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

Monocular unsupervised depth estimation method based on CBAM

A depth estimation, unsupervised technology, applied in neural learning methods, computing, computer components, etc., can solve problems such as restricting generalization ability, and achieve the effect of retaining depth details and improving depth estimation accuracy.

Pending Publication Date: 2021-06-11
SOUTHEAST UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, supervised learning requires a large number of groundtruth data samples for training, which greatly restricts its generalization ability

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
  • Monocular unsupervised depth estimation method based on CBAM
  • Monocular unsupervised depth estimation method based on CBAM
  • Monocular unsupervised depth estimation method based on CBAM

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] Below in conjunction with accompanying drawing and specific embodiment, further illustrate the present invention, it should be understood that following specific embodiment is only for illustrating the present invention and is not intended to limit the scope of the present invention.

[0053] A CBAM-based monocular unsupervised depth estimation method described in the present invention, such as figure 1 As shown, firstly, CBAM is combined with Resblock to form Resblock-CBAM, then a depth estimation network with attention mechanism is designed based on Resblock-CBAM, and finally the depth estimation network is implemented for photometric reconstruction, parallax smoothing and left-right parallax consistency of stereo image pairs. Training, and complete the depth estimation of the monocular image; including the following specific steps:

[0054] Step 1), introducing CBAM and combining Resblock into Resblock-CBAM, including the following specific steps:

[0055] a), set t...

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 monocular unsupervised depth estimation method based on CBAM. Depth estimation is one of key technologies for a robot to realize surrounding environment perception, a distance measurement value obtained by a laser radar and other sensors is processed by a depth estimation method based on supervised learning and then is used as a true value for training, but the process occupies a large amount of manpower and computing resources, so that the application of the depth estimation in cross-scene is limited. Under a depth estimation framework based on unsupervised learning, a convolution block attention module is introduced, photometric reconstruction, parallax smoothing and left and right parallax consistency training of a stereo image pair are carried out, and scale depth estimation is carried out on a monocular image. By using the method provided by the invention, the depth details of the objects in the surrounding environment can be reserved, the overall depth estimation precision is improved, and the generalization ability under the cross-scene condition can also be guaranteed.

Description

technical field [0001] The invention belongs to the field of intelligent body autonomous navigation and environment perception, in particular to a CBAM-based monocular unsupervised depth estimation method. Background technique [0002] In order to achieve safe and reliable autonomous navigation, the agent needs to have a complete environment perception function, which includes the depth estimation of the environment around the agent. Depth estimation based on 3D lidar can obtain more accurate depth estimation results, but it is expensive and can only obtain sparse depth estimation. Depth estimation based on RGBD camera is simple to operate, but the range of depth estimation is limited and its use in outdoor environments is limited. Depth estimation based on stereo cameras is not limited for indoor and outdoor use, but it occupies a large amount of computing resources, and the range of depth estimation is limited due to the short baseline. Depth estimation based on monocula...

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): G06T7/593G06K9/62G06N3/04G06N3/08
CPCG06T7/593G06N3/08G06T2207/10028G06T2207/20081G06T2207/20084G06N3/045G06F18/214G06F18/25
Inventor 潘树国魏建胜高旺赵涛
Owner SOUTHEAST UNIV
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