Vehicle behavior prediction method based on real-time vision

Abstract

The invention discloses a vehicle behavior prediction method based on real-time vision, and belongs to the technical field of vehicle behavior prediction, and the method comprises the steps: 1, detecting lane lines around a current vehicle according to a lane line detection algorithm, building a lane line diagram, collecting the information of a to-be-analyzed vehicle, and carrying out the detection of the lane lines; inputting the to-be-analyzed vehicle into the lane line schematic diagram according to the collected to-be-analyzed vehicle information; 2, carrying out dimension removal assignment on vehicle speed information, lane information and signal lamp information in the to-be-analyzed vehicle information, and establishing a to-be-analyzed vehicle information set; and step 3, clustering all to-be-analyzed vehicle information sets, the deep learning network is only used in the initial establishment stage and is not used in the subsequent analysis process, so that more system resources can be saved, the operation speed can be improved, and the problem that the existing vehicle behavior prediction method usually uses a large number of neural network models, and the prediction efficiency is high is solved. And more system resources are consumed, so that the operation speed of the system is low.

Description

technical field [0001] The invention belongs to the technical field of vehicle behavior prediction, in particular to a real-time vision-based vehicle behavior prediction method. Background technique [0002] The automatic driving function refers to the function of guiding and making decisions on the driving tasks of the vehicle without the need for the driver to perform physical driving operations on the self-driving vehicle, and instead of testing the driver's control behavior to enable the vehicle to complete the safe driving function. In the context of autonomous driving, vehicle behavior prediction is the judgment of the driving trend of surrounding vehicles in the next few seconds. Since the driving behavior that the vehicle can take on the road is not unique, it can predict a variety of different behaviors of the vehicle. [0003] In the prior art, a method based on deep learning is generally used to predict vehicle behavior, for example, by collecting a large number o...

AI Technical Summary

Problems solved by technology

[0003] In the prior art, a method based on deep learning is generally used to predict vehicle behavior, for example, by collecting a large number of samples of vehicle driving paths, training a neural network model, and then using the trained neural network model to perform real-time prediction on the future driving path of the vehicle. Forecasting; however, the method of predicting vehicle behavior using a neural network based on a deep learning algorithm consumes more system resources and has a slow calculation speed. Therefore, it is necessary to propose a more efficient solution to realize the prediction of vehicle behavior

Images