CN107240127A - The image registration appraisal procedure of distinguished point based mapping - Google Patents

The image registration appraisal procedure of distinguished point based mapping Download PDF

Info

Publication number
CN107240127A
CN107240127A CN201710255374.1A CN201710255374A CN107240127A CN 107240127 A CN107240127 A CN 107240127A CN 201710255374 A CN201710255374 A CN 201710255374A CN 107240127 A CN107240127 A CN 107240127A
Authority
CN
China
Prior art keywords
registration
error
reference picture
characteristic point
image
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.)
Pending
Application number
CN201710255374.1A
Other languages
Chinese (zh)
Inventor
王雪
于福江
徐国靖
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.)
China Aeronautical Radio Electronics Research Institute
Original Assignee
China Aeronautical Radio Electronics Research Institute
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 China Aeronautical Radio Electronics Research Institute filed Critical China Aeronautical Radio Electronics Research Institute
Priority to CN201710255374.1A priority Critical patent/CN107240127A/en
Publication of CN107240127A publication Critical patent/CN107240127A/en
Pending legal-status Critical Current

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
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Probability & Statistics with Applications (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of image registration accuracy appraisal procedure of distinguished point based mapping, comprise the steps of：1) some characteristic points of reference picture and image subject to registration are automatically extracted；2) select suitable geometric space transformation model to calculate transformation matrix, complete the registration of reference picture and image subject to registration；3) characteristic point of images after registration is mapped on reference picture；4) the distance between the characteristic point being mapped on reference picture and the characteristic point extracted from reference picture error are counted；5) the distance between characteristic point error is analyzed by K means clustering methods, calculates the registration error of image.The method that the present invention is provided can effectively assess the precision of image registration, help researcher targetedly to reduce data error, and innovatory algorithm performance improves image registration accuracy, to obtain high-definition picture.

Description

The image registration appraisal procedure of distinguished point based mapping
Technical field
The invention belongs to digital image processing techniques field, it is adaptable to the qualitative assessment of image registration effect.
Background technology
In image processing field, the judge for processing result image is often leading using the subjective assessment of people, is seen The influence of the person's of examining subjective factor is larger, it is sometimes desirable to coordinate subjective assessment with objective quantizating index.
It is an important content in image registration field to the qualitative assessment of registration effect.Real image registration is usually Due to not accurate criterion, considerable assessment can not be obtained by causing the validity of algorithm, so as to influence at follow-up image Reason analysis and effect.
Registration accuracy is an important indicator of image registration, generally in units of pixel, is reflected between registering image The uniformity of deformation, directly affects the treatment effects such as follow-up image co-registration, splicing, and target identification.Due to exist translation, Tend not to realize complete accuracy registration between image and reference picture after the deformation such as scaling, rotation, registration, but have one Fixed error, trueness error is smaller, shows that the location of pixels in registration result estimates, registration effect more consistent with reference picture Better；Vice versa.
The content of the invention
The goal of the invention of the present invention is to propose a kind of image registration accuracy appraisal procedure of distinguished point based mapping, is somebody's turn to do The distance between characteristic point and reference picture characteristic point of the method by calculating images after registration error, objective judgement image registration Effect.
The goal of the invention of the present invention is achieved through the following technical solutions：
A kind of image registration accuracy appraisal procedure of distinguished point based mapping, is comprised the steps of：
Step 1) automatically extract some characteristic points of reference picture and image subject to registration；
Step 2) select suitable geometric space transformation model to calculate transformation matrix, complete reference picture and figure subject to registration The registration of picture；
Step 3) utilize step 2) used in transformation matrix the characteristic point of images after registration is mapped on reference picture；
Step 4) statistics the characteristic point being mapped on reference picture and the characteristic point extracted from reference picture between away from From error；
Step 5) setting error threshold, the distance between characteristic point error is divided into two by K-means clustering methods Cluster, (1) is if the range error of two cluster centres is above error threshold, then it is assumed that registration failure；(2) if in two clusters The range error of the heart is both less than error threshold, then it is assumed that all Characteristic points match successes, takes the range error of two cluster centres Average value be used as Images Registration；(3) if the range error of a cluster centre is more than error threshold, and another is clustered The range error at center is less than error threshold, then it is assumed that the larger characteristic point of range error belongs to error hiding characteristic point, takes smaller Range error be used as Images Registration.
The beneficial effects of the present invention are：The method that the present invention is provided can effectively assess the precision of image registration, help Researcher targetedly reduces data error, and innovatory algorithm performance improves image registration accuracy, to obtain high resolution graphics Picture；Simultaneously as this method uses the registration Algorithm based on Automatic Feature Extraction, it is to avoid the interference of artificial reconnaissance and limitation, Also contribute to improve the operation efficiency of algorithm for checking aspect, reached the purpose for improving discrimination and accuracy, permitted It is significant in many practical applications.
Brief description of the drawings
Fig. 1 is the schematic flow sheet for the image registration accuracy appraisal procedure that distinguished point based maps.
Embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.
As shown in figure 1, a kind of image registration accuracy appraisal procedure of distinguished point based mapping is by calculating shown in the present embodiment Machine program realizes that main contents are as follows：
Step 1) characteristic point that completes reference picture and image subject to registration automatically extracts；
Step 2) transformation matrix is calculated according to the suitable geometric space transformation model of image-forming condition selection, complete subject to registration The spatial registration of image and reference picture；
Step 3) utilize step 2) in transformation matrix the characteristic point of the image after registration is mapped in reference picture；
Step 4) characteristic point that is mapped on reference picture of statistics and step 1) in the characteristic point extracted on reference picture it Between range error (unit：Pixel)；
Step 5) step-up error threshold value T1, error threshold T1Selection and reference picture and image subject to registration resolution ratio, figure As registration accuracy requires relevant, it can be set according to actual conditions, usual registration accuracy error is 1~2 pixel.Utilize K- Means clustering methods, are divided into two clusters by the distance between all characteristic points error, and respectively in the cluster of two clusters The range error ε of the heart1、ε2Carry out analysis judgement：(1) if the range error ε of two cluster centres1、ε2Above error threshold T1, Then think that it fails to match；(2) if the range error ε of two cluster centres1、ε2Both less than error threshold T1, then it is assumed that all features Point matching success, registration error takes the two average value(3) if the range error ε of a cluster centre1It is more than Error threshold T1, and the range error ε of another cluster centre2Less than error threshold T1, then it is assumed that the larger feature of range error Point ε1Belong to error hiding characteristic point, do not pay attention to, take less range error as Images Registration, i.e. ε=ε2。
It is understood that for those of ordinary skills, can be with technique according to the invention scheme and its hair Bright design is subject to equivalent substitution or change, and all these changes or replacement should all belong to the guarantor of appended claims of the invention Protect scope.

Claims (1)

1. a kind of image registration accuracy appraisal procedure of distinguished point based mapping, is comprised the steps of：
Step 1) automatically extract some characteristic points of reference picture and image subject to registration；
Step 2) the suitable geometric space transformation model of selection calculates transformation matrix, completes reference picture and image subject to registration Registration；
Step 3) utilize step 2) used in transformation matrix the characteristic point of images after registration is mapped on reference picture；
Step 4) count the distance between the characteristic point being mapped on reference picture and the characteristic point extracted from reference picture by mistake Difference；
Step 5) setting error threshold, the distance between characteristic point error is divided into by two clusters by K-means clustering methods, (1) if the range error of two cluster centres is above error threshold, then it is assumed that registration failure；(2) if two cluster centres Range error is both less than error threshold, then it is assumed that the success of all Characteristic points matchs, take two cluster centres range error it is flat Average is used as Images Registration；(3) if the range error of a cluster centre is more than error threshold, and another cluster centre Range error be less than error threshold, then it is assumed that the larger characteristic point of range error belongs to error hiding characteristic point, take it is less away from Images Registration is used as from error.
CN201710255374.1A 2017-04-19 2017-04-19 The image registration appraisal procedure of distinguished point based mapping Pending CN107240127A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710255374.1A CN107240127A (en) 2017-04-19 2017-04-19 The image registration appraisal procedure of distinguished point based mapping

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710255374.1A CN107240127A (en) 2017-04-19 2017-04-19 The image registration appraisal procedure of distinguished point based mapping

Publications (1)

Publication Number Publication Date
CN107240127A true CN107240127A (en) 2017-10-10

Family

ID=59983726

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710255374.1A Pending CN107240127A (en) 2017-04-19 2017-04-19 The image registration appraisal procedure of distinguished point based mapping

Country Status (1)

Country Link
CN (1) CN107240127A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109360203A (en) * 2018-10-30 2019-02-19 京东方科技集团股份有限公司 Method for registering images, image registration device and storage medium
CN111180048A (en) * 2019-12-30 2020-05-19 上海研境医疗科技有限公司 Tumor component labeling method, device, equipment and storage medium
CN112329848A (en) * 2020-11-04 2021-02-05 昆明理工大学 Image space mapping method based on advection vector field clustering
CN113741682A (en) * 2020-05-29 2021-12-03 北京七鑫易维信息技术有限公司 Method, device and equipment for mapping fixation point and storage medium

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101567051A (en) * 2009-06-03 2009-10-28 复旦大学 Image matching method based on characteristic points
CN101620619A (en) * 2009-08-07 2010-01-06 北京航空航天大学 System and method for processing gross error of measuring data based on clustering method
CN101877140A (en) * 2009-12-18 2010-11-03 北京邮电大学 Panorama-based panoramic virtual tour method
EP2538387A1 (en) * 2011-06-22 2012-12-26 The Boeing Company Image registration
CN103310219A (en) * 2012-03-08 2013-09-18 佳能株式会社 Method and equipment for precision evaluation of object shape registration, and method and equipment for registration
CN104574421A (en) * 2015-01-29 2015-04-29 北方工业大学 Large-breadth small-overlapping-area high-precision multispectral image registration method and device
CN105825504A (en) * 2016-03-11 2016-08-03 中山大学 MSCR-region-feature-based image duplicating detection method
CN106355197A (en) * 2016-08-24 2017-01-25 广东宝乐机器人股份有限公司 Navigation image matching filtering method based on K-means clustering algorithm

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101567051A (en) * 2009-06-03 2009-10-28 复旦大学 Image matching method based on characteristic points
CN101620619A (en) * 2009-08-07 2010-01-06 北京航空航天大学 System and method for processing gross error of measuring data based on clustering method
CN101877140A (en) * 2009-12-18 2010-11-03 北京邮电大学 Panorama-based panoramic virtual tour method
EP2538387A1 (en) * 2011-06-22 2012-12-26 The Boeing Company Image registration
CN103310219A (en) * 2012-03-08 2013-09-18 佳能株式会社 Method and equipment for precision evaluation of object shape registration, and method and equipment for registration
CN104574421A (en) * 2015-01-29 2015-04-29 北方工业大学 Large-breadth small-overlapping-area high-precision multispectral image registration method and device
CN105825504A (en) * 2016-03-11 2016-08-03 中山大学 MSCR-region-feature-based image duplicating detection method
CN106355197A (en) * 2016-08-24 2017-01-25 广东宝乐机器人股份有限公司 Navigation image matching filtering method based on K-means clustering algorithm

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HIDEKATA HONTANI 等: "Point-Based Non-Rigid Surface Registration with Accuracy Estimation", 《2010 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 *
王天召 等: "基于K-means聚类和RANSAC的图像配准算法研究", 《计算机工程与科学》 *
王瑞: "基于K-均值聚类图像配准算法", 《价值工程》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109360203A (en) * 2018-10-30 2019-02-19 京东方科技集团股份有限公司 Method for registering images, image registration device and storage medium
US11120565B2 (en) 2018-10-30 2021-09-14 Beijing Boe Optoelectronics Technology Co., Ltd. Image registration method, image registration device and storage medium
CN109360203B (en) * 2018-10-30 2021-12-03 京东方科技集团股份有限公司 Image registration method, image registration device and storage medium
CN111180048A (en) * 2019-12-30 2020-05-19 上海研境医疗科技有限公司 Tumor component labeling method, device, equipment and storage medium
CN111180048B (en) * 2019-12-30 2024-02-06 上海镜观生物科技有限公司 Tumor component labeling method, device, equipment and storage medium
CN113741682A (en) * 2020-05-29 2021-12-03 北京七鑫易维信息技术有限公司 Method, device and equipment for mapping fixation point and storage medium
CN112329848A (en) * 2020-11-04 2021-02-05 昆明理工大学 Image space mapping method based on advection vector field clustering
CN112329848B (en) * 2020-11-04 2022-07-29 昆明理工大学 Image space mapping method based on advection vector field clustering

Similar Documents

Publication Publication Date Title
CN112818988B (en) Automatic identification reading method and system for pointer instrument
WO2021138990A1 (en) Adaptive detection method for checkerboard sub-pixel corner points
CN107240127A (en) The image registration appraisal procedure of distinguished point based mapping
CN107679535B (en) Automatic reading identification system and method for pointer type water meter based on template matching
CN104484648B (en) Robot variable visual angle obstacle detection method based on outline identification
CN111476827B (en) Target tracking method, system, electronic device and storage medium
CN106023151B (en) Tongue object detection method under a kind of open environment
CN103048331B (en) Printing defect detection method based on flexible template registration
CN109584281B (en) Overlapping particle layering counting method based on color image and depth image
CN105957041B (en) A kind of wide-angle lens infrared image distortion correction method
CN104168478B (en) Based on the video image color cast detection method of Lab space and relevance function
CN106709500B (en) Image feature matching method
CN105046701B (en) Multi-scale salient target detection method based on construction graph
CN106127205A (en) A kind of recognition methods of the digital instrument image being applicable to indoor track machine people
WO2015131468A1 (en) Method and system for estimating fingerprint pose
CN104517101A (en) Game poker card recognition method based on pixel square difference matching
CN102982524B (en) Splicing method for corn ear order images
CN105488475B (en) Method for detecting human face in mobile phone
CN107545550B (en) Cell image color cast correction method
CN106023187A (en) Image registration method based on SIFT feature and angle relative distance
Chen et al. Method on water level ruler reading recognition based on image processing
CN102915540A (en) Image matching method based on improved Harris-Laplace and scale invariant feature transform (SIFT) descriptor
CN102663384A (en) Curve identification method based on Bezier control point searching and apparatus thereof
CN106682674A (en) Pointer type meter automatic reading method based on intelligent tour-inspection robot
CN103714550A (en) Image registration automatic optimization algorithm based on matching of curve characteristic evaluation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20171010