Supercharge Your Innovation With Domain-Expert AI Agents!

A method and device for dynamic target 3D detection

A dynamic target, 3D technology, applied in the application field of computer vision technology, can solve problems such as low accuracy, environmental perception information cannot meet the requirements of automatic driving scene perception, etc., to achieve guaranteed effects, reduce time and space complexity, and reduce complexity Effect

Active Publication Date: 2020-05-29
WUHAN ZHONGHAITING DATA TECH CO LTD
View PDF8 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The environmental perception information provided by 2D object detection cannot meet the perception requirements of autonomous driving scenarios
At present, the accuracy of pure visual 3D object detection is relatively low.

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
  • A method and device for dynamic target 3D detection
  • A method and device for dynamic target 3D detection
  • A method and device for dynamic target 3D detection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0076] An embodiment of the present invention provides a method for dynamic target 3D detection, such as figure 1 shown, including the following steps:

[0077] S1, synchronously acquire camera image data and lidar point cloud data through time stamp alignment, and cut the lidar point cloud data;

[0078] S2, performing voxel rasterization processing on the clipped point cloud data and projecting it into a 6-channel bird's-eye view;

[0079] S3, performing feature extraction on the image data and the 6-channel bird's-eye view image, and acquiring image data feature layers and bird's-eye view feature layers;

[0080] S4, combined with the prior 3D anchor frame, select the candidate area in the image data feature layer and the bird's-eye view feature layer, extract the feature vector of the candidate area and fuse it, and generate a 3D candidate anchor frame;

[0081] S5, using the 3D candidate anchor frame to respectively extract the feature crops of the image feature layer a...

Embodiment 2

[0133] An embodiment of the present invention provides a device for dynamic target 3D detection, including:

[0134] The data acquisition module acquires camera image data and lidar point cloud data synchronously through time stamp alignment, and cuts the lidar point cloud data;

[0135] A rasterization processing module, which performs voxel rasterization processing on the cut point cloud data and projects it into a 6-channel bird's-eye view;

[0136] The feature layer generation module extracts the features of the image data and the 6-channel bird's-eye view, and obtains the image data feature layer and the bird's-eye view feature layer;

[0137] The 3D candidate anchor box generation module fuses the image data feature layer and the bird's-eye view feature layer, and combines the prior 3D anchor box to generate a candidate area and obtain a 3D candidate anchor box;

[0138]A feature vector generation module, using the 3D candidate anchor frame to extract the feature crops ...

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 relates to a method and a device for dynamic target 3D detection. The method comprises the following steps: firstly, point cloud data of a specific area is acquired in a point cloud cutting manner; and a special aerial view projection mode is adopted to reduce the complexity of the point cloud data and improve the calculation efficiency; down-sampling is carried out in a mode of combining Depth-wise convolution and Point-wise convolution, so that the time and space complexity of convolution operation is reduced while the feature extraction effect is ensured; up-sampling is performed on the obtained feature vectors through a feature pyramid network to obtain a high-representation performance feature layer with the same size as the original image. The application of the feature pyramid network enhances the feature extraction effect of the small object. The feature extraction network not only reduces the time and space complexity of feature extraction, but also ensures theeffect of feature extraction, and greatly improves the precision of small object detection and more accurate position information.

Description

technical field [0001] The invention relates to the application field of computer vision technology, in particular to a method and a device for detecting 3D dynamic objects (such as automobiles and pedestrians). Background technique [0002] Computer vision refers to the extraction of surrounding environmental information from the computer by processing the raw data of various sensors such as cameras, lidar, and millimeter-wave radar. Its function is similar to the human visual nervous system, which recognizes all existing objects in the environment and determines their position, size, orientation, speed and other information. [0003] Computer vision is of great use in the field of self-driving cars and mobile robotics. Computer vision technology can provide self-driving cars with information about the environment around the vehicle, such as the position, orientation, and speed of roads, obstacles, vehicles, and pedestrians. These perceptual information can provide powerf...

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): G06K9/00G06K9/32G06K9/62G06T7/80G06T7/90
CPCG06T7/85G06T7/90G06T2207/10028G06V20/64G06V10/25G06F18/23213G06F18/214
Inventor 漆梦梦陶靖琦杨贵施忠继刘奋
Owner WUHAN ZHONGHAITING DATA TECH CO LTD
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com