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

Image positioning method based on data modal missing and Embedded Net

An image positioning and modality technology, applied in the field of positioning and navigation technology and machine learning applications, can solve the problems of lack of multi-modal data, it is difficult to ensure the robustness of the positioning performance of the solution, and avoid the reduction of positioning accuracy and overcome the problem of signal attenuation , the effect of broad application prospects

Pending Publication Date: 2021-12-21
NANJING UNIV OF POSTS & TELECOMM
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For the lack of multimodal data that is common in most current deep learning network models, the common solution is to reuse the obtained values, use default values ​​or use interpolation methods to fill in missing data, but such When the processing method is applied to indoor positioning, it is difficult to ensure the robustness of the positioning performance of the solution

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
  • Image positioning method based on data modal missing and Embedded Net
  • Image positioning method based on data modal missing and Embedded Net
  • Image positioning method based on data modal missing and Embedded Net

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] Such as figure 1 As shown, the present invention discloses an image localization method based on data mode loss and Embrace Net, which consists of two stages of offline modeling and online positioning, as follows.

[0036] The offline modeling phase includes the following steps:

[0037] S1, using the camera to collect training images (ie original images), classify the training images after image preprocessing, and use them as training data sets;

[0038] S2. Perform feature extraction and feature fusion on the training images in the training data set, and then perform offline classification learning to obtain a feature extraction network, feature fusion network and position classification model with optimal parameters.

[0039] Further, said S1 includes the following steps:

[0040] S11. Divide the area to be positioned, determine multiple reference points in the area to be positioned, and use multiple cameras to separately collect training images of the target at ea...

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 positioning method based on data modal missing and Embedded Net, which is composed of two stages of off-line modeling and on-line positioning, and comprises the following steps: S1, collecting training images, preprocessing the training images, carrying out category division, and taking the training images as a training data set; S2, extracting features of the training images and fusing with the features; performing offline classification learning to obtain a feature extraction network, a feature fusion network and a position classification model of optimal parameters; S3, collecting a target image, and performing image preprocessing on the target image; and S4, performing feature extraction and feature fusion on each target image, and then obtaining a position estimation value of the target. The multi-image measurement technology and the machine learning technology are combined, target positioning for the indoor complex environment is achieved, follow-up positioning work is conducted through the images collected by the camera, the problem of signal attenuation is solved, and the implementation cost of the scheme is reduced.

Description

technical field [0001] The invention relates to an indoor target positioning method, in particular to an image positioning method based on data mode loss and EmbraceNet, and belongs to the field of positioning navigation technology and machine learning applications. Background technique [0002] In recent years, with the development and popularization of computer communication technology and mobile smart terminals, people are increasingly relying on their electronic devices to obtain their own physical location information, thereby satisfying various service requirements based on location information. [0003] At this stage, the development of positioning technology for outdoor environments is very mature, and various satellite navigation systems are widely used, such as the Beidou satellite navigation system independently developed by my country, GPS in the United States, GLONASS in Russia, and GALILEO in Europe. Although the above-mentioned satellite navigation systems gen...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/214G06F18/253
Inventor 颜俊朱洪柳曹艳华
Owner NANJING UNIV OF POSTS & TELECOMM
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