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

Autonomous obstacle avoidance method based on deep learning and stereoscopic vision

A technology of deep learning and obstacle avoidance, applied in neural learning methods, character and pattern recognition, two-dimensional position/channel control, etc., can solve complicated 3D reconstruction and path planning problems, and avoid 3D reconstruction and path planning problems , Improve the effect of accuracy and speed

Inactive Publication Date: 2021-06-25
NANJING UNIV OF SCI & TECH
View PDF5 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide an autonomous obstacle avoidance method based on deep learning and stereo vision, which takes deep learning algorithm as the core, integrates the depth information of traditional machine vision, and solves the problem that a single deep learning algorithm can only obtain two-dimensional coordinate information Insufficient to obtain complete 3D coordinates, realize end-to-end learning of obstacle avoidance scenarios, and avoid complicated 3D reconstruction and path planning problems

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
  • Autonomous obstacle avoidance method based on deep learning and stereoscopic vision
  • Autonomous obstacle avoidance method based on deep learning and stereoscopic vision
  • Autonomous obstacle avoidance method based on deep learning and stereoscopic vision

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach

[0027] Step 4. On the basis of the autonomous obstacle avoidance model, the depth information of obstacles is fused to realize obstacle detection in three-dimensional space, obtain the three-dimensional coordinates of obstacles, and solve the problem that a single deep learning algorithm can only obtain two-dimensional coordinate information. , the specific implementation is as follows:

[0028] In order to ensure the one-to-one mapping relationship between the detection image and the depth image of the obstacle avoidance scene in two-dimensional space, coordinate decoding is performed on the detection image to obtain the detection result under the left eye view of the original size, and the coordinate information of the obstacle area is obtained according to the above detection result , obtain the depth information of the obstacle through the coordinate index, and use the minimum depth in the area as the distance between the obstacle and the robot, and then obtain the three-di...

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 autonomous obstacle avoidance method based on deep learning and a binocular stereoscopic vision technology. Environment video information is collected in real time through a binocular camera, an obstacle three-dimensional detection algorithm is constructed based on the target detection in deep learning in combination with the binocular stereoscopic vision technology, and a robot performs autonomous obstacle avoidance according to an obtained obstacle avoidance steering decision, the three-dimensional coordinates of an obstacle and the obstacle avoidance angle. According to the method, deep learning target detection serves as a core, depth information of traditional machine vision is fused, the defect that only two-dimensional coordinate information can be obtained through a single deep learning algorithm is overcome, and complete three-dimensional coordinates are obtained. According to the invention, end-to-end learning in an obstacle avoidance scene is realized, complicated three-dimensional reconstruction and path planning problems are avoided, and the system cost is saved.

Description

technical field [0001] The invention belongs to the field of robot obstacle avoidance, in particular to an autonomous obstacle avoidance method based on deep learning and stereo vision. Background technique [0002] With the continuous development of science and technology, mobile robots have been widely used in various fields, including life services, industrial production, military, entertainment and so on. Robotics design control, machinery, computer and other disciplines. The robot's navigation and obstacle avoidance capabilities are important indicators that reflect the intelligence of the robot. [0003] In the article "Real-time Obstacle Avoidance for Manipulator AndMobileRobots", Khatib builds an artificial potential field, so that obstacles and target points can generate abstract repulsion and attraction to the robot in the artificial potential field, and jointly control the robot to avoid obstacles. The path planned by the artificial potential field method has th...

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): G05D1/02G06T7/73G06T7/80G06N3/04G06N3/08G06K9/46
CPCG05D1/0251G06T7/85G06T7/73G06N3/08G05D1/0223G05D1/0214G05D1/0221G05D1/028G05D1/0276G06T2207/10012G06T2207/20081G06V10/44G06N3/045
Inventor 焦浩李远平何博侠张鹏辉田德旭
Owner NANJING UNIV OF SCI & 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