Unlock instant, AI-driven research and patent intelligence for your innovation.

Intelligent automobile-oriented traffic scene semantic modeling device and modeling method and positioning method

A traffic scene and smart car technology, applied in the field of smart cars, can solve problems such as dynamic target interference, achieve the effect of solving the difficulty of extracting special assistance, improving map construction and positioning efficiency, and improving positioning accuracy and efficiency

Pending Publication Date: 2021-09-21
JIANGSU UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Among them, the traffic semantic layer uses deep learning to carry out semantic recognition of traffic elements, and removes dynamic objects such as pedestrians and vehicles in the scene to solve the problem of dynamic object interference; the road position layer describes the positional relationship between scenes; scene features layer is to fully describe the traffic scene on the basis of minimizing data storage
The above three solve the problem of fine description of traffic scenes by high-precision maps

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Intelligent automobile-oriented traffic scene semantic modeling device and modeling method and positioning method
  • Intelligent automobile-oriented traffic scene semantic modeling device and modeling method and positioning method
  • Intelligent automobile-oriented traffic scene semantic modeling device and modeling method and positioning method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] Embodiment 1. Smart car-oriented traffic scene semantic modeling device

[0056] The device can accurately collect traffic scenes, and after calibration with the calibration board, it can realize multi-sensor data fusion and finally realize multi-level and multi-scale semantic modeling of traffic scenes, which has the characteristics of reducing storage space and improving modeling accuracy. The schematic diagram of the system structure is figure 1 As shown, it includes a multi-source heterogeneous data acquisition system, a multi-source heterogeneous sensor calibration and fusion system, and a feature processing system. Among them, the multi-source heterogeneous data acquisition system includes three lidars 1, 2, 3, Beidou system 4, differential Beidou base station 8 and inertial navigation system 5. The lidars 1, 2, and 3 can be installed at any position outside the smart car. For example, they are installed at the front, middle, and rear positions of the vehicle in...

Embodiment 2

[0057] Embodiment 2. Traffic scene semantic modeling method for smart cars

[0058] The method includes data collection and multi-scale semantic traffic scene modeling, and the flow chart of the method is as follows figure 2 As shown, the specific steps are as follows:

[0059] (1) Data collection

[0060] (1.1) For data collection, drive the smart car to any road, turn on all sensor systems, and ensure that the relative positions of the laser radar and the Beidou system remain unchanged during data collection.

[0061] (1.2) Place the calibration board on the opposite side of the laser radar, so that the three laser radars scan the calibration board at the same time, transmit the data to the industrial computer, fit the plane through the industrial computer, and then calculate the plane equation of the calibration board, The plane equations of the three laser radars are shown in the following formulas:

[0062] a" 1 x+b″ 1 y+c″ 1 z+d″ 1 =0

[0063]

[0064]

[...

Embodiment 3

[0081] Embodiment 3. High-precision positioning method based on traffic scene semantic modeling

[0082] This embodiment applies the constructed traffic scene semantic model, and its main application field is high-precision positioning. The high-precision positioning process is as follows: Figure 7 shown.

[0083] (1) Turn on the smart car sensor to collect Beidou information and laser point cloud information.

[0084] (2) Extract the road location layer in the semantic model and match it with the Beidou information to obtain the nearest node, and draw a circle with the node as the center and r as the radius. The circle is the positioning range, and all nodes in the circle are candidates point.

[0085] (3) Extract the semantic layer in the semantic model, match it with the laser point cloud in the smart car, and identify and match it through the Tri-net neural network to obtain the nearest positioning node.

[0086] (4) Extract the scene feature layer in the model, and ma...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an intelligent automobile-oriented traffic scene semantic modeling device and modeling method and a positioning method. According to the intelligent automobile-oriented traffic scene semantic modeling device and modeling method, aiming to realize high modeling precision, multi-scale and multi-level refined representation is performed on a traffic scene; the traffic scene semantic modeling device comprises a road position layer, a scene feature layer and a traffic semantic layer, wherein the traffic semantic layer adopts a deep learning mode to perform semantic recognition on traffic elements, and eliminates dynamic targets such as pedestrians and vehicles in a scene, so that the problem of dynamic target interference is solved; the road position layer describes the position relation between scenes; the scene feature layer fully describes a traffic scene on the basis of reducing data storage to the maximum extent. With the above three layers adopted, the problem of fine description of a traffic scene by a high-precision map can be solved.

Description

technical field [0001] The invention belongs to the technology of smart cars, and in particular relates to a traffic scene semantic modeling device, a modeling method and a positioning method for smart cars. Background technique [0002] With the advancement of science and technology, smart cars have gradually become a hot topic of research at home and abroad. High-precision maps in this field are one of the most critical issues in realizing vehicle intelligence. High-precision map construction is the basis for high-precision positioning, environmental perception, decision-making planning, and execution control. The construction of this map is different from ordinary maps. Ordinary maps only need to provide high-precision latitude and longitude information. However, only providing latitude and longitude information cannot meet the driving needs of smart cars. Traffic scenes need to be described on the basis of ordinary maps. Sufficient or not directly determines the accura...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G01C21/32G06T17/05
CPCG01C21/32G06T17/05G06T2200/08
Inventor 李祎承陆子恒蔡英凤王海朱镇蒋卓一冯锋
Owner JIANGSU UNIV