Parking assisting method

Abstract

The invention provides a parking assisting method. The method includes: adopting a camera for online real-time image data acquisition, correcting image data to generate an aerial view, and segmentingthe aerial view; detecting out all straight lines according to a straight line detection algorithm; adopting an exclusion algorithm for intersection point neighborhood binary horizontal and perpendicular gradient calculation of all intersection points of the straight lines, and judging whether gradients are larger than a second threshold or not; obtaining parking space data according to parking space angle points; acquiring the vehicle speed, vehicle position and posture and coordination information, calculating world coordinates of a parking space according to a relative position relation between the parking space and a vehicle, performing path planning according to the world coordinates, and entering a parking mode. In a vehicle parking process, real-time correction of the parking spaceis performed according to an image processing algorithm, and a high-accuracy high-success-rate parking assisting function is realized. In addition, due to adoption of a novel parking space detection algorithm, parking space detection errors can be effectively reduced.

Description

technical field [0001] The invention belongs to the field of automobile intelligent driving, and in particular relates to a parking assistance method. Background technique [0002] With the rapid development of science and technology, the automobile industry is facing major innovations in recent years, and autonomous driving has become a research hotspot for major automobile manufacturers and universities. According to the classification of autonomous driving technology, L1-level assisted driving is currently the most feasible technical solution. Among them, the parking assistance system has gradually become the standard configuration in new products of major automobile brands. [0003] However, due to the traditional parking assistance system solution, the parking space perception is mainly performed through ultrasonic radar. Due to the limitations of the working principle of the ultrasonic radar, the solution is poorly adaptable to the parking space scene. For example, i...

AI Technical Summary

Problems solved by technology

Due to the limitations of the working principle of the ultrasonic radar, the solution is poorly adaptable to the parking space scene. For example, in a parking lot with no cars in the front and rear, the solution cannot detect the parking space, which makes the parking function based on the ultrasonic radar invalid.

In addition, due to the poor anti-interference characteristics of the ultrasonic radar, the detection accuracy of the parking space size and coordinates is also poor, which further weakens the applicability of the scheme

[0004] At present, based on multi-sensor parking assistance system, the existing schemes based on ultrasonic radar and multi-camera mostly use ultrasonic radar to detect parking spaces, and use cameras to generate surrounding views, so the anti-interference characteristics are poor. The detection accuracy of the coordinates is also poor

And although there is also a multi-sensor cooperative work scheme using ultrasonic radar and camera in the prior art, which technically uses a camera to detect parking spaces, but on the one hand, because the intersection of straight lines satisfying a certain angle is used as the corner point of the parking space, the accuracy of the algorithm is low. , which can easily lead to false detection of corners of parking spaces; on the other hand, its parking space detection technology only detects through the camera during the driving process of the vehicle, and does not perform real-time correction of the parking space information during the parking process, which also leads to low accuracy.

Images