Automatic driving preceding vehicle cut-in scene extraction method based on signal rule

Abstract

The invention discloses an automatic driving preceding vehicle cut-in scene extraction method based on a signal rule. The method comprises the following steps: S1, acquiring driving data and position data of a vehicle, position relation data of a target vehicle and the vehicle, and video data synchronized with the driving data of the vehicle; s2, extracting a signal segment in a preset first time range according to a preset signal rule, and extracting a video segment, corresponding to the signal segment, in the video data in a preset second time range; and S3, performing scene marking on the video clip and storing the video clip. According to the invention, based on the automatic driving controller signal, the driving track of the target vehicle in the signal is identified by using the rule, so that whether the target vehicle enters the lane is judged and the video meeting the signal rule time period is extracted, thereby improving the extraction efficiency, liberating human resources and achieving higher extraction accuracy.

Description

technical field [0001] The invention belongs to the technical field of automatic driving algorithms, and more specifically relates to a signal rule-based method for extracting a scene where a vehicle in front of automatic driving cuts in. Background technique [0002] Recently, autonomous driving has gradually become an important field of research in the automotive industry. With the improvement of the level of autonomous driving, more and more computers are involved in driving, and more and more preparations are required to ensure the safety of autonomous driving. According to statistics, a large number of traffic accidents occur when vehicles change lanes. Therefore, the recognition of the front vehicle cut-in during automatic driving is particularly important. In order to train the vision algorithm that can recognize the front car and the person, a large amount of video data of the corresponding scene is required for training, testing, simulation, etc. At present, the vi...

AI Technical Summary

Problems solved by technology

It is based on the sudden change in the longitudinal distance of the target vehicle to judge whether there is a vehicle cutting into or out of the driving lane, but this method cannot rule out the sudden change in the longitudinal distance caused by the sudden deceleration or acceleration of the target vehicle, resulting in inaccurate extraction results

Images