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

Robot, monocular depth estimation method and system and storage medium

A technology of depth estimation and storage medium, applied in neural learning methods, instruments, computing, etc., can solve problems such as large field of view and inability to take into account the depth of images, and achieve the effect of saving computing power

Pending Publication Date: 2021-04-27
SHANGHAI MROBOT TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In order to solve the technical problem that the depth of the image cannot be fast, cheap, and have a large field of view in the mobile control of the robot in the prior art, the present invention provides a robot, a monocular depth estimation method, system and storage medium, specifically The technical scheme is as follows:

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
  • Robot, monocular depth estimation method and system and storage medium
  • Robot, monocular depth estimation method and system and storage medium
  • Robot, monocular depth estimation method and system and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053]Such asfigure 1 As shown, the present embodiment discloses a single depth estimation method, including:

[0054]S1: Get a single-graphic image and a single line laser depth data;

[0055]S2: Import the single-graph image and the single-line laser depth data into a pre-training estimate model and output depth maps; the estimation model is based on the high-resolution net.

[0056]In conventional techniques, the RGB-D camera or binocular camera is generally used for the consideration of avoidance obstacles and environmental interactions, however the RGB-D is usually small in the process of actual use, while the depth of the binary cameras It is estimated that the computational complexity is high, and there is a large strength, resulting in a poor effect in the actual use.

[0057]In this embodiment, the use of single-line laser binding single-graph images, using a high resolution neural network model to achieve construction of depth maps, on the one hand, the field of view is much higher th...

Embodiment 2

[0061]This embodiment discloses a single-optical depth estimate, including:

[0062]S0: Training the estimation model:

[0063]S1: Get a single-graphic image and a single line laser depth data;

[0064]S2: Import the single-graph image and the single-line laser depth data into a pre-training estimate model and output depth maps; the estimation model is based on the high-resolution net.

[0065]Specifically include:

[0066]S0-1: Initialize the estimation model;

[0067]S0-2: Get a single-line laser depth data set, a single-graph image data set and depth map data set; the single-line laser depth data set, the single-graph image data set and the elements in the depth map data set correspond;

[0068]S0-3: Import the single-line laser depth data set and the individual image data set into the estimation model; the single-line laser depth data is an anchor point in the vertical direction;

[0069]S0-4: The output result set of the estimated model is set to the depth map data set and generate a deviation evaluat...

Embodiment 3

[0075]Such asfigure 2 As shown, the present embodiment discloses a single depth estimation method, including:

[0076]S0-1: Initialize the estimation model;

[0077]S0-2: Get a single-line laser depth data set, a single-graph image data set and depth map data set; the single-line laser depth data set, the single-graph image data set and the elements in the depth map data set correspond;

[0078]S0-3: Import the single-line laser depth data set and the single-graph image data set into the estimation model;

[0079]S0-4: The output result set of the estimated model is set to the depth map data set and generate a deviation evaluation result, and if the deviation evaluation result satisfies the threshold, enter S0-5, otherwise enter S1;

[0080]S0-5: Correct the estimation model based on the deviation evaluation result, returning to S0-3;

[0081]S1: Get a single-graphic image and a single line laser depth data;

[0082]S2-1: Use the estimated model to classify the single-graph image, divide the content of ...

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 robot, a monocular depth estimation method, a monocular depth estimation system and a storage medium. The monocular depth estimation method comprises the steps of acquiring a monocular image and single-line laser depth data; importing the monocular image and the single-line laser depth data into a pre-trained estimation model and outputting a depth map, wherein the estimation model is constructed on the basis of the High Solutions Net. The method has the technical effects that the single-line laser is combined with the monocular image, the high-resolution neural network model is used for realizing the construction of the depth map, and the technical effect of quickly constructing the depth map with a wide visual field range under the condition of not increasing the local computing power is achieved.

Description

Technical field[0001]The present invention relates to the field of robots, and in particular, to a robot and a single-optical depth estimate, system, and storage medium.Background technique[0002]In conventional techniques, the RGB-D camera or binocular camera is generally used for the consideration of avoidance obstacles and environmental interactions, however the RGB-D is usually small in the process of actual use, while the depth of the binary cameras It is estimated that the computational complexity is high, and there is a need for more integration force, resulting in a poor effect in the actual use, and if it is necessary to improve the use of binary cameras, it will need more The power is meant to be more expensive.Inventive content[0003]To solve the technical problem of the image of the image in the mobile control in the prior art, it is not possible to simultaneously balance the technical problem of rapid, cheap, and visual fields, and the present invention provides a robot a...

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): G06K9/00G01B11/22G06K9/62G06N3/08
CPCG01B11/22G06N3/08G06V20/10G06F18/24
Inventor 王增
Owner SHANGHAI MROBOT TECH CO LTD
Features
  • Generate Ideas
  • 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