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

A Global Search Method for Template Image Based on Mask Matrix and Fast Fourier Transform

A technology of Fourier transform and template image, which is applied in the field of image processing, can solve the problems of high computational complexity, achieve the effects of overcoming complexity, saving processing time, and improving computing speed

Active Publication Date: 2019-06-21
XIDIAN UNIV
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The present invention is to overcome the deficiency that the computational complexity of the sweeping window method is too high, and proposes a template image global search method based on mask matrix and fast Fourier transform

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
  • A Global Search Method for Template Image Based on Mask Matrix and Fast Fourier Transform
  • A Global Search Method for Template Image Based on Mask Matrix and Fast Fourier Transform
  • A Global Search Method for Template Image Based on Mask Matrix and Fast Fourier Transform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0049] In view of the problems that the template search method in the field of digital image processing has too high algorithm complexity, repeated calculations cannot be used multiple times, and certain calculation accuracy is lost, the present invention proposes a template image global search based on mask matrix and fast Fourier transform method, see figure 1 , including the following steps:

[0050] (1) Copy the template map and full map: read the original template map and full map data, where the original template map is as follows image 3 , whose size is smaller than the full image, which is the image of lena in the gallery. See figure 2 .

[0051] (2) Use the original template image to generate a mask image:

[0052] (2a) Generate a template with the original Figure 1 The full 1 matrix of sample size is the mask map;

[0053] (2b) When it is not necessary to match the target background, the element at the position representing the background area in the mask mat...

Embodiment 2

[0079] The template image global search method based on mask matrix and fast Fourier transform is the same as embodiment 1, and the template expansion graph described in step (6) and the full graph are used to formulate as follows to do circular convolution:

[0080]

[0081] Among them, F -1 Represents inverse fast Fourier transform, C stands for conjugation, F stands for fast Fourier transform, subscript 2 stands for two-dimensional transformation, t_ext stands for template expansion map, img stands for full map, and corr is the output template full map correlation matrix.

[0082] In the actual project, it can be implemented in matlab language. If it is implemented in matlab, the above formula can be expressed in functional form as:

[0083] corr=ifft2(conj(fft2(t_ext)).*fft2(img))

[0084] Among them, fft2 is the fast two-dimensional Fourier transform function in matlab, ifft2 is the fast two-dimensional inverse Fourier transform function, conj is the conjugate functi...

Embodiment 3

[0090] The template image global search method based on mask matrix and fast Fourier transform is the same as embodiment 1-2, and the energy matrix of calculating the whole image described in step (7) is expressed as follows:

[0091]

[0092] Among them, F -1 Represents inverse fast Fourier transform, C stands for conjugation, F stands for fast Fourier transform, subscript 2 stands for two-dimensional transformation, mask_ext stands for mask extension map, img stands for full image, img_energy stands for output full image energy matrix.

[0093] If you use matlab language, the above formula can be expressed as:

[0094] img_energy=ifft2(conj(fft2(mask_ext)).*fft2(img.*img))

[0095] Among them, fft2 is the fast two-dimensional Fourier transform function in matlab, ifft2 is the fast two-dimensional inverse Fourier transform function, and conj is the conjugate function to calculate the energy matrix of the whole image.

[0096] If you use the C++ language, you can write y...

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 template image global search method based on mask matrix and fast Fourier transform. It mainly solves the problem that the operation complexity of the existing template search method is too high. The present invention includes copying the template graph and the full graph, using the template graph to generate a mask graph, processing the original template graph with the mask graph, calculating the energy of the template graph, expanding the template graph and the mask graph, performing circular convolution on the template expanded graph and the full graph , calculate the energy matrix of the whole image, bring it into the template matching formula, find out the maximum response position, and realize the steps of searching the whole image. The present invention realizes the one-time calculation of the response matrix through the expansion of the template graph and the generation of the mask graph; realizes dense matching; calculates circular convolution through Fourier transform, reduces the complexity of the template search operation, and greatly reduces the operation time; The "circular direction" of the circular convolution is guaranteed by conjugation. The invention can be applied to situations such as video tracking, template position search, mechanical parts positioning and the like.

Description

[0001] Technical field: [0002] The invention belongs to the technical field of image processing, and mainly relates to image matching, and furthermore, it is a method for searching a smaller template image in a larger image, specifically a method of searching for a template image in a whole image at one time using a mask matrix method. The invention can be applied to situations such as video tracking, template position search, mechanical parts positioning and the like. [0003] Background technique: [0004] The search of the template map has important engineering value, such as the tracking of the target, and the positioning of the mechanical parts, using the template map to search in the whole map, so as to get the area or target to be searched as soon as possible. The template map represents the target to be searched, generally a smaller map; the full map represents the entire search range, and there are one or several small areas in the full map, these small areas are re...

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 Patents(China)
IPC IPC(8): G06K9/62
CPCG06F18/22
Inventor 李云松王康康
Owner XIDIAN 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