An image cloud computing method and system based on online deep learning slam

A deep learning and cloud computing technology, applied in the field of image processing research, can solve problems such as imperfection, low sensor accuracy, and time-consuming, to achieve the effect of improving efficiency and accuracy, improving training efficiency, and reducing training time

Active Publication Date: 2022-03-25
SOUTH CHINA UNIV OF TECH
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the existing technology, due to problems such as low sensor accuracy and large calculation load, it will take a lot of time, and it is not perfect, and the effect is not very ideal. The development of SLAM based on three-dimensional vision has encountered certain resistance

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
  • An image cloud computing method and system based on online deep learning slam
  • An image cloud computing method and system based on online deep learning slam
  • An image cloud computing method and system based on online deep learning slam

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0055] An image cloud computing method based on online deep learning SLAM is as follows figure 1 shown, including the following steps:

[0056] The first step: the image data acquisition layer obtains the RGBD image and the depth image through the RGBD camera, collects the image data, and uses the image stream of the streaming media server to store the image data in the memory;

[0057] The second step: extract key frames from the image data in the memory, and upload the key frames to the cloud computing platform;

[0058] Step 3: Construct a dataset from the historical data on the cloud computing platform, use MapReduce to train the convolutional neural network to train the dataset, and obtain the optimal convolutional neural network parameters;

[0059] The MapReduce training convolutional neural network trains the dataset, specifically: the input stage: the data to be processed is divided into fixed-size segments, and each segment is further decomposed into key-value pairs...

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 image cloud computing method based on online deep learning SLAM, comprising the following steps: collecting and storing image data; extracting and uploading key frames; constructing a data set from image data and performing training to obtain optimal convolutional neural network parameters ; Extract real-time image feature points for identification, and perform feature point matching on adjacent frame images; the image feature points are iterated to obtain the best matching transformation matrix, and the position and attitude information is used to correct it to obtain the camera pose transformation; through point cloud data Registration and position and posture information to obtain the optimal pose estimation; transform the pose information to a coordinate system through matrix transformation to obtain map information; repeat the previous steps for areas with insufficient precision; the client displays the results and performs online adjustment at the same time; The invention parallelizes image processing, deep learning training and SLAM using cloud computing technology to improve the efficiency and accuracy of image processing, positioning and mapping.

Description

technical field [0001] The invention relates to the field of image processing research, in particular to an image cloud computing method and system based on online deep learning SLAM. Background technique [0002] At present, with the development of mobile robots, people's demand for them has gradually increased, such as: unmanned driving, sweeping robots, 3D printing, criminal investigation scene records, etc., which greatly facilitates people's lives, but at the same time there are also emerging some new questions. In the prior art, due to the problems of low sensor accuracy and large amount of calculation, it will take a lot of time, and it is not perfect, and the effect is not very satisfactory. The development of SLAM based on 3D vision has been hindered to a certain extent. [0003] In recent years, deep learning has developed rapidly and achieved good results in chess games and some simulation games. The emergence of cloud computing makes it possible to collect and ...

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 Patents(China)
IPC IPC(8): G06T7/73G06V10/764G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06T7/73G06T2207/20081G06T2207/20084G06N3/045G06F18/24
Inventor 李迪楚英王世勇杨啸
Owner SOUTH CHINA UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap