An automatic deviation correction method based on UAV in parking lot scene

Abstract

The invention discloses an automatic deviation correction method based on an unmanned aerial vehicle in a parking lot scene. The method is to obtain an image of a vehicle driving on the road below through an unmanned aerial vehicle, combine the Hog features of the image, and use the SVM image classification model to classify the image into Class to a certain image type, and then map the image type to the current required deviation correction control command through the image type-deflection angle mapping table and deflection angle-control command mapping table, and send it to the vehicle, and the vehicle translates the control command into vehicle control Command and execute to realize vehicle deviation correction. This method relies on the UAV and its visual sensor, and does not require significant modification of the vehicle, which reduces the complexity and cost of the vehicle. In the process of deviation correction, from the UAV to collect the image to the target vehicle to execute the control command, there is no need to Manual participation can realize automatic deviation correction in the true sense.

Description

technical field [0001] The invention belongs to the technical field of automatic driving, and in particular relates to an automatic deviation correction method based on a drone in a parking lot scene. Background technique [0002] In recent years, drones have developed rapidly and have been widely used in various fields such as transportation and agriculture. During the driving process of the vehicle, the drone can provide the car with a global perspective, and these visual images contain the environmental information around the vehicle. UAV-based car navigation will abandon the traditional way of relying on vehicle-mounted monocular / binocular camera navigation, and replace the first perspective with a third perspective, bringing a more global vision to the car and bringing new possibilities to autonomous driving. sex. [0003] In the field of autonomous driving, people generally use on-board cameras to provide visual information for vehicles to achieve the purpose of navi...

AI Technical Summary

Problems solved by technology

[0006] (1) The entire device is placed on the vehicle, and the vehicle needs to be refitted, adding many additional devices, increasing the complexity of the vehicle structure and the cost of the vehicle;

[0007] (2) The deviation correction method involved in the device needs to establish a perfect reversing knowledge base in advance, and the establishment of the knowledge base is a complicated and lengthy work;

[0008] (3) The reversing deviation correction of this device can only provide the reversing trajectory line, and the real deviation correction process still needs to be completed manually by the driver, which cannot realize automation in the true sense

[0011] (1) The system can guide the vehicle through the drone, but the vehicle is still driven manually, and the driver still needs to correct the offset during driving;

[0012] (2) The drones of this system mainly rely on ultrasonic waves, without visual sensors or visual positioning systems, which cannot guarantee accurate indoor positioning and stable indoor fixed-point hovering of drones

Images