mounted intelligent driving assistance systems and unmanned driving systems, and the improvement of the performance of most intelligent driving assistance systems depends on the accuracy of dynamics control. The design of the control system needs to obtain real-time and accurate road surface information. The estimator based on the dynamic model can obtain real-time and accurate road adhesion coefficient estimates, but this dynamic estimation method largely depends on the accuracy of the vehicle model and tire model. Moreover, certain driving incentive conditions...

AI Technical Summary

Problems solved by technology

[0002] With the development of automobile intelligence, users have higher and higher performance requirements for vehicle-mounted intelligent driving assistance systems and unmanned driving systems, and the improvement of the performance of most intelligent driving assistance systems depends on the accuracy of dynamics control. The design of the control system needs to obtain real-time and accurate road surface information. The estimator based on the dynamic model can obtain real-time and accurate road adhesion coefficient estimates, but this dynamic estimation method largely depends on the accuracy of the vehicle model and tire model. And it needs to meet certain driving incentive conditions. In addition, the dynamic estimation results reflect the road adhesion coefficient at the contact mark between the tire and the road surface, which has a certain hysteresis, and it is difficult to obtain the predicted value of the road adhesion coefficient in advance.

[0003] At this stage, more and more smart cars are equipped with cameras and other equipment to obtain road information and surrounding vehicle information, which brings new opportunities for the research of road adhesion coefficient recognition methods. Its advantage is that it can perceive the road conditions ahead, so it has A certain predictive ability enables intelligent driving vehicles to adjust the control strategy in advance in the event of a sudden change in the road surface, improving the ability to respond to dangerous conditions, but how to use image sensing information to obtain road surface adhesion coefficient information is still a challenge

Images