CN107909578A - Light field image refocusing method based on hexagon stitching algorithm - Google Patents
- ️Fri Apr 13 2018
CN107909578A - Light field image refocusing method based on hexagon stitching algorithm - Google Patents
Light field image refocusing method based on hexagon stitching algorithm Download PDFInfo
-
Publication number
- CN107909578A CN107909578A CN201711032913.1A CN201711032913A CN107909578A CN 107909578 A CN107909578 A CN 107909578A CN 201711032913 A CN201711032913 A CN 201711032913A CN 107909578 A CN107909578 A CN 107909578A Authority
- CN
- China Prior art keywords
- image
- light
- refocusing
- hexagon
- pixel Prior art date
- 2017-10-30 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/67—Focus control based on electronic image sensor signals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
Abstract
The present invention relates to a kind of light field image refocusing method based on hexagon stitching algorithm, carries out background process first, reduces or even eliminate the aberration produced in traditional refocusing method;Select hexagon to replicate amplified grand pixel center, then spliced again with duplicating image, then reduce, you can obtain refocusing image of the digital focus in a certain depth location.Since the center of hexagon is smaller apart from difference with each marginal position, effectively avoid choosing the image that difference is larger formed by single lenticule marginal position, quality is relatively low in due to microlens array.It is that the phenomenon of several fixed discrete depths progress digital focus can only be chosen caused by when must be using single pixel as base unit not have to change selected window size in traditional algorithm.The method of the present invention realizes that any depth location continuous number focuses in measurement range, this causes depth resolution when light-field camera measures with measurable section position only by the hardware limitation of measuring system.
Description
Technical field
The present invention relates to a kind of image processing techniques, more particularly to a kind of light field image weight based on hexagon stitching algorithm Focus method.
Background technology
Light-field camera imaging technique is grown rapidly in recent years, it achieves tremendous development and extensive use in civil field, Many research institutions and the concern and research of scientific research personnel are obtained in scientific research field.Particularly, light-field camera measuring method is made For the single-lens image measuring method that image processing techniques can be used to be calculated measurement result, analyzed only is used only, Its measurement accuracy, speed, information depth are outside light-field camera development in itself, also pushing as image processing techniques development It is dynamic.Since light-field camera measuring method measuring speed is fast, can be with the advantage such as On-line sampling system, low to environmental requirement, it is in essence In close fields of measurement, especially particle sizing field, there are huge development potentiality and Research Prospects.
In light-field camera measuring method, the refocusing mistake for the image for original image is restored to be calculated, analyzing Journey is one of most important step in whole light-field camera measurement.The quality and precision of refocusing determine follow-up calculating, analysis Quality and precision the upper limit.But since in accurate measurement, commonly used microlens array is honeycomb distributed architecture, The microlens array of the transverse and longitudinal arrangement of this and conventional light-field camera is the light-field cameras of different therefore traditional square splicings Refocusing method can not be applied in scientific research now the light-field camera equipped with honeycomb profile microlens array widely applied.Separately On the one hand, for the light-field camera equipped with honeycomb profile microlens array, researcher uses parallelogram splicing side Method, but this joining method can increase influence of the lenticule marginal portion aberration to measurement result, and can the largely upper limit System splicing subgraph size, so as to limit the range of choice of refocusing depth distance.
Meanwhile existing refocusing algorithm is the institute in units of single pixel since it splices subgraph size in selection Several fixed discrete depths can only be chosen in traditional refocusing method and carry out digital focus, this is to be difficult to receive in accurate measurement 's.Therefore, at present in the field of precision measurement of light-field camera, especially particle three-dimensional measurement, it is also necessary to a kind of low error, height Precision, can any depth location continuous number focuses in measurement range light-field camera refocusing algorithm.
The content of the invention
The present invention be directed to traditional refocusing method there are the problem of, it is proposed that a kind of light based on hexagon stitching algorithm Field picture refocusing method, while realizing that any depth location continuous number focuses in measurement range, with respect to conventional method, Improve picture quality.
The technical scheme is that:A kind of light field image refocusing method based on hexagon stitching algorithm, specific bag Include following steps:
1) light-field camera used is demarcated using transparent calibrated thing first before, measuring, obtains the reality that pixel represents Border size;
2) region to be measured, is illuminated using light source, position and the intensity size of light source is adjusted, makes bias light uniform;
3), region income to be measured is shot using light-field camera, is obtained without the background image for treating object;
4), hollow glass bead particulate samples to be measured are placed in region to be measured, region to be measured is carried out using light-field camera Shooting, obtains the particle image for having object under test;
5), particle image and the gray value of each pixel of background image are subtracted each other, obtains and new removes background image;
6), each grand pixel image in background image will be gone to amplify N times, N is the direct proportion letter of refocusing depth distance d Number, proportionality coefficient is by test calibration or the amplification factor N that optical field imaging systematic parameter used is calculated and depth The respective function relation of distance d obtains;
7) figure, is carried out to amplified each grand pixel image center with the regular hexagon window of a fixed dimension As replicating;
8), each grand pixel image center image that will be copied in step 7) with regular hexagon window, according to former grand picture The opposite link position of element carries out image mosaic, and with reference to the amplification factor N in step 6, stitching image reduced n times, n be N just Proportion function, proportionality coefficient are determined by required final image size, you can obtain digital focus depth location d in step 6) Refocusing image.
The beneficial effects of the present invention are:Light field image refocusing method of the invention based on hexagon stitching algorithm, can To significantly reduce or even eliminate the aberration produced in these traditional refocusing methods, and largely improve accurate imaging and Measure quality.The method of the present invention any depth location continuous number can focus in measurement range, this cause light-field camera into Depth resolution and measurable section position during row measurement only by the hardware limitation of measuring system, this for accurate measurement and Imaging is of great significance.
Brief description of the drawings
Fig. 1 carries out duplication schematic diagram using hexagonal window openings for the present invention to grand pixel image;
Fig. 2 is that the present invention does not remove the refocusing image of background and removes background refocusing image comparison figure;
Fig. 3 is Conventional parallel quadrangle refocusing method and hexagon refocusing method comparison diagram of the present invention.
Embodiment
Light field image refocusing method based on hexagon stitching algorithm, be one kind go background, can be in measurement range Any depth location continuous number focuses on, the light field image refocusing method especially suitable for accurate measurement or imaging.It is specific real Existing mode includes the following steps:
1st, the light-field camera used is demarcated using transparent calibrated thing first before measuring, obtains the reality that pixel represents Size;
2nd, region to be measured is illuminated using light source, adjusts position and the intensity size of light source, make bias light uniform;
3rd, region income to be measured is shot using light-field camera, obtained without the background image for treating object;
4th, hollow glass bead particulate samples to be measured are placed in region to be measured, region to be measured is clapped using light-field camera Take the photograph, obtain the particle image for having object under test;
5th, particle image and the gray value of each pixel of background image are subtracted each other, obtains and new remove background image;
6th, each grand pixel image in background image will be gone to amplify N times, N is the direct proportion letter of refocusing depth distance d Number, proportionality coefficient is by test calibration or the amplification factor N that optical field imaging systematic parameter used is calculated and depth The respective function relation of distance d obtains.
7th, figure is carried out to amplified each grand pixel image center with the regular hexagon window of a fixed dimension As replicating, duplication schematic diagram is carried out to grand pixel image using hexagonal window openings as shown in Figure 1.Wherein, regular hexagon window The length of side is the half of adjacent grand pixel center distance.
8th, each grand pixel image center image that will be copied in step 7 with regular hexagon window, according to former grand pixel Image mosaic is carried out with respect to link position, with reference to the amplification factor N in step 6, stitching image is reduced n times, n is the direct ratio of N Example function, proportionality coefficient determines by required final image size, you can the weight of acquisition digital focus depth location d in step 6 Focusedimage.
Hollow glass bead particle in the above embodiment is replaced with to the scaling board of dark circular patterns, remaining measurement Step is same, can detect focusing effect.
Traditional refocusing process can due to environment influence, source mass, main lens optical property, microlens array into The influence of the factors such as the interference of picture and largely reduce the quality of refocusing image.Especially microlens array can cause original Each grand pixel is significantly reduced by the gray scale at center to surrounding in beginning image, and the gray-tone response of each different grand pixels also has not Negligible difference.These influences can cause the aberration of refocusing image to significantly increase so as to influence final imaging and measurement Quality.And do not remove the refocusing image of background as shown in Figure 2 and remove background refocusing image comparison figure, the left side is not remove the back of the body Scape refocusing image, background refocusing image is is removed in the right, it is seen that the method for the present invention go context process can significantly reduce or even The aberration produced in these traditional refocusing methods is eliminated, and largely improves accurate imaging and measurement quality.
The size for splicing subgraph in the present invention is not by the size of conversion selected window, but by the grand picture of amplification original image Element is got, due to image magnification can with negated integer, in no traditional algorithm change selected window size be necessary The phenomenon that several fixed discrete depths carry out digital focus can only be chosen caused by during using single pixel as base unit.The present invention In method any depth location continuous number can be focused in measurement range, this causes depth when light-field camera measures Resolution ratio is spent with measurable section position only by the hardware limitation of measuring system, this has important for accurate measurement and imaging Meaning.
The selected window shape that the present invention uses is hexagon and unconventional rectangle or parallelogram.Due to six sides The center of shape is smaller apart from difference with each marginal position, therefore in the case where window area is identical, hexagonal window openings Grand pixel image distal-most edge selected by than rectangle or parallelogram-shaped window is farther apart from grand pixel edge, effectively avoids Choose due to the image that difference is larger, quality is relatively low formed by single lenticule marginal position in microlens array.The opposing party Face, on the premise of without departing from grand pixel edge, the maximum pixel area selected by hexagonal window openings is higher than traditional rectangle Or parallelogram-shaped window, this causes the refocusing method of the present invention to have the Depth Imaging scope of bigger.If Fig. 3 left sides are biography Image obtained by parallelogram refocusing of uniting method, the right image obtained by hexagon refocusing method of the present invention, hence it is evident that this hair Image obtained by bright method has the Depth Imaging scope of bigger, and image is also apparent.
Claims (1)
-
A kind of 1. light field image refocusing method based on hexagon stitching algorithm, it is characterised in that specifically comprise the following steps:
1), measurement before the light-field camera used is demarcated using transparent calibrated thing first, acquisition pixel representative actual ruler It is very little;
2), region to be measured illuminated using light source, adjust position and the intensity size of light source, make bias light uniform;
3), using light-field camera region income to be measured is shot, obtain without the background image for treating object;
4), hollow glass bead particulate samples to be measured are placed in region to be measured, region to be measured is shot using light-field camera, Obtaining has the particle image of object under test;
5), particle image and the gray value of each pixel of background image subtracted each other, obtain and new remove background image;
6), each grand pixel image that will go in background image amplify N times, N is the direct proportion function of refocusing depth distance d, Proportionality coefficient is by test calibration or the amplification factor N that optical field imaging systematic parameter used is calculated and depth distance The respective function relation of d obtains;
7), image carried out to amplified each grand pixel image center with the regular hexagon window of a fixed dimension answer System;
8), by step 7)The middle each grand pixel image center image copied with regular hexagon window, according to former grand pixel phase Image mosaic is carried out to link position, with reference to the amplification factor N in step 6, stitching image is reduced n times, n is the direct proportion of N Function, proportionality coefficient are determined by required final image size, you can obtain the reunion of digital focus depth location d in step 6) Burnt image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711032913.1A CN107909578A (en) | 2017-10-30 | 2017-10-30 | Light field image refocusing method based on hexagon stitching algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711032913.1A CN107909578A (en) | 2017-10-30 | 2017-10-30 | Light field image refocusing method based on hexagon stitching algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107909578A true CN107909578A (en) | 2018-04-13 |
Family
ID=61842047
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711032913.1A Pending CN107909578A (en) | 2017-10-30 | 2017-10-30 | Light field image refocusing method based on hexagon stitching algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107909578A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108921781A (en) * | 2018-05-07 | 2018-11-30 | 清华大学深圳研究生院 | A kind of light field joining method based on depth |
CN110244469A (en) * | 2019-06-19 | 2019-09-17 | 中国人民解放军陆军装甲兵学院 | A kind of determination method and system of directional scattering device position and diffusion angle |
CN111679337A (en) * | 2019-10-15 | 2020-09-18 | 上海大学 | A method for suppressing scattering background in underwater active laser scanning imaging system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014016965A (en) * | 2012-07-11 | 2014-01-30 | Toshiba Corp | Image processor, image processing method and program, and imaging device |
CN104463949A (en) * | 2014-10-24 | 2015-03-25 | 郑州大学 | Rapid three-dimensional reconstruction method and system based on light field digit refocusing |
CN104899870A (en) * | 2015-05-15 | 2015-09-09 | 清华大学深圳研究生院 | Depth estimation method based on light-field data distribution |
CN106303228A (en) * | 2016-08-04 | 2017-01-04 | 深圳市未来媒体技术研究院 | The rendering intent of a kind of focus type light-field camera and system |
-
2017
- 2017-10-30 CN CN201711032913.1A patent/CN107909578A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014016965A (en) * | 2012-07-11 | 2014-01-30 | Toshiba Corp | Image processor, image processing method and program, and imaging device |
CN104463949A (en) * | 2014-10-24 | 2015-03-25 | 郑州大学 | Rapid three-dimensional reconstruction method and system based on light field digit refocusing |
CN104899870A (en) * | 2015-05-15 | 2015-09-09 | 清华大学深圳研究生院 | Depth estimation method based on light-field data distribution |
CN106303228A (en) * | 2016-08-04 | 2017-01-04 | 深圳市未来媒体技术研究院 | The rendering intent of a kind of focus type light-field camera and system |
Non-Patent Citations (2)
Title |
---|
NICLAS ZELLER 等: "Depth estimation and camera calibration of a focused plenoptic camera for visual odometry", 《ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING》 * |
YUANCHI MA 等: "A Depth Estimation Algorithm of Plenoptic Camera for the Measurement of Particles", 《2017 IEEE INTERNATIONAL CONFERENCE ON IMAGING SYSTEMS AND TECHNIQUES (IST)》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108921781A (en) * | 2018-05-07 | 2018-11-30 | 清华大学深圳研究生院 | A kind of light field joining method based on depth |
CN108921781B (en) * | 2018-05-07 | 2020-10-02 | 清华大学深圳研究生院 | Depth-based optical field splicing method |
CN110244469A (en) * | 2019-06-19 | 2019-09-17 | 中国人民解放军陆军装甲兵学院 | A kind of determination method and system of directional scattering device position and diffusion angle |
CN110244469B (en) * | 2019-06-19 | 2021-03-23 | 中国人民解放军陆军装甲兵学院 | Method and system for determining position and diffusion angle of directional diffuser |
CN111679337A (en) * | 2019-10-15 | 2020-09-18 | 上海大学 | A method for suppressing scattering background in underwater active laser scanning imaging system |
CN111679337B (en) * | 2019-10-15 | 2022-06-10 | 上海大学 | A method for suppressing scattering background in underwater active laser scanning imaging system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Boult et al. | 1992 | Correcting chromatic aberrations using image warping. |
CN107084794B (en) | 2021-06-22 | Flame three-dimensional temperature field measuring system and method based on light field layered imaging technology |
CN104296968B (en) | 2016-12-07 | The modulation transfer function test method of multichannel CCD |
CN107403177A (en) | 2017-11-28 | Brightness measurement method based on industrial camera |
CN112767364B (en) | 2022-06-17 | Image detection system for gate blade surface corrosion and corrosion area rapid measuring and calculating method |
CN104406988A (en) | 2015-03-11 | Method for detecting defects inside glass |
CN105204173A (en) | 2015-12-30 | View synthesis correction method and device |
CN105738292A (en) | 2016-07-06 | Cell phone imaging method with colorimetric sensor array |
Wolf et al. | 2007 | Quantitative analysis of digital microscope images |
CN107909578A (en) | 2018-04-13 | Light field image refocusing method based on hexagon stitching algorithm |
CN109325927A (en) | 2019-02-12 | Brightness compensation method for photogrammetry images of industrial cameras |
CN102620668A (en) | 2012-08-01 | Method and device for measuring pixel pitch of image sensor based on point-target image splicing technology |
CN101907490B (en) | 2012-02-08 | Measuring method of tiny light spot intensity distribution based on two-dimensional subdivision method |
CN103091993B (en) | 2015-02-11 | Test mark and measuring method used in lithography machine lens thermal effect measuring |
CN109557091B (en) | 2021-09-28 | Disordered medium channel analysis device and method |
CN110108230B (en) | 2021-04-16 | Binary grating projection defocus degree evaluation method based on image difference and LM iteration |
US9542969B2 (en) | 2017-01-10 | Optical recording medium and optical information playback method |
CN102620669B (en) | 2014-08-13 | Method and device for measuring pixel pitch of image sensor by utilizing constant moving point target |
CN114234846B (en) | 2022-11-08 | A Fast Nonlinear Compensation Method Based on Double Response Curve Fitting |
CN115451820A (en) | 2022-12-09 | Three-channel polarization information acquisition system |
Li et al. | 2018 | Local error and its identification for microlens array in plenoptic camera |
CN106468562A (en) | 2017-03-01 | A kind of color camera radial direction aberration calibration steps based on absolute phase |
CN102607444B (en) | 2014-07-16 | Method and device for measuring space of pixels of image sensor by using linear light source |
CN110095472B (en) | 2019-12-13 | HDRI-based high-reflectivity metal surface defect detection method and system |
CN102620667A (en) | 2012-08-01 | Method and device for measuring pixel pitch of image sensor based on point-target image splicing technology |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
2018-04-13 | PB01 | Publication | |
2018-04-13 | PB01 | Publication | |
2018-05-08 | SE01 | Entry into force of request for substantive examination | |
2018-05-08 | SE01 | Entry into force of request for substantive examination | |
2020-12-29 | WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180413 |
2020-12-29 | WD01 | Invention patent application deemed withdrawn after publication |