An Image Stitching Method Based on Planar Area Consistency

An image stitching and plane technology, which is applied to the details of image stitching, image enhancement, image analysis, etc., can solve problems such as the inability to produce accurate and reliable stitching results, the inability to guarantee the extraction of plane areas, and the confusion of features, and achieve accurate and efficient segmentation. The effect of high training and segmentation accuracy

Active Publication Date: 2022-04-22
NANJING UNIV
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the current method cannot guarantee the extraction of significant planar areas in the scene, resulting in the mixing and mutual influence of features in each planar area during processing, and cannot produce accurate and reliable stitching results.

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
  • An Image Stitching Method Based on Planar Area Consistency
  • An Image Stitching Method Based on Planar Area Consistency
  • An Image Stitching Method Based on Planar Area Consistency

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0143] The experimental hardware environment of the present embodiment is: Intel (R) Core (TM) i7-6850K CPU@3.60GHz,

[0144] GeForce GTX 1080Ti, 12G video memory, software development using Python 3.7, C++11, OpenCV3.4.2 and SciPy. The test image comes from the dataset in the article Scene parsing through ade20k dataset.

[0145] The invention discloses an image mosaic method based on consistent plane regions, the core of which is to find the plane regions in the input image; find the consistent regions in different images; calculate the matching relationship and similar transformation relationship according to the consistent regions; solve the optimized network Grid, texture mapping and fusion of images to be stitched, including the following steps:

[0146] Step 1, train the planar region segmentation network of the encoder-decoder structure;

[0147] Step 2, perform planar region segmentation on the input image;

[0148] Step 3, use the SIFT algorithm to find the featur...

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 present invention provides an image mosaic method based on consistent planar regions, comprising: using an encoder-decoder architecture to train a planar region segmentation network; using the network to extract the planar regions of an input image; using the SIFT algorithm to find feature points of the input image and Perform matching; use the results of plane region segmentation and feature points to obtain consistent plane regions in multi-channel images and reliable matching feature point pairs belonging to this region; use the above feature point pairs to calculate matching points and local similarity transformations in overlapping regions relationship; use the grid deformation architecture to solve the grid coordinates after stitching; use texture mapping to obtain the warped image; linearly fuse the warped image to obtain the final stitching result.

Description

technical field [0001] The invention relates to the fields of computer vision, semantic segmentation, grid deformation, etc., and in particular to an image mosaic method based on consistent plane regions. Background technique [0002] With the development of science and technology, people enjoy the convenience of modern image acquisition technology, which also leads to the need to stitch low-resolution, small-angle images into high-resolution, wide-angle images. Image stitching technology can not only produce beautiful wide-angle images, but also effectively reduce storage, and is widely used in video surveillance, space detection, virtual reality, medical imaging, meteorology, military, remote sensing and many other fields. [0003] General-purpose image stitching algorithms often have strict requirements on the scene of the input image, such as requiring it to be approximated as a plane or comply with conic projection constraints. These algorithms often produce stitching ...

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): G06T3/40G06T5/50G06T7/11G06T15/04G06V10/75G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06T3/4038G06T5/50G06T7/11G06T15/04G06N3/084G06T2200/32G06T2207/10024G06T2207/20221G06T2207/20081G06T2207/20084G06V10/757G06N3/045
Inventor 李奡程郭延文
Owner NANJING UNIV
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
Try Eureka
PatSnap group products