Unlock instant, AI-driven research and patent intelligence for your innovation.

Real-time Harris corner extraction method based on improved Gauss model

A corner point extraction and model technology, which is applied in the fields of image matching, target recognition and target tracking, can solve problems such as difficult implementation and multiple floating point numbers, and achieve the effects of reducing difficulty, improving corner point extraction speed, and reducing clock delay

Inactive Publication Date: 2021-03-02
LUOYANG INST OF ELECTRO OPTICAL EQUIP OF AVIC
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

From the Gaussian kernel distribution above, it can be seen that there are many floating-point numbers in the Gaussian filter model, which is very difficult for FPGA implementation.

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
  • Real-time Harris corner extraction method based on improved Gauss model
  • Real-time Harris corner extraction method based on improved Gauss model
  • Real-time Harris corner extraction method based on improved Gauss model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0033] The principle of Harris corner extraction algorithm is:

[0034] Step 1: By calculating the grayscale change of the surrounding window of a certain point (x, y) in the image I and the surrounding window of the point after it moves (Δx, Δy), the position of the corner point is detected, and the change is determined by the autocorrelation function formula (1) express:

[0035]

[0036] Among them, w(u, v) is the Gaussian weighting function, W(x, y) is a square window centered on the point (x, y), and the Taylor expansion of the formula (1) is used to obtain the first-order approximation of the formula (2) :

[0037]

[0038] Among them, M is a second-order real symmetric matrix, and the specific form is shown in the following formula (3):

[0039]

[0040] Where G is the filtering model, and the improved Gauss model is used to replace t...

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 provides a real-time Harris corner extraction method based on an improved Gauss model, and the method comprises the steps of calculating the gray scale change of a certain point in an image, detecting the position of the corner, calculating the response value of the corner according to M, and obtaining the point as the Harris corner in the image when the R value of a pixel point is greater than a threshold value. According to the invention, an improved Gaussian model is provided, the improved Gaussian model is more suitable for logic calculation, a traditional Gaussian filteringmodel is optimized according to logic implementation characteristics, complex floating-point number multiplication and division convolution operation conversion data shift is achieved while the cornerpoint extraction precision is reserved, and the logic implementation difficulty is greatly reduced. A traditional 7 * 7 floating point decimal model is simplified, data shifting operation can be conveniently used for replacing multiplication and division operation during logic implementation, clock delay is reduced, and the angular point extraction speed is increased.

Description

technical field [0001] The invention relates to the fields of image matching, target recognition, target tracking and the like, in particular to a Harris corner point extraction method. Background technique [0002] The traditional Harris corner point extraction process uses the classic Gauss model to fuzzy the partial derivative calculation of image gray information, according to the calculation formula of two-dimensional Gauss can get σ 2 = 1 continuous Gaussian response such as figure 1 shown. [0003] Theoretical analysis shows that Gaussian weights distribute non-negative values ​​in all domains of definition, so an infinite convolution kernel is required when processing images. but from figure 1 As can be seen in , the main weight values ​​are distributed in the domain around the center position. In fact, it is only necessary to take weight values ​​within 3 times the variance to establish a Gaussian kernel. figure 1 σ 2 =1 Gaussian model, so as shown in the tr...

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): G06K9/46G06F17/16
CPCG06F17/16G06V10/44
Inventor 王浩郭许生姚群磊
Owner LUOYANG INST OF ELECTRO OPTICAL EQUIP OF AVIC