Parking space detection method for autonomous vehicle

Abstract

The invention discloses a parking space detection method for an autonomous vehicle, which comprises the steps of firstly establishing a parking scene local point cloud map, then screening a parking available area, extracting ground point cloud in the parking available area, converting the ground point cloud into an intensity picture, then operating a parking space mark line detection algorithm inthe intensity picture, and finally calculating and outputting relative coordinates of a parking space relative to a vehicle body. The parking space detection method of the invention has the beneficialeffects that: the parking space detection method is not influenced by the change of illumination conditions, and can still work normally under the condition of weak light; the whole algorithm has small computation amount, can run on an ARM processor with low power consumption and low cost in real time, and the whole system has low cost; meanwhile, reflection intensity information in laser radar data is adopted to detect a parking space mark line, and the output parking space coordinates are based on the parking space mark line in a three-dimensional point cloud space, accuracy and stability are achieved.

Description

technical field [0001] The invention belongs to the field of automatic driving, and in particular relates to a parking space detection method for automatic driving vehicles. Background technique [0002] In recent years, with the improvement of sensor performance and the continuous evolution of multi-sensor fusion algorithms, autonomous driving technology has made great progress. From the traditional lane departure warning based on monocular vision to the current lane keeping based on neural network lane line extraction, from the past cruise control based on wheel speedometer and monocular vision to today's technology based on millimeter wave radar and visual sensor fusion Adaptive cruise control. In recent years, the research and development of automatic parking functions for autonomous vehicles has been a hot topic in both academia and industry. In the development process of the automatic parking function, how to accurately identify the parking space is the primary chall...

AI Technical Summary

Problems solved by technology

However, due to the use of a considerable number of sensors and V2X base stations, the one-time investment cost of parking lot construction has increased significantly, and subsequent equipment maintenance is also a cost

At the same time, due to the use of a large number of position sensors, the location coordinates of each sensor need to be calibrated during the construction of the parking lot. The calibration process is cumbersome and complicated. The expansion of the parking lot in the later stage also requires the expansion of the original position sensor network.

Images