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

SAR image segmentation method based on shape completion area chart and feature coding

A technology of image segmentation and feature encoding, which is applied in the field of image processing and can solve problems such as unclosed homogeneous regions

Active Publication Date: 2015-03-25
XIDIAN UNIV
View PDF4 Cites 19 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention proposes a new area map extraction method based on shape-based sketch line completion for the segmentation of SAR images, which solves the problem of unclosed homogeneous areas encountered in the area extraction of the above-mentioned existing methods.

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
  • SAR image segmentation method based on shape completion area chart and feature coding
  • SAR image segmentation method based on shape completion area chart and feature coding
  • SAR image segmentation method based on shape completion area chart and feature coding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] refer to figure 1 , the implementation steps of the present invention are as follows:

[0048] Step 1. For the input SAR image, the piperiver image used in the simulation experiment, such as Figure 6 As shown, the initial sketch image of the SAR image is extracted to obtain its sketch image. The extraction steps are as follows:

[0049] First, filter the original SAR image using a set of filters composed of a dictionary of image primitives, fuse the filtering results and perform non-maximum suppression to obtain an unprocessed sketch; then perform a greedy method on the unprocessed sketch The tracking method sequentially adds new image primitives to form line segment outlines to obtain unrepaired sketches; finally, based on the unrepaired sketches, space constraints are added according to the Gestalt aggregation principle, and a set of operations in graphics is used The operator repairs and corrects the sketch lines in the unrepaired sketch, and finally obtains the ...

Embodiment 2

[0067] The SAR image segmentation method based on the shape complement area map and feature coding is the same as in embodiment 1, wherein the specific steps of the shape-based sketch line complement method refer to image 3 Can be expressed as:

[0068] 2.4.1. For each sketch line except the gathering area in the sketch map, carry out the following process, for an endpoint P1 of the current sketch line, first search whether there are other sketch line endpoints in its 5*5 square neighborhood, If so, select the endpoint of the longest sketch line, and use a line segment to connect P1 with the endpoint of the longest sketch line, and then continue this process for the other endpoint of the newly connected longest sketch line until endpoint 5 *5 Until there are no new sketch line endpoints within the square neighborhood. When there is no new sketch line endpoint in the 5*5 square neighborhood, repeat the same process as P1 for another endpoint P2 of the current sketch line.

...

Embodiment 3

[0073] The SAR image segmentation method based on the shape-completion area map and feature coding is the same as that in Embodiment 1-2, wherein the merging strategy in step 7 is specifically:

[0074] (7.1) For each small area in the structural area, first, according to the spatial neighbor relationship, if it is adjacent to a certain aggregation area and the number of pixels is less than 200, it will be placed in the corresponding aggregation area; If the regions are adjacent and the difference in gray value is not more than 25, they will be placed in the corresponding homogeneous region.

[0075] (7.2) For the remaining unmarked small area, if the number of pixels is greater than 36, the class label of the area is designated as the class label of the most similar marked area according to the gray mean value; if the number of pixels is less than or equal to 36, directly set it Collocated into the spatially closest region.

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 SAR image segmentation method based on a shape completion area chart and feature coding. The method solves the technical problem of closure extraction of SAR image areas. The method includes the steps that a sketch is obtained by utilizing an SAR image initial sketch model; the area chart is obtained by utilizing the area chart completed through sketch lines and based on shapes; the area chart is mapped into an original SAR image space, and then a clustering area, a homogeneous area and a structural area are obtained; characteristic learning and hierarchical clustering are carried out on the clustering area; characteristic learning and hierarchical clustering are carried out on the homogeneous area; segmentation is carried out on the structural area; results of the clustering area, the homogeneous area and the structural area are combined, line targets are marked, and a final SAR image segmentation result is obtained. According to the method, the homogeneous area with the better closure and homogeneity can be extracted, further clustering of the clustering area and the homogeneous area can be well achieved in the characteristic learning and clustering methods applied to the areas, and the method can be used for SAR image segmentation.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to a region extraction method of a SAR image, in particular to a SAR image segmentation method based on a shape-completion region map and feature coding, which can be used for SAR image segmentation. Background technique [0002] SAR image segmentation is one of the key links in SAR image processing and interpretation. During the imaging process of SAR images, due to the different heights of ground objects and different imaging angles, the areas where ground objects gather show the characteristics of alternating light and dark in SAR images. The sketch of the SAR image proposed in "Local maximal homogenous region search for SAR speckle reduction with sketch-based geometrical kernel function" is a sparse representation of the SAR image, which describes the grayscale change information in the SAR image in the form of sketch lines . The sketch line classification method based ...

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): G06T7/00
CPCG06T7/11G06T2207/10044G06T2207/20112
Inventor 刘芳马文萍刘佳李玲玲焦李成郝红侠杨淑媛孙涛张向荣尚荣华
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