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

A shape-optimized rectangular panoramic image construction method based on feature selection

A panoramic image and feature selection technology, which is applied in the field of image processing technology and panoramic imaging, can solve the problems of unnatural rotation of spliced ​​images, achieve realistic reproduction, avoid information loss, and overcome shape distortion

Inactive Publication Date: 2019-05-17
YANTAI UNIV
View PDF2 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these methods are transitions from perspective transformation to similarity transformation, which may cause unnatural rotation of stitched images

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 shape-optimized rectangular panoramic image construction method based on feature selection
  • A shape-optimized rectangular panoramic image construction method based on feature selection
  • A shape-optimized rectangular panoramic image construction method based on feature selection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053] The embodiment of the present invention proposes a shape optimization method of a rectangular panorama based on feature selection, see figure 2 , the method includes the following steps:

[0054] 101: Stitching two overlapping images taken by a common camera in a non-fixed manner to obtain a rectangular panoramic image with no distortion in shape and no loss of information;

[0055] Wherein, the operation of this step breaks the normal shooting method of panorama image construction that must be fixed horizontal movement and fixed point rotation.

[0056] 102: First perform feature point matching on the input image, and then use a clustering algorithm to screen out mismatched feature points and isolated feature points;

[0057] Wherein, the step 102 is specifically:

[0058] For the feature selection of two overlapping images taken in a non-fixed manner, first detect the SIFT feature points matched by the two images, and use the Ransac (Random Sample Consensus, random...

Embodiment 2

[0097] Below in combination with specific calculation formulas and examples, the scheme in Embodiment 1 is further introduced, see the following description for details:

[0098] 1. Introduction to Moving Direct Linear Transformation (MDLT)

[0099] Suppose f and f' are a pair of matching points on overlapping images. They can be mapped by projective transformations or homography. For example, f'=Hf is a method of global projection transformation, which will cause misalignment. Zaragoza et.al proposed the MDLT valuation method to calculate the local H. MDLT transforms each mesh vertex using locally correlated homography. Let mesh vertices near feature points be assigned higher weights; projective transformations have a better effect on local structures. This expression is as follows

[0100]

[0101]

[0102]

[0103] 2. The method proposed in this paper

[0104] MDLT methods can achieve highly accurate calibrations by maintaining global projection and allowing ...

Embodiment 3

[0140] In order to verify the effectiveness of this method, in this section, experiments are carried out on images of different views captured by accident, and the results of generating panoramas are compared with other three stitching methods, namely, Photoshop CS6, APAP transformation method [4] , Stereo stitching image shape optimization method [17] and other methods to draw cylindrical panoramas. The comparison results are as Figure 5 shown. In addition, more results obtained by this method are shown. The APAP transformation utilizes a local projective transformation to make the image precisely calibrated. The gradient warping method adopts the gradient idea to avoid shape distortion.

[0141] Figure 5 (a) is a cylindrical panorama. There are certain problems such as shape distortion and information loss after image cropping. Such as Figure 5 As shown in (b), when the lack of shape constraints cannot produce good panoramic images, the APAP method can well calibr...

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 shape-optimized rectangular panoramic image construction method based on feature selection, and the method comprises the steps of carrying out the splicing of two overlappedimages which are shot in a non-fixed mode based on a common camera, and obtaining a rectangular panoramic image which is not distorted in shape and is not lost in information; taking the shot image asan input image, performing feature point matching on the input image, and screening mismatched feature points and isolated feature points by using a clustering algorithm; performing image registration operation on the input image overlapping area by utilizing the screened feature points; gradually changing distortion transformation is carried out in the non-overlapping area so as to reduce perspective distortion generated by the image; and establishing a grid optimization model based on boundary constraint and central region reservation, obtaining an optimal rectangular grid vertex, and finally obtaining a final rectangular panoramic image through an interpolation method. According to the invention, the rectangular panoramic image is generated under the condition that the information lossis reduced as much as possible through non-fixed shooting of the image with the overlapped area.

Description

technical field [0001] The present invention relates to the field of image processing technology and panoramic imaging, in particular to a rectangular panoramic image construction method for shape optimization based on feature selection for two overlapping images captured by a common camera in a non-fixed manner. Background technique [0002] With the development of virtual reality technology, people's demand for large viewing angle images is increasing. However, due to the limitations of imaging equipment, it is difficult to capture large viewing angle images in single-lens shooting. Therefore, researchers have proposed some research methods to generate large-view images. [0003] Image stitching can combine multiple adjacent single small-view images ( figure 1 (a)) Stitching into a single image with large viewing angles. Image stitching is to transform the perspective of images by projective distortion (such as cylinder, sphere or perspective) to obtain registered image...

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): G06T3/00G06T3/40
Inventor 阎维青
Owner YANTAI 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