Autonomous parking system and method based on multi-vision inertial navigation integration

Abstract

The invention discloses an autonomous parking system based on multi-vision inertial navigation integration. The autonomous parking system comprises a binocular vision SLAM module, a round looking parking detection module and a decision and control module, wherein the binocular vision SLAM module is used for generating a parking lot simulating map by acquired parking lot images through the multi-vision inertial navigation integration, obtaining real-time positioning information of a vehicle under a coordinate system of the parking lot simulating map, and transmitting the real-time positioning information to the decision and control module; the round looking parking detection module is used for acquiring current position data of the vehicle for current environmental distinction detection, and transmitting the testing result to the decision and control module; and the decision and control module is used for making a vehicle autonomous parking strategy according to the received geographicpositioning information of the vehicle in the parking lot simulating map and the current environmental distinction parameters. The invention also discloses an autonomous parking method based on the multi-vision inertial navigation integration. The autonomous parking method is used for realizing the autonomous parking system.

Description

technical field [0001] The invention relates to the technical field of vehicle autonomous parking, in particular to an autonomous parking system and method based on multi-visual inertial navigation fusion. Background technique [0002] Under the background of such rapid popularization of automobiles, the intelligentization of automobiles has become a major trend in the development of the automobile industry. Due to the cumbersome parking process, high technical requirements for drivers, and accidents in the parking process, the autonomous parking system is even more Attract major automobile manufacturers to invest a lot of money in research and development. So far, most of the autonomous parking systems equipped with high-end models on the market rely on radar and ultrasonic sensors to achieve the purpose of parking by obtaining the distance between the car and surrounding objects through radar and ultrasonic sensors. However, both radar and ultrasonic sensors have certain ...

AI Technical Summary

Problems solved by technology

However, both radar and ultrasonic sensors have certain blind spots and cannot completely collect data

At the same time, such sensors are expensive and are generally only equipped on high-end models of high-end brands, which cannot play a positive role in the popularization of smart cars

[0003] At the same time, due to the rapid increase in the number of vehicles, it is difficult to find a parking space in the city during busy hours. The process of finding a parking space is boring and extremely troublesome. Unfortunately, most of the autonomous parking systems on the market require a driver to The driver finds the parking space by himself and drives the car near the space to realize self-parking. Some self-parking systems also require the driver to mark the parking space

very inhumane

[0004] In addition, visual SLAM technology is rarely used in the market to provide positioning information for autonomous parking systems. It is difficult for visual SLAM technology to track and calculate the current position information stably. The scene and speed requirements are extremely high, and it is not suitable for truly changeable parking. surroundings

Images