摘要:
The invention relates to a method for automatically detecting a driving maneuver of a motor vehicle (A), in particular during a passing maneuver or an evasive maneuver, wherein the vehicle environment is captured and an electronic image is produced thereof, the electronic image is used to detect a lane and/or road and objects (B, C) in the surroundings of the vehicle, longitudinal and transverse dynamic movement information ( Ψ, a y , δ H , ω FL , ω FR , ω RL , ω RR ) of the motor vehicle (A) is determined, and the position (X) of the motor vehicle (A) is estimated odometrically on the basis of the data (b Lane / Y Lane / θ , C 0 ) from the lane and/or road detection and/or from the movement information ( Ψ, a y , δ H , ω FL , ω FR , ω RL , ω RR ) of the motor vehicle (A). According to the invention, a) the following indicator variables are formed from the estimated position data (X) of the motor vehicle (A): a lateral distance value (LO L , LO R ) of the motor vehicle (A) from a lane or road marking (L), a time-to-collision value (TTC A,B ) relative to the distance ( d ) from the object (B) located in the driving direction, in particular from the vehicle ahead (B), a longitudinal dynamic passing or evasive maneuver (I), formed from the indicator variable (TTC A/B ) of the time-to-collision value and a value corresponding to the positioning (FPS) of the gas pedal of the motor vehicle (A), and b) threshold values (I th , TTC A, B, th ) are determined for said indicator variables (LO L , LO R , TTC A,B , I) to be used as criteria both for detecting partial maneuvers of a passing or evasive process, in particular of a following process, a lane change, passing a stationary or moving object (B) and changing into the lane of the passed object (B), and for detecting transitions between said partial maneuvers.
摘要:
The invention relates to a method and to an object detection device (4) for analyzing surrounding objects and/or surrounding scenes using a data processing and/or evaluation device (6), wherein in the data processing and/or evaluation device (6) image data (x t ) are evaluated based on a CRF model and wherein the CRF model provides additional object nodes (o t n ) that take into account the information from an object detector.