Method for perspective perception of intelligent vehicle front road environment on basis of V2V

Abstract

The invention provides a method for perspective perception of intelligent vehicle front road environment on the basis of V2V. The method comprises the following steps of: obtaining a front vehicle and a rear vehicle running on a road, and establishing connection between the front vehicle and the rear vehicle through a cooperative communication protocol; obtaining front video data of the front vehicle by the rear vehicle so as to establish a front vehicle video image; detecting a road target in real time by the rear vehicle and obtaining a positioning parameter of the detected road target; establishing a rear vehicle video image by the rear vehicle according to front video data of the rear vehicle by utilizing a time label, and synchronizing the front vehicle video image and the rear vehicle video image; extracting SIFT feature points in the front vehicle video image and the rear vehicle video image by the rear vehicle by adoption of an SIFT algorithm, and obtaining matching points; establishing a 3D perspective model; carrying out optimal screening on the matching points; calculating an affine transformation parameter; and fusing the road target in the front vehicle video image into the rear vehicle video image by the rear vehicle so as to obtain an image for carrying out perspective perception of the front road by the rear vehicle. According to the method, the road environment perception ability of vehicles can be effectively strengthened.

Description

technical field [0001] The invention relates to the technical field of intelligent transportation, in particular to a V2V (Vehicle-to-vehicle, automobile-automobile)-based perspective perception method for road environment ahead of an intelligent vehicle. Background technique [0002] In recent years, with the increase in vehicle ownership, the traffic accident rate has also increased year by year. According to the 2016 global road accident statistics report organized by WHO, about 1.25 million people die in traffic accidents every year, and 20 to 50 million people are injured in traffic accidents, causing economic losses of hundreds of billions. Preventing and reducing the occurrence of traffic accidents has become a hot topic in the development of intelligent transportation systems in my country and even in the world. The development of intelligent vehicles with active safety technology has become the common goal of governments, research institutions and automobile manufa...

AI Technical Summary

Problems solved by technology

[0003] According to the statistics of the National Traffic Safety Report of the United States, 31% of traffic accidents come from collisions between front and rear vehicles. When the vehicle is about to overtake the vehicle in front, it collides with a vehicle or pedestrian that suddenly enters the road ahead of the vehicle in front, resulting in a serious traffic accident

These types of front-to-back collision accidents are caused by the rear vehicle's inability to understand the road conditions in front of the vehicle in front, and the current road perception system of intelligent vehicles can only perceive the road environment and road objects within the "visible" range. , but for the occluded road target, it cannot perceive

Images