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

G2o and random ferns-based dense map creation method

A map creation, random fern technology, applied in image analysis, mapping and navigation, image data processing, etc., can solve the problems of accumulation, unable to build a globally consistent map, achieve good accuracy and robustness, reduce system The amount of computation and the effect of increasing the correction accuracy

Active Publication Date: 2018-06-08
BEIJING UNIV OF TECH
View PDF6 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Since the original TSDF model only considers the correlation in adjacent time, the error will inevitably accumulate to the next moment, and it is impossible to build a globally consistent map

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
  • G2o and random ferns-based dense map creation method
  • G2o and random ferns-based dense map creation method
  • G2o and random ferns-based dense map creation method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0036] The invention is based on a Kinect sensor, collects data of the experimental environment through a map creation algorithm, and obtains corresponding depth maps and color maps. Optimize the depth map and calculate the rotation and translation matrix from the collected data information to create a camera pose model. The observation state with the smallest error is obtained by constructing the iterative function of g2o, using the GPU to process the ICP matching projection algorithm in parallel, constructing a model-to-model matching strategy, and using random ferns for closed-loop matching. Establish a TSDF model to represent the surface of a three-dimensional object and perform multi-frame fusion. The overall flow of the method involved is attached figure 1 As shown, the specific implementation process is divided into the following steps...

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 a g2o and random ferns algorithm-based dense map creation method, and belongs to the field of robot real-time localization and map creation. The method comprises the steps of firstly, building a camera attitude model and a TSDF model of weighted fusion 3D point truncation information, wherein the models are used for accurately representing the surface of a created object; secondly, proposing an improved loop closure detection method, and combining the method with a random ferns color image codification strategy, thereby optimizing the TSDF model; and finally, resolvinga constraint function by using a g2o graph optimization library, and establishing optimization edges between data sets. An experimental result shows that a mixed optimization attitude model can establish a global SLAM map more quickly and effectively identify reached regions.

Description

technical field [0001] The invention belongs to the field of robot real-time positioning and map creation. Use g2o to optimize the local and global loop detection tasks, use the random fern algorithm for closed loop matching, and build a TSDF (Truncated Signed Distance Function, TSDF) model. Background technique [0002] The present invention mainly solves the problem of robot V-SLAM (Vision-only Simultaneous Localization and Mapping, V-SLAM) using visual sensors in an indoor environment. The main difficulty of SLAM is that precise map creation relies on an accurate estimate of the robot's position, which in turn comes from the sensor's perception of the outside world, that is, the landmarks in the map (Landmark). Therefore, "localization" and "mapping", as two coupled and interdependent problems, are susceptible to each other's noise. In addition, common phenomena in the SLAM process, such as Data Association, Loop Closure Detection, Global Optimization, etc., are also im...

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): G06T7/73G06T7/33G01C21/20
CPCG01C21/206G06T7/344G06T7/75
Inventor 贾松敏李柏杨张国梁李秀智张祥银
Owner BEIJING UNIV OF TECH