CN104392447A - Image matching method based on gray scale gradient - Google Patents

Image matching method based on gray scale gradient Download PDF

Info

Publication number
CN104392447A
CN104392447A CN201410700097.7A CN201410700097A CN104392447A CN 104392447 A CN104392447 A CN 104392447A CN 201410700097 A CN201410700097 A CN 201410700097A CN 104392447 A CN104392447 A CN 104392447A
Authority
CN
China
Prior art keywords
degree
pixel
window
degrees
matched
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201410700097.7A
Other languages
Chinese (zh)
Other versions
CN104392447B (en
Inventor
张华�
刘桂华
史晋芳
王时丽
张静
刘满禄
王坤朋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest University of Science and Technology
Original Assignee
Southwest University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest University of Science and Technology filed Critical Southwest University of Science and Technology
Priority to CN201410700097.7A priority Critical patent/CN104392447B/en
Publication of CN104392447A publication Critical patent/CN104392447A/en
Application granted granted Critical
Publication of CN104392447B publication Critical patent/CN104392447B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses an image matching method based on a gray scale gradient. The image matching method based on the gray scale gradient includes: selecting the pixel dot to be matched from the reference image, selecting the matching window by taking the pixel dot to be matched as the centre; calculating the gradient quadratic sum of the pixel dot to be matched in the matching window along the direction of 0 degree, 45 degrees, 90 degrees and 135 degrees; selecting an object window being the same to the matching window on size and shape by taking some pixel dot in the object image as the centre; calculating the gradient quadratic sum of the pixel dot in the object window along the direction of 0 degree, 45 degrees, 90 degrees and 135 degrees; respectively calculating the gradient quadratic sum of the matching window along the direction of 0 degree, 45 degrees, 90 degrees and 135 degrees and the sum of the absolute value of the difference value of a plurality of windows of the object image along the direction of 0 degree, 45 degrees, 90 degrees and 135 degrees as the metric value; selecting the centre dot of the object window with minimum metric value as the matching pixel dot. The matching accuracy rate of the method is obviously higher than that of the SSD method and SAD method by combining the gray scale information and the gray scale gradient information and the method is effective for matching the precision dot pair and the correlation use.

Description

A kind of image matching method based on shade of gray
Technical field
The invention belongs to field of machine vision, be specifically related to the image matching method of a kind of combining image half-tone information and gradient information.
Background technology
Along with the development of science and technology, image matching technology has become a very important link in image procossing research.Image matching technology is widely used at numerous areas such as scene rebuilding, target following, aircraft navigation, medical diagnosiss.
The method of images match roughly can be divided three classes: the first kind is the image matching method based on gray scale, and Equations of The Second Kind is the image matching method of feature based, and the 3rd class is based on the matching process to image understanding and explanation.Wherein, coupling based on gray scale is using the half-tone information of each pixel local domain of stereo image pair as coupling basis, utilize certain similarity measurement, as related function, covariance function, difference quadratic sum, difference absolute value and etc. estimate extreme value, judge the corresponding relation of stereo image pair.Traditional matching process based on half-tone information has the absolute value SAD(Sum of Absolute Differences of respective pixel difference in the quadratic sum SSD (Sum of Squared Differences) of respective pixel difference in image sequence, image sequence) etc.
But, traditional matching process based on gray scale is mainly from the summation statistically considering the grey scale change in a small neighbourhood, what describe is grey similarity between pixel region, do not consider the spatial coherence between unique point, similar for grey scale change but the region that intensity profile is different is insensitive, easily cause error hiding.
Summary of the invention
The object of this invention is to provide a kind of image matching method based on shade of gray, combine half-tone information and shade of gray information, matching accuracy rate, apparently higher than SSD method and SAD method, can be effective to the right coupling of Accurate Points and related application.
An embodiment provides a kind of image matching method based on shade of gray, comprising: in reference picture, choose pixel to be matched, and choose square match window centered by pixel to be matched; Calculate pixel to be matched gradient quadratic sum along 0 degree, 45 degree, 90 degree and 135 degree four direction in match window; Ordinal selection matched pixel point in target image, and the target window all identical with match window size and shape is chosen centered by matched pixel point; Calculate matched pixel point in target window along the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction; Calculate the absolute value sum 0 degree, 45 degree, 90 degree and 135 degree equidirectional gradient sum-of-squares difference of multiple target window in match window and target image respectively as metric; And the central point choosing the minimum target window of metric is as matched pixel point.
Calculating pixel in match window or target window along the method for the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction can be:
0 degree of direction: ;
45 degree of directions: ;
90 degree of directions: ;
135 degree of directions: ;
Wherein (xi, yi) is the pixel in reference picture or target image, represent the grey scale pixel value at (xi, yi) place in reference picture or target image.
Image matching method based on shade of gray provided by the invention is based on following principle: if there is the pixel pair of coupling in two width figure, and the pixel value intensity profile change that so space of these pixels is adjacent is also similar or identical.Therefore, consider the spatial coherence utilized between unique point, judge whether pixel mates according to the absolute difference sum of certain pixel all directions corresponding grey scale gradient in two images.In conjunction with half-tone information and shade of gray information, matching accuracy rate, apparently higher than SSD method and SAD method, can be effective to the right coupling of Accurate Points and related application.
Accompanying drawing explanation
Figure 1 shows that the process flow diagram of an embodiment of the image matching method based on shade of gray of the present invention;
Figure 2 shows that in calculating match window of the present invention or target window, pixel is along the schematic diagram of the method for the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction;
Figure 3 shows that method of the present invention mates the design sketch obtained to Reindeer figure.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, below in conjunction with the specific embodiment of the invention and corresponding accompanying drawing, technical solution of the present invention is clearly and completely described.
With reference to figure 1, Figure 1 shows that the process flow diagram of an embodiment 100 of the image matching method based on shade of gray of the present invention.Embodiment 100 comprises the steps 101 to 106.
In a step 101, in reference picture, choose pixel to be matched, and choose square match window centered by pixel to be matched.The square match window chosen should be symmetrical and symmetrical up and down based on pixel to be matched.
In a step 102, pixel to be matched gradient quadratic sum along 0 degree, 45 degree, 90 degree and 135 degree four direction in match window is calculated.
As shown in Figure 2, in one embodiment of the invention, calculating pixel in window to be matched along the method for the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction can be:
0 degree of direction: ;
45 degree of directions: ;
90 degree of directions: ;
135 degree of directions: ;
Wherein, (x i, y i) be the pixel in reference picture, represent (x in reference picture i, y i) grey scale pixel value at place.
In step 103, ordinal selection matched pixel point in target image, and the target window all identical with match window size and shape is chosen centered by matched pixel point.The target window chosen should be symmetrical and symmetrical up and down based on matched pixel point.
In one embodiment of the invention, all target windows in target image can be traveled through with match window.
At step 104, matched pixel point is calculated in target window along the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction.
In one embodiment of the invention, calculating pixel in target window along the method for the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction can be:
0 degree of direction: ;
45 degree of directions: ;
90 degree of directions: ;
135 degree of directions:
Wherein, (x i, y i) be the pixel in target image, represent (x in target image i, y i) grey scale pixel value at place.
In step 105, the absolute value sum 0 degree, 45 degree, 90 degree and 135 degree equidirectional gradient sum-of-squares difference of multiple target window in match window and target image is calculated respectively as metric.The formula of the absolute value sum of compute gradient sum-of-squares difference can be:
Wherein, V li(x, y)represent the gradient quadratic sum of the pixel in reference picture along certain direction, V ri(x, y)represent the gradient quadratic sum of the pixel in target image along equidirectional.
In step 106, the central point of the minimum target window of metric is chosen as matched pixel point.
In one embodiment of the invention, target image can be traveled through with match window, the match point of pixel in search matching image.
To the image matching method based on shade of gray that described herein according to the embodiment of the present invention.The inventive method combines half-tone information and shade of gray information, and matching accuracy rate, apparently higher than SSD method and SAD method, can be effective to the right coupling of Accurate Points and related application.
Inventor has carried out programming realization with MATLAB to the method on computers, image to be matched is to being Reindeer, Moebius, Art tri-width image coming from standard picture storehouse, and its size is respectively: 447x370,463x370,463x370, the match window of employing is the square window of 5x5.Fig. 3 illustrates method of the present invention and Reindeer is mated to the design sketch obtained.
For verifying the accuracy that this method is mated further, the parallax value that the present invention obtains three kinds of methods and the known standard parallax value of Reindeer, Moebius, Art compare, carry out coupling required time two aspects from matching accuracy rate and a pair unique point and carried out detailed comparisons, result as shown in Table 1.Can be drawn by table one, method of the present invention in matching accuracy apparently higher than SSD and SAD two kinds of methods.
 
Table one three kinds of algorithm accuracys rate and time compare
Although with above-mentioned preferred embodiment to invention has been detailed description, not limit the present invention with above-described embodiment.Those skilled in the art should recognize when the technical characteristic do not departed from given by technical solution of the present invention and scope, the increase done technical characteristic, with the replacement of some same contents of this area, all should belong to protection scope of the present invention.

Claims (2)

1. based on an image matching method for shade of gray, it is characterized in that, comprising:
In reference picture, choose pixel to be matched, and choose square match window centered by described pixel to be matched;
Calculate described pixel to be matched gradient quadratic sum along 0 degree, 45 degree, 90 degree and 135 degree four direction in match window;
Ordinal selection matched pixel point in target image, and the target window all identical with match window size and shape is chosen centered by described matched pixel point;
Calculate described matched pixel point gradient quadratic sum along 0 degree, 45 degree, 90 degree and 135 degree four direction in target window;
Calculate the absolute value sum 0 degree, 45 degree, 90 degree and 135 degree equidirectional gradient sum-of-squares difference of multiple target window in match window and target image respectively as metric; And
Choose the central point of the minimum target window of metric as matched pixel point.
2. method according to claim 1, is characterized in that, in described calculating match window or target window, pixel along the method for the gradient quadratic sum of 0 degree, 45 degree, 90 degree and 135 degree four direction is:
0 degree of direction: ;
45 degree of directions: ;
90 degree of directions: ;
135 degree of directions: ;
Wherein (x i, y i) be the pixel in reference picture or target image, represent (x in reference picture or target image i, y i) grey scale pixel value at place.
CN201410700097.7A 2014-11-28 2014-11-28 A kind of image matching method based on shade of gray Expired - Fee Related CN104392447B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410700097.7A CN104392447B (en) 2014-11-28 2014-11-28 A kind of image matching method based on shade of gray

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410700097.7A CN104392447B (en) 2014-11-28 2014-11-28 A kind of image matching method based on shade of gray

Publications (2)

Publication Number Publication Date
CN104392447A true CN104392447A (en) 2015-03-04
CN104392447B CN104392447B (en) 2017-10-13

Family

ID=52610345

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410700097.7A Expired - Fee Related CN104392447B (en) 2014-11-28 2014-11-28 A kind of image matching method based on shade of gray

Country Status (1)

Country Link
CN (1) CN104392447B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109816619A (en) * 2019-01-28 2019-05-28 努比亚技术有限公司 Image interfusion method, device, terminal and computer readable storage medium
WO2020001034A1 (en) * 2018-06-30 2020-01-02 华为技术有限公司 Image processing method and device
CN111866493A (en) * 2020-06-09 2020-10-30 青岛小鸟看看科技有限公司 Image correction method, device and equipment based on head-mounted display equipment
CN111866492A (en) * 2020-06-09 2020-10-30 青岛小鸟看看科技有限公司 Image processing method, device and equipment based on head-mounted display equipment
CN115200797A (en) * 2022-09-19 2022-10-18 山东超华环保智能装备有限公司 Leakage detection system for zero leakage valve

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080232654A1 (en) * 2006-06-22 2008-09-25 Rudolf Maarten Bolle Fingerprint representation using localized texture feature
CN102402691A (en) * 2010-09-08 2012-04-04 中国科学院自动化研究所 Method for tracking gestures and actions of human face

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080232654A1 (en) * 2006-06-22 2008-09-25 Rudolf Maarten Bolle Fingerprint representation using localized texture feature
CN102402691A (en) * 2010-09-08 2012-04-04 中国科学院自动化研究所 Method for tracking gestures and actions of human face

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
HEIKO HIRSCHMULLER 等: "Evaluation of Stereo Matching Costs on Images with Radiometric Differences", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 *
丁雄飞 等: "基于Moravec算子和改进的SIFT算法的图像匹配", 《合肥学院学报（自然科学版）》 *
荆鑫: "计算机视觉中双目立体匹配技术的研究", 《中国优秀硕士学位论文全文数据集 信息科技辑》 *
陈跃峰 等: "基于内容检索的图像系统的关键技术及其实现", 《计算机工程与应用》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020001034A1 (en) * 2018-06-30 2020-01-02 华为技术有限公司 Image processing method and device
CN109816619A (en) * 2019-01-28 2019-05-28 努比亚技术有限公司 Image interfusion method, device, terminal and computer readable storage medium
CN111866493A (en) * 2020-06-09 2020-10-30 青岛小鸟看看科技有限公司 Image correction method, device and equipment based on head-mounted display equipment
CN111866492A (en) * 2020-06-09 2020-10-30 青岛小鸟看看科技有限公司 Image processing method, device and equipment based on head-mounted display equipment
CN111866493B (en) * 2020-06-09 2022-01-28 青岛小鸟看看科技有限公司 Image correction method, device and equipment based on head-mounted display equipment
CN115200797A (en) * 2022-09-19 2022-10-18 山东超华环保智能装备有限公司 Leakage detection system for zero leakage valve
CN115200797B (en) * 2022-09-19 2022-12-16 山东超华环保智能装备有限公司 Leakage detection system for zero leakage valve

Also Published As

Publication number Publication date
CN104392447B (en) 2017-10-13

Similar Documents

Publication Publication Date Title
Zhou et al. Semi-dense 3D reconstruction with a stereo event camera
Wang et al. Mvdepthnet: Real-time multiview depth estimation neural network
Zhang et al. Structure and motion from line correspondences: Representation, projection, initialization and sparse bundle adjustment
CN109478330B (en) Tracking system based on RGB-D camera and method thereof
CN104392447A (en) Image matching method based on gray scale gradient
Pinggera et al. Know your limits: Accuracy of long range stereoscopic object measurements in practice
CN103900583A (en) Device and method used for real-time positioning and map building
Antunes et al. Unsupervised vanishing point detection and camera calibration from a single manhattan image with radial distortion
Raposo et al. Plane-based odometry using an RGB-d Camera.
Knorr et al. Online extrinsic multi-camera calibration using ground plane induced homographies
Li et al. Co-planar parametrization for stereo-SLAM and visual-inertial odometry
Hansen et al. Online continuous stereo extrinsic parameter estimation
Gu et al. Camera calibration based on the back projection process
Xia et al. Global calibration of multi-cameras with non-overlapping fields of view based on photogrammetry and reconfigurable target
Zhou et al. A new algorithm for the extrinsic calibration of a 2D LIDAR and a camera
Heng et al. Semi-direct visual odometry for a fisheye-stereo camera
Ying et al. Radial distortion correction from a single image of a planar calibration pattern using convex optimization
Zhou et al. Semi-dense visual odometry for RGB-D cameras using approximate nearest neighbour fields
Wang et al. Accurate detection and localization of curved checkerboard-like marker based on quadratic form
US10134136B2 (en) Image processing apparatus and image processing method
Yang et al. Eccentricity error compensation for geometric camera calibration based on circular features
Mu et al. Visual navigation features selection algorithm based on instance segmentation in dynamic environment
Gu et al. Calibration of stereo rigs based on the backward projection process
Lee et al. Switch-slam: Switching-based lidar-inertial-visual slam for degenerate environments
Ahmadi et al. HDPV-SLAM: Hybrid depth-augmented panoramic visual SLAM for mobile mapping system with tilted LiDAR and panoramic visual camera

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20171013

Termination date: 20191128