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

Image Speckle Detection Method Based on Anisotropic Gaussian Kernel and Gradient Search

An anisotropic and gradient search technology, applied in image analysis, image data processing, instruments, etc., can solve the problems of accuracy impact, inability to detect spots well, and high computational complexity, and achieve high accuracy and solve computational problems. The complexity is affected by the detection accuracy and the effect of improving the operation efficiency

Active Publication Date: 2020-04-07
XIDIAN UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Common differential methods include Gaussian Laplacian algorithm, Gaussian scale difference algorithm and Hessian determinant algorithm, etc. These algorithms all use isotropic Gaussian kernel, so they are invariant or covariant to image translation, rotation and coordinate scale transformation , but only detects circular blobs
In a practical computer vision system, the input image is often subjected to affine transformation, and the shapes of the spots are also varied, and none of the above methods can detect the spots well.
In this regard, scholars have proposed an affine adaptive differential speckle detection method and a speckle detection algorithm based on the generalized Gaussian Laplacian operator. Describe the spots, but the former needs continuous iteration, the latter has high computational complexity, and since both of them perform parameter search in discrete space, the accuracy of detection will be affected by the fineness of parameter discretization

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 Speckle Detection Method Based on Anisotropic Gaussian Kernel and Gradient Search
  • Image Speckle Detection Method Based on Anisotropic Gaussian Kernel and Gradient Search
  • Image Speckle Detection Method Based on Anisotropic Gaussian Kernel and Gradient Search

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0029]Speckle detection is an important part of the image feature detection technology field, a special case of region detection, and an important preprocessing link in many feature generation, target recognition and other methods. Compared with other image features, speckles provide area information that edges, contours and corners cannot provide. Compared with pure corners, it has better stability and stronger anti-noise ability, so it can be used in images. Target recognition and tracking, texture analysis and texture recognition and other fields. While the existing Gaussian Laplacian algorithm can only detect circular spots, although the affine adaptive differential spot detection method and the generalized Gaussian Laplacian algorithm can simultaneously detect circular and elliptical spots, and can better Describe the spots, but the former requires continuous iteration, and the latter has high computational complexity, and since both of them perform parameter search in di...

Embodiment 2

[0043] The image speckle detection method based on anisotropic Gaussian kernel and gradient search is the same as in embodiment 1, and the screening candidate spots in step (3) of the present invention obtains initial spots, specifically, the spot screening process is carried out according to the scale and position estimation of initial spots For two spots with close positions and high overlap rate, compare their scale estimates, keep the spot with larger scale estimate, and delete the spot with smaller scale estimate, because the larger scale contains more information, Therefore, it is also more stable; all candidate spots are screened using the above screening process, and the result obtained is the initial spot.

Embodiment 3

[0045] The image spot detection method based on anisotropic Gaussian kernel and gradient search is the same as embodiment 1-2, and the anisotropic Gaussian Laplacian filter in the step (5) of the present invention, its form is as follows:

[0046]

[0047] Among them, ▽ 2 Represents the Laplacian operator, det represents the determinant operation of the matrix, g(x; Σ) represents the anisotropic Gaussian kernel, (x, y) represents the two-dimensional plane coordinates, (u, v) represents the anisotropy The central coordinates of the filter, Σ represents the covariance matrix, ρ represents the anisotropy factor, σ represents the scale parameter of the anisotropic Gaussian kernel, and the superscript T represents the transpose. The advantage of this filter form is that the central coordinate (u,v) controls the position of the filter, thereby controlling the relative position of the filter center and the local image center, the anisotropy factor ρ controls the shape of the filte...

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 spot detection method based on the anisotropic Gaussian kernel and gradient search. The invention solves problems that a conventional method is not accurate in estimation of the spot shape and is complex in calculation, and the method comprises the steps: converting an input image into a gray scale image through MATLAB; obtaining candidate spots through Gaussian-Laplacian spot detection; screening the candidate spots according to the spot overlapping rate, and obtaining initial spots; selecting a local image for each initial spot; generating a normalized on anisotropic Gaussian-Laplacian filter through MATLAB, carrying out the filtering of the local image, obtaining the response, and taking the response of the central position as a target function; searching the parameter corresponding to the maximum value of the target function as the final spot detection result through the gradient search method; displaying and comparing the spot detection result through MATLAB. The method is high in detection precision, is low in calculation complexity, and can be used for image feature detection, image registering and image recognition.

Description

technical field [0001] The invention belongs to the technical field of image feature detection, and relates to a speckle detection method, in particular to an image speckle detection method based on anisotropic Gaussian kernel and gradient search, which can be used for speckle detection in grayscale images. Background technique [0002] Speckle detection is an important part of the field of image feature detection technology. The purpose of the speckle detection method is to detect areas with a certain geometric shape that are brighter or darker than the surrounding areas in the image, and these areas are called speckles. Blob detection is a special case of region detection, and it is an important preprocessing link in many feature generation, object recognition and other methods. Compared with other image features, blobs provide regional information that edges, contours, and corners cannot, so blob detection plays a very important role in image registration and stereo visi...

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): G06T7/00G06T7/62G06T7/73
CPCG06T7/0002G06T7/62G06T7/73
Inventor 水鹏朗李藕
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