On-line calibration method of multi-beam laser radar at the field end

Abstract

The invention discloses an online calibration method for multi-beam laser radar at the field end, comprising: marking the laser radar to be calibrated at the field end; specifying initial external parameters for the laser radar to be calibrated; specifying obstacles for the laser radar to be calibrated, and recording the obstacles Relative to the reference parameters of the laser radar to be calibrated; the laser radar to be calibrated scans obstacles according to the preset frequency under the initial external reference, and compares the obtained obstacles relative to the parameters of the laser radar to be calibrated with the reference parameters, and according to the parameters Determine whether to update the external parameters based on the difference between them and the reference parameters; save the updated external parameters as the new initial external parameters. It adjusts the external parameters in real time through the online calibration of the laser radar, which ensures that the field-side radar can still operate stably even if it is interfered by itself or external factors, and always maintains high accuracy. It inputs accurate external parameters for positioning and perception segmentation to ensure positioning. And the accuracy requirements of the results of perception segmentation, that is, to ensure the smooth progress of automatic parking.

Description

Technical field [0001] The present invention relates to the field of automatic driving vehicles, and more particularly to an online calibration method of a multi-wire beam laser radar. Background technique [0002] Automatic parking refers to automatic parking spaces that do not require manual control, that is, automatic parking can help drivers automatically park. Different automatic parking systems use different methods to detect objects around the car. For many drivers, the monocular park is a painful experience. The parking space of large cities is limited, and the space to enter the small space has become a must-have skill. There are very few people who have stopped the car, and the parking may cause traffic to block, the nerve exhaustion and the bumper are hit. Therefore, the automatic parking function is born, and this technology is currently realized by the following ways: [0003] 1. Enter the inductor around the bumper before and after the car, then use the computer on ...

AI Technical Summary

Problems solved by technology

[0006] The above-mentioned methods 1 and 2 both need to install additional equipment in the car, and then need to undergo repeated debugging before they can be truly applied, which greatly increases the cost of the vehicle. Although method 3 can avoid installing expensive sensors in the car, thus Effectively reduce costs, but the calibration of multiple lidars at the field end has become a primary problem that must be solved. The accuracy of calibration is directly related to the accuracy of perception segmentation and positioning accuracy, which affects whether the entire automatic parking function can be realized smoothly. At present, people only Pay attention to the calibration accuracy of the lidar after it is installed at the field end, but seldom pay attention to the accuracy of the radar after installation and use. However, the external parameter accuracy of the calibrated field end radar will change with the time and temperature of use. For example, long-term vibration or thermal expansion and contraction will cause changes in external parameters, thereby affecting the realization of the entire automatic parking function.

Images