Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Underwater robot target tracking method based on deep learning and monocular vision

An underwater robot, deep learning technology, applied in instruments, image data processing, computing and other directions, can solve problems such as difficulties

Active Publication Date: 2019-09-17
NANJING INST OF TECH
View PDF3 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This can cause difficulties in tasks such as segmentation, tracking, or classification because they use color indirectly or directly

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
  • Underwater robot target tracking method based on deep learning and monocular vision
  • Underwater robot target tracking method based on deep learning and monocular vision
  • Underwater robot target tracking method based on deep learning and monocular vision

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0123] (1) Deep learning architecture

[0124] Object recognition based on deep learning outperforms traditional machine learning with matching big data. Deep learning is a convolutional neural network (CNN) that has deep layers consisting of convolutional, round-robin, and fully-connected layers.

[0125] The most commonly used structure of neural network is figure 2 As shown, it consists of three layers, called input layer, hidden layer, and output layer, and each layer consists of one or more nodes represented by small circles. Narrow lines between nodes indicate the flow of information from one node to the next. The output layer has four nodes, four classes in the case of object classification. The nodes in the hidden and output layers are called active nodes, while the nodes in the input layer are called passive nodes. Every value from the input layer is copied and sent to all hidden nodes. Known as a fully interconnected structure. Such as image 3 As shown, the ...

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 belongs to the technical field of underwater robots, and discloses an underwater robot target tracking method based on deep learning and monocular vision, which comprises the following steps of inputting the images from a video sequence, estimating an underwater transmission map by using a deep learning neural network for each input image, and determining a target orientation; and establishing a direction and a control scheme of the target motion estimation through a transmission graph obtained through a network. According to the novel underwater robot monocular vision target tracking method based on deep learning, the transmission of an underwater image is calculated by a monocular image acquisition method in an underwater environment. For each incoming video frame and an environment without the priori knowledge, a previously trained convolutional neural network is creatively introduced to compute a transmission map, which provides the depth-dependent estimation. According to the method provided by the invention, the target area can be found, and a tracking direction is established.

Description

technical field [0001] The invention belongs to the technical field of underwater robots, in particular to an underwater robot target tracking method based on deep learning and monocular vision. Background technique [0002] Currently, the closest prior art: [0003] Underwater robots have been widely used in various underwater tasks, such as maintenance and inspection of underwater structures, installation of sensors, and sample retrieval for scientific exploration, etc. These tasks are usually performed by human-controlled remote control, so operations rely on human perception (mainly Vision), the operating conditions, experience and skills of the operator seriously affect the quality of operation. Therefore, in order to ensure the accuracy of underwater work, automatic control is required, and high-precision underwater target tracking is a key requirement. [0004] There are many attempts to use vision sensors for underwater navigation, however, the proposed navigation ...

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): G06T7/20
CPCG06T7/20G06T2207/10016G06T2207/20081G06T2207/20084
Inventor 陈国军陈巍
Owner NANJING INST 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
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