Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A semantic SLAM object association method based on a hierarchical topic model

A technology of hierarchical topics and objects, applied in character and pattern recognition, image data processing, instruments, etc., can solve problems such as unreliability, achieve accurate object association, and promote the effect of camera pose estimation

Active Publication Date: 2019-03-29
ZHEJIANG UNIV OF TECH
View PDF6 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But in practice, with the movement of the robot, the information captured by the sensor is noisy, and it is unreliable to estimate the movement of the robot only by the information of the sensor, so it needs the assistance of various optimization algorithms

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 semantic SLAM object association method based on a hierarchical topic model
  • A semantic SLAM object association method based on a hierarchical topic model
  • A semantic SLAM object association method based on a hierarchical topic model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0030] The present invention will be further described below in conjunction with the accompanying drawings.

[0031] refer to Figure 1 ~ Figure 3 , a semantic SLAM object association method based on a hierarchical topic model, comprising the following steps:

[0032] 1) Perform internal reference calibration on the camera to obtain the camera's distortion parameters and internal reference matrix

[0033]

[0034] Among them, [x, y] are the coordinates of the normalized plane point, [x distorted ,y distorted ] is the distorted coordinate, k 1 , k 2 , k 3 ,p 1 ,p 2 is the distortion term;

[0035]

[0036] P is the internal reference matrix of the camera, where f is the focal length of the camera, [O x , O y ] is the main optical axis point;

[0037] 2) Use Single Shot MultiBox Detector (SSD) and Convolutional NeuralNetwork (CNN) to build a deep learning network, train a deep learning model, and complete object recognition and object pose prediction tasks;

[...

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

A semantic SLAM object association method based on a hierarchical topic model uses a depth learning model to detect objects in key frames and predict their positions and postures, When processing eachframe object, the Gibbs sampling method is used to sample the real environment object set with potential associated object according to the principle of view overlap, and the object association method is used to calculate for each object in the current frame, and the maximum posteriori probability is used to judge whether the object is associated or not. Factor maps are constructed for objects, cameras, and map points, and observations between them are used as edges to optimize object posture, camera posture, and map point positions. Finally, a complete semantic map including object information and camera trajectory is constructed. The invention can realize object association with high precision and avoids redundant object association. It can promote the camera pose estimation of semanticSLAM, and the optimized object pose can make the object association more accurate, so as to build a more accurate semantic map.

Description

technical field [0001] The invention relates to technical fields such as robot vision, deep learning, and statistics, and specifically relates to a semantic SLAM object association method based on a hierarchical topic model. Background technique [0002] Simultaneous localization and mapping (SLAM) is an important issue in robot applications, such as autonomous driving, autonomous navigation and other fields. Constructing an accurate environmental map is a specific form of SLAM. Traditional SLAM technology relies on low-level geometric features, such as points, lines, and surfaces. This technology is prone to failure in open or repetitive texture environments. Semantic SLAM utilizes the advanced semantic information in the environment, which can effectively make up for the shortcomings of traditional SLAM, and can create a readable and more application-worthy semantic map. [0003] Object association and object pose optimization are two crucial components in semantic SLAM. ...

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
IPC IPC(8): G06T7/73G06T7/80G06K9/00
CPCG06T7/73G06T7/80G06T2207/10004G06T2207/20081G06T2207/20084G06T2207/30244G06V20/10
Inventor 张剑华贵梦萍王其超刘儒瑜徐浚哲陈胜勇
Owner ZHEJIANG UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com