CN106846383A - High dynamic range images imaging method based on 3D digital micro-analysis imaging systems - Google Patents
High dynamic range images imaging method based on 3D digital micro-analysis imaging systems Download PDFInfo
- Publication number
- CN106846383A CN106846383A CN201710057799.1A CN201710057799A CN106846383A CN 106846383 A CN106846383 A CN 106846383A CN 201710057799 A CN201710057799 A CN 201710057799A CN 106846383 A CN106846383 A CN 106846383A
- Authority
- CN
- China
- Prior art keywords
- image
- high dynamic
- dynamic range
- sequence
- range images
- 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
Links
- 238000003384 imaging method Methods 0.000 title claims abstract description 40
- 238000004452 microanalysis Methods 0.000 title claims abstract description 11
- 238000000034 method Methods 0.000 claims abstract description 58
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 13
- 230000009466 transformation Effects 0.000 claims abstract description 4
- 230000000694 effects Effects 0.000 claims abstract description 3
- 230000011218 segmentation Effects 0.000 claims abstract description 3
- 238000009499 grossing Methods 0.000 claims abstract 2
- 230000008569 process Effects 0.000 claims description 25
- 239000011159 matrix material Substances 0.000 claims description 20
- 230000015556 catabolic process Effects 0.000 claims description 11
- 238000005286 illumination Methods 0.000 claims description 11
- 238000013507 mapping Methods 0.000 claims description 10
- 238000001514 detection method Methods 0.000 claims description 8
- 238000005192 partition Methods 0.000 claims description 6
- 230000004044 response Effects 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 230000007812 deficiency Effects 0.000 abstract description 3
- 238000001914 filtration Methods 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 7
- 230000007547 defect Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000004438 eyesight Effects 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 238000000386 microscopy Methods 0.000 description 2
- 206010021703 Indifference Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000001000 micrograph Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012372 quality testing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G02—OPTICS
- G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
- G02B21/00—Microscopes
- G02B21/36—Microscopes arranged for photographic purposes or projection purposes or digital imaging or video purposes including associated control and data processing arrangements
- G02B21/365—Control or image processing arrangements for digital or video microscopes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
- G06T2207/10061—Microscopic image from scanning electron microscope
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- General Physics & Mathematics (AREA)
- Optics & Photonics (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Microscoopes, Condenser (AREA)
Abstract
The present invention relates to a kind of high dynamic range images imaging method of base 3D digital micro-analysis imaging systems, by generating the high dynamic range images of object to be observed and obtaining the original high dynamic multi-focus sequence image of sample to be observed, the movement in image registration and super-pixel level is carried out using phase matching method and Fourier transformation, then is split target object by prospect background dividing method；Make quadtree decomposition treatment, the picture rich in detail block in mark image sequence for the image after segmentation, and record the elevation information corresponding to per piece image；The picture rich in detail block that will finally mark is fused into the three-dimensional shape of logistics to be observed, and the three-dimensional shape for generating is filtered using medium filtering, to eliminate the sawtooth effect that three-dimensional shape causes by sample frequency deficiency, so that the 3 D stereo of the object to be observed of generation is formed more smoothing.
Description
Technical field
The present invention relates to high definition high accuracy micro-imaging detection technique field, more particularly to it is a kind of based on 3D digital micro-analysis into
As the high dynamic range images imaging method of system.
Background technology
Many focal length 3D technologies (Shape from Focus, abbreviation SFF) are normal in current analysis of digital microscopy images process field
3D technology.Only need to be obtained with using traditional monocular microscope the three-dimensional shaped of observation sample due to many focal length 3D technologies
Shape and obtain the extensive concern of experts and scholars.It is different from stereovision technique and obtains depth information, many focal lengths using binocular head
3D technology is only by mobile, the distance of observed objects to camera lens, clear area in detection image, so as to just can be gone out with restoration and reconstruction
The depth information of object.
But, the major defect of many focal length 3D technologies is when observation sample has the reflective situation of height, due to collection
The dynamic range of the image for obtaining is not enough and cause the image detail in some regions not enough, or even there be not image in some regions
Details, thus leverages the accuracy rate of the object three-dimensional form after rebuilding.However, many scientific researches at present are still
The influence that (ionospheric) focussing factor rebuilds accuracy rate to object three-dimensional form is principally dedicated to, original image is but have ignored in dynamic range side
Influence of the quality in face to object three-dimensional form reconstructed results.
In order to overcome the influence of dynamic range deficiency factor in resulting image, high dynamic range imaging technology is suggested.
High dynamic range images (High-Dynamic Range, abbreviation HDR) can be obtained using high dynamic range imaging technology.Pass through
Demarcate, the image for the different exposure time of same scene is merged, and can obtain the high dynamic range of 32 of the scene
Enclose illumination spectrum.The dynamic range that these 32 illumination spectrogram pictures accurately, truly can reflect in scene, then by part
These 32 illumination spectrogram pictures are mapped to the normal image of 8 for tone mapping, consequently facilitating conventional display apparatus show with
Preserve these normal images of 8.But, due to the computation complexity high of high dynamic range imaging technology, at present on the market
Micro- 3D method for reconstructing is still present limitation realizing this technical elements.
The content of the invention
The technical problems to be solved by the invention are directed to above-mentioned prior art and provide a kind of based on the imaging of 3D digital micro-analysis
The high dynamic range images imaging method of system.The high dynamic range images imaging method can solve the problem that conventional images imaging method
The defect of aloof from politics and material pursuits dynamic scene cannot be clapped, and the three-dimensional shape of object to be observed can be simultaneously accurately generated, so that
Enjoyed for observer provides comprehensive 3D stereoscopic visions.
The present invention solve the technical scheme that is used of above-mentioned technical problem for：Height based on 3D digital micro-analysis imaging systems is moved
State range image imaging method, it is characterised in that comprise the following steps：
Step 1, for the object to be observed on microscope carrier, by adjusting the height of objective table, and utilizes camera
The high dynamic multiple focussing image of each aspect being obtained from the top of object bottom to be observed to object to be observed, it is three-dimensional vertical to obtain
Original high dynamic multi-focus sequence image needed for body imaging；
Step 2, registration is carried out using phase matching method to the original high dynamic multi-focus sequence image of gained, to cause institute
Locus, the zoom scale for stating front and rear connected image pair in original high dynamic multi-focus sequence image are corresponding with picture size
Unanimously, the high dynamic multi-focus sequence image good so as to obtain registration；
Step 3, for the good high dynamic multi-focus sequence image of registration, the prospect background dividing method accumulated using background
Extraction needs to generate the observation sample region of 3 D stereo；
Step 4, is split to the observation sample region using Quadtree Partition method, and detection high dynamic multi-focus
Clear part in every piece image of sequence image, and record the elevation information corresponding to per piece image；
Step 5, partly merges, to clear in each width image for detecting so as to generate the three of object to be observed
Dimension three-dimensional shape.
Further, the process of the high dynamic range images for obtaining each aspect using camera in the step 1 includes：
The response curve of (a) calibration for cameras；B () obtains the image of exposure values different in Same Scene；C () is using mark
The response curve of fixed camera, generates the illumination spectrogram of 32 of the scene；D () will be described using local tone mapping
The illumination spectrogram of 32 maps to the normal image of 8, and it is that computer can show and store to preserve the normal image
Form.
Further, in step 1, the acquisition process of the original high dynamic multi-focus sequence image includes：First, lead to
Cross the height of moving stage, change the distance between object to be observed and microscopical object lens, realize that monocular microscope is different
Focussing plane image sequence；Secondly, the requirement of each width focussing plane picture altitude information is recorded；Again, it is poly- for each width
Image focal plane is focused detection, and records the pixel with maximum focus degree in each width focussing plane image
Point, rebuilds for follow-up three-dimensional shape.
Specifically, in step 2, the phase matching method is matched somebody with somebody to the original high dynamic multi-focus sequence image of gained
Accurate process includes：
First, in the original poly- sequence image of high dynamic poly, for each image pair being connected before and after every two width, will
Each image of image pair is converted to gray level image, so as to obtain gray level image pair；
Secondly, the gray level image centering using complex bandpass filters from after conversion extracts the phase letter of each frequency range
Breath；
Again, using the phase information extracted, realize the gray level image in super-pixel by Fourier transformation
Movement in level, with the uniformity of the position of connected two images before and after ensureing；
Finally, for each group of image pair in original high dynamic multi-focus sequence image, the process is repeated, until height is dynamic
The zoom scale of all images and displacement are consistent in state multiple focussing image sequence.
Specifically, included using the process in Quadtree Partition method segmentation observation sample region in the step 4：
First, it is input to original high dynamic multi-focus sequence image as one layer of quaternary tree root in quaternary tree；
Secondly, picture breakdown condition is set, and whether each tomographic image in quaternary tree is met at decomposition condition
Reason：
If meeting described picture breakdown condition for a tomographic image, four fork decomposition are carried out to this tomographic image, and it is defeated
Enter the next layer to quaternary tree；The like, until image sequence be decomposed gained minimum image block be all unsatisfactory for it is described
Picture breakdown condition, then terminate quadtree decomposition process；Wherein, the picture breakdown condition for setting as：
Its (ionospheric) focussing factor maximum difference is calculated respectively for the image block that each layer in image sequence in quaternary tree is decomposed
Value MDFM and gradient disparities value SMDG；Wherein, the computing formula of (ionospheric) focussing factor maximum different value MDFM and gradient disparities value SMDG
It is as follows respectively：
MDFM=FMmax-FMmin；
Wherein, FMmaxRepresent the maximum of focometry, FMminRepresent the minimum value of focometry；gradmax(x, y) table
Show greatest gradient value, gradmin(x, y) represents minimal gradient value；
A tomographic image block in for quaternary tree, if meeting MDFM >=0.98 × SMDG, in the upper layer images sequence of surface
In the presence of the image block being fully focused, then the tomographic image block will not continue decomposition downwards；Conversely, the tomographic image block is continued to point
Solution is gone down, until all images are all broken down into the subimage block that cannot be decomposed in quaternary tree.
Specifically, the maximum FM of the focometrymax, focometry minimum value FMminAcquisition process be：
First, the gradient matrix of each pixel in a tomographic image of quaternary tree root is calculated, computing formula is：
GMi=gradient (Ii), i=1,2 ..., n；
Wherein, IiIt is i-th original high dynamic multiple focussing image, GMiIt is and IiCorresponding gradient matrix；N is original height
Image total number in dynamic multi-focus sequence image；
Secondly, the gradient matrix of maximum in all gradient matrixs of this tomographic image every bit and the gradient of minimum are found
Matrix, formula is as follows：
GMmax=max (GMi(x, y)), i=1,2 ..., n；
GMmin=min (GMi(x, y)), i=1,2 ..., n；
Again, calculate this tomographic image gradient matrix sum a little, computing formula is as follows：
FMi=ΣxΣygradi(x, y), i=1,2 ..., n；
Finally, the maximum and minimum value of above-mentioned gradient matrix sum are found respectively, and computing formula is as follows：
FMmax=max { FMi, i=1,2 ..., n；
FMmin=min { FMi, i=1,2 ..., n.
Specifically, the clear process for partly being merged in the step 5 for each width image includes：For gained institute
The clear part having records its elevation information, and all of clearly subimage block is melted respectively as clearly subimage block
Synthesize the three-dimensional image of the complete observation sample of a width.
With improvement, also include in the step 5：The three-dimensional shape for generating is filtered using median filter method
Ripple, to eliminate the sawtooth effect that three-dimensional shape causes by sample frequency deficiency, so that the 3 D stereo shape of generation
Into more smooth.
Compared with prior art, the advantage of the invention is that：
First, high dynamic range images imaging method provided by the present invention employs high dynamic range imaging, three-dimensional and stands
Body is imaged and many depth image integration technologies, while obtaining the image sequence of the different exposure time of Same Scene, generates scene
The illumination spectrogram of 32, the illumination spectrogram of 32 is then mapped into the normal image of 8 using local tone, and preserve into
The generation high dynamic range video form that computer can show, store, using tone mapping technique, display in real time and transmission are high
Dynamic range microscopy video, is easy to observer to dynamically watch object to be observed in real time；
Secondly as the computation complexity that the application of high dynamic range video technology is related to is higher, the present invention uses phase
Position matching, the method such as Quadtree Partition, can real-time processing vision signal, the real-time microscopy video of generation shows, so as to reduce
Computation complexity；
Again, the high dynamic range images imaging method in the present invention can observe object to be observed with real-time high-definition, gram
Having taken current picture imaging techniques cannot be while sees the defect in reflective and non-reflective region clearly to high-contrast sample；
Finally, high dynamic range images imaging method of the invention be obtained in that by the different image of focus synthesize it is complete
Focus image；In the different image process for the treatment of focus, by the automatic depth for obtaining image midpoint, so as to recover image
The three-dimensional coordinate put on surface, ensures for brand-new material quality testing provides strong auxiliary.
Brief description of the drawings
Fig. 1 is the high dynamic range images imaging method stream based on 3D digital micro-analysis imaging systems in the embodiment of the present invention one
Journey schematic diagram；
Fig. 2 is the schematic diagram of 3D digital micro-analysis imaging systems in the embodiment of the present invention one；
Original high dynamic multi-focus sequence images of the Fig. 3 corresponding to metallic screw in embodiment one；
Fig. 4 is the high dynamic range images and common automatic exposure image comparison of metallic screw acquired in embodiment one
Figure；Wherein, left side one is classified as corresponding high dynamic range images, and right side one is classified as corresponding common automatic exposure image；
Fig. 5 is the schematic diagram of extraction foreground image in embodiment one；
Fig. 6 a are the 3D stereo-pictures mapped without image texture generated using high dynamic range images in embodiment one；
Fig. 6 b are the 3D stereo-pictures for having image texture to map generated using high dynamic range images in embodiment one；
Fig. 6 c are the 3D stereograms mapped without image texture generated using original automatic exposure image in embodiment one
Picture；
Fig. 6 d are the 3D stereo-pictures for having image texture to map generated using original automatic exposure image in embodiment one；
Fig. 6 e are the true value figure of the 3D stereo-pictures for not having image texture mapping in embodiment one；
Fig. 6 f are the true value figure of the 3D stereo-pictures for having image texture mapping in embodiment one；
Fig. 7 a are the 3D stereo-pictures mapped without image texture generated using high dynamic range images in embodiment two；
Fig. 7 b are the 3D stereo-pictures for having image texture to map generated using high dynamic range images in embodiment two；
Fig. 7 c are the 3D stereograms mapped without image texture generated using original automatic exposure image in embodiment two
Picture；
Fig. 7 d are the 3D stereo-pictures for having image texture to map generated using original automatic exposure image in embodiment two；
Fig. 7 e are the true value figure of the 3D stereo-pictures for not having image texture mapping in embodiment two；
Fig. 7 f are the true value figure of the 3D stereo-pictures for having image texture mapping in embodiment two；
Fig. 8 is to generate in embodiment two 3D three-dimensional shapes method and not using high dynamic using high dynamic range images
The square root error comparison diagram of range image；
Fig. 9 a are the 3D stereo-pictures mapped without image texture generated using high dynamic range images in embodiment three；
Fig. 9 b are the 3D stereo-pictures for having image texture to map generated using high dynamic range images in embodiment three；
Fig. 9 c are the 3D stereograms mapped without image texture generated using original automatic exposure image in embodiment three
Picture；
Fig. 9 d are the 3D stereo-pictures for having image texture to map generated using original automatic exposure image in embodiment three；
Fig. 9 e are the true value figure of the 3D stereo-pictures for not having image texture mapping in embodiment three；
Fig. 9 f are the true value figure of the 3D stereo-pictures for having image texture mapping in embodiment three；
Figure 10 is that high dynamic range images generate 3D three-dimensional shapes method and original automatic exposure image generation 3D is three-dimensional
Method On Shape generates square root error comparison diagram corresponding to 3D stereo-pictures.
Specific embodiment
The present invention is described in further detail below in conjunction with accompanying drawing embodiment.
Embodiment one
As shown in Fig. 2 the 3D digital micro-analysis imaging systems employed in the present embodiment one include conventional optical microscope,
Automatic carrier, CMOS cameras and the computer that can be moved on any direction of X-axis, Y-axis and Z axis.Wherein, the present embodiment
Object to be observed in one is metallic screw, and metallic screw is placed on automatic carrier.Shown in Fig. 1, the present embodiment
The high dynamic range images imaging method based on 3D digital micro-analysis imaging systems comprises the following steps in one：
Step 1, for the object to be observed on microscope carrier, i.e. metallic screw, by adjusting the height of objective table,
So that CMOS cameras are focused in each aspect of object to be observed, that is, in each aspect of metallic screw, and utilize phase
Machine is obtained from the high dynamic multiple focussing image of each aspect at the top of metallic screw bottom to metallic screw, to obtain 3 D stereo
Original high dynamic multi-focus sequence image needed for imaging；For metallic screw original high dynamic multi-focus sequence image referring to
Shown in Fig. 3；Wherein, the high dynamic multiple focussing image process of acquisition object to be observed includes high dynamic range images and obtains and many
Focusedimage obtains two processes；Specifically, the process for obtaining the high dynamic range images of each aspect of object to be observed includes：
The response curve of (a) calibration for cameras；B () obtains the image of exposure values different in Same Scene；C () is using mark
The response curve of fixed camera, generates the illumination spectrogram of 32 of scene；D () is mapped using local tone, by the illumination of 32
Spectrogram maps to the normal image of 8, and preserves the form that normal image can show and store for computer.Metallic screw institute
Corresponding every layer high dynamic range images arrange shown referring to the left side one in Fig. 4；
Step 2, registration is carried out using phase matching method to the original high dynamic multi-focus sequence image of gained, to cause original
The locus of front and rear connected image pair, zoom scale corresponding with picture size one in beginning high dynamic multi-focus sequence image
Cause, the high dynamic multi-focus sequence image good so as to obtain registration；Wherein, phase matching method is to the original high dynamic poly of gained
The process that burnt sequence image carries out registration includes：
First, in the original poly- sequence image of high dynamic poly, for each image pair being connected before and after every two width, by image
Each image of centering is converted to gray level image, so as to obtain gray level image pair；
Secondly, the gray level image centering using complex bandpass filters from after conversion extracts the phase letter of each frequency range
Breath；
Again, using the phase information extracted, realize gray level image in super-pixel level by Fourier transformation
It is mobile, with the uniformity of the position of connected two images before and after ensureing；
Finally, for each group of image pair in original high dynamic multi-focus sequence image, the process is repeated, until height is dynamic
The zoom scale of all images and displacement are consistent in state multiple focussing image sequence.
Step 3, for the good high dynamic multi-focus sequence image of registration, the prospect background dividing method accumulated using background
Extraction needs to generate the observation sample region of 3 D stereo；It is shown in Figure 5, i.e., using inter-frame difference, by metallic screw correspondence
Background extracting out, then enter row threshold division, so as to obtain foreground image；
Step 4, is split to observation sample region using Quadtree Partition method, and detection high dynamic multi-focus sequence
Clear part in every piece image of image, and record the elevation information corresponding to per piece image；Wherein,
It is described as follows for the Quadtree Partition method in the present embodiment one：
First, it is input to original high dynamic multi-focus sequence image as one layer of quaternary tree root in quaternary tree；
Secondly, picture breakdown condition is set, and whether each tomographic image in quaternary tree is met at decomposition condition
Reason：
If meeting the picture breakdown condition for a tomographic image, four fork decomposition are carried out to this tomographic image, and be input to
Next layer of quaternary tree；The like, until image sequence be decomposed gained minimum image block be all unsatisfactory for picture breakdown bar
Part, then terminate quadtree decomposition process；Wherein, it is as follows for picture breakdown condition stub：
Its (ionospheric) focussing factor maximum difference is calculated respectively for the image block that each layer in image sequence in quaternary tree is decomposed
Value MDFM and gradient disparities value SMDG；Wherein, the computing formula of (ionospheric) focussing factor maximum different value MDFM and gradient disparities value SMDG
It is as follows respectively：
MDFM=FMmax-FMmin；
Wherein, FMmaxRepresent the maximum of focometry, FMminRepresent the minimum value of focometry；gradmax(x, y) table
Show greatest gradient value, gradmin(x, y) represents minimal gradient value；For the maximum FM of focometrymax, focometry most
Small value FMminCalculated case be：
First, the gradient matrix of each pixel in a tomographic image of quaternary tree root is calculated, computing formula is：
GMi=gradient (Ii), i=1,2 ..., n；
Wherein, IiIt is i-th original high dynamic multiple focussing image, GMiIt is and IiCorresponding gradient matrix；N is original height
Image total number in dynamic multi-focus sequence image；
Secondly, the gradient matrix of maximum in all gradient matrixs of this tomographic image every bit and the gradient of minimum are found
Matrix, formula is as follows：
GMmax=max (GMi(x, y)), i=1,2 ..., n；
GMmin=min (GMi(x, y)), i=1,2 ..., n；
Again, calculate this tomographic image gradient matrix sum a little, computing formula is as follows：
FMi=Σx∑ygradi(x, y), i=1,2 ..., n；
Finally, the maximum and minimum value of above-mentioned gradient matrix sum are found respectively, and computing formula is as follows：
FMmax=max { FMi, i=1,2 ..., n；FMmin=min { FMi, i=1,2 ..., n.
It is described as follows for the clear partial routine in every piece image of detection high dynamic multi-focus sequence image：
For each image block sequence in quaternary tree, an image of gradient matrix maximum in image block sequence is found
Block, and image block that this has greatest gradient matrix is recorded in the position of image sequence and its elevation information；
I=1,2 ..., n；Wherein, fmi(x, y) represents i-th figure in image sequence
The gradient matrix of picture.
Step 5, partly merges, to clear in each width image for detecting so as to generate the three of object to be observed
Dimension three-dimensional shape, that is, metallic screw three-dimensional image；Wherein, the corresponding three-dimensional image mark of setting metallic screw
It is designated as Z：
Z (x, y)=zi(x, y), zi(x, y) represents the clearly image block of i-th in image sequence.
In order to use high dynamic during tradition is generated into 3D three-dimensional shapes method and the present invention using original automatic exposure image
Range image generation 3D three-dimensional shape methods are compared, and the present embodiment one gives metallic screw and is utilized respectively above two 3D
Three-dimensional shape method generates comparing figure corresponding to stereo-picture, shown in Fig. 6.In the present invention, will be using former
Start from dynamic exposure image technology and be designated as Normal SFF, HDR-SFF will be designated as using high dynamic range images technology.Wherein：
In order to contrast the accuracy rate of above two 3D three-dimensional shape generation methods, the embodiment of the present invention one is by introducing square
Root error, to weigh two kinds of 3D three-dimensional shapes generation methods under the same conditions, the gap between true value：
Wherein, GT(i, j) represents true value, and Z (i, j) represents Normal SFF or HDR-SFF values.
Table 1 gives two kinds of 3D three-dimensional shapes generation methods and is using 22 kinds of different (ionospheric) focussing factors, and it is flat that correspondence is obtained
Square error.Be can be seen that for same (ionospheric) focussing factor by the result in contrast table 1, given birth to using high dynamic range images
Into the corresponding (ionospheric) focussing factor square root error amount of 3D three-dimensional shapes be less than not using high dynamic range images generation 3D
The corresponding (ionospheric) focussing factor square root error amount of three-dimensional shape.Result in table 1 shows, uses high dynamic range images in the present invention
The 3D three-dimensional shapes of generation are more accurate than without the 3D three-dimensional shapes generated using high dynamic range images.
Table 1
Embodiment two
It is small with one on bank card using a kind of bank card of plastic material as object to be observed in the present embodiment two
Write English alphabet " d ".Wherein, the step of generating its three-dimensional image for the bank card and metallic screw in embodiment one
The generation step of three-dimensional image is identical, and here is omitted.
In the present embodiment two, in order to verify the accuracy and robust of high dynamic range images imaging method in the present invention
Property, the present embodiment two gives the bank card and generates corresponding high dynamic range images, referring specifically to shown in Fig. 7 a~Fig. 7 f.
Fig. 8 is, for the bank card in the present embodiment two, to generate 3D three-dimensional shapes method using high dynamic range images and do not make
With the square root error comparison diagram of high dynamic range images.
As seen from Figure 8, for same (ionospheric) focussing factor, the 3D three-dimensional shapes pair generated using high dynamic range images
The 3D three-dimensional shapes that the (ionospheric) focussing factor square root error amount answered is less than not using high dynamic range images generation are corresponding poly-
Burnt factor square root error amount.It can be seen that, the 3D three-dimensional shapes ratio generated using high dynamic range images in the present invention is not used
The 3D three-dimensional shapes of high dynamic range images generation are more accurate.
Embodiment three
Using metal chip as object to be observed in the present embodiment three.Wherein, generated for the metal chip thirdly
The step of dimension stereo-picture, is identical with the generation step of metallic screw three-dimensional image in embodiment one, and here is omitted.
Figure 10 is that high dynamic range images generate 3D three-dimensional shapes method and original automatic exposure image generation 3D is three-dimensional
Method On Shape generates square root error comparison diagram corresponding to 3D stereo-pictures.
As seen from Figure 10, for same (ionospheric) focussing factor, the 3D three-dimensional shapes generated using high dynamic range images
The 3D three-dimensional shapes that corresponding (ionospheric) focussing factor square root error amount is less than not using high dynamic range images generation are corresponding
(ionospheric) focussing factor square root error amount.It can be seen that, the 3D three-dimensional shapes ratio generated using high dynamic range images in the present invention is not made
The 3D three-dimensional shapes generated with high dynamic range images are more accurate.
Claims (8)
1. the high dynamic range images imaging method of 3D digital micro-analysis imaging systems is based on, it is characterised in that comprised the following steps：
Step 1, for the object to be observed on microscope carrier, by adjusting the height of objective table, and is obtained using camera
The high dynamic multiple focussing image of each aspect at the top of from object bottom to be observed to object to be observed, with obtain 3 D stereo into
Original high dynamic multi-focus sequence image as needed for；
Step 2, registration is carried out using phase matching method to the original high dynamic multi-focus sequence image of gained, to cause the original
The locus of front and rear connected image pair, zoom scale corresponding with picture size one in beginning high dynamic multi-focus sequence image
Cause, the high dynamic multi-focus sequence image good so as to obtain registration；
Step 3, for the good high dynamic multi-focus sequence image of registration, the prospect background dividing method accumulated using background is extracted
Need the observation sample region of generation 3 D stereo；
Step 4, is split to the observation sample region using Quadtree Partition method, and detection high dynamic multi-focus sequence
Clear part in every piece image of image, and record the elevation information corresponding to per piece image；
Step 5, partly merges, to clear in each width image for detecting so that the three-dimensional for generating object to be observed is stood
Shape.
2. high dynamic range images imaging method according to claim 1, it is characterised in that phase is utilized in the step 1
The process that machine obtains the high dynamic range images of each aspect includes：
The response curve of (a) calibration for cameras；B () obtains the image of exposure values different in Same Scene；C () is using demarcation
The response curve of camera, generates the illumination spectrogram of 32 of the scene；D () is using local tone mapping by described 32
Illumination spectrogram map to the normal image of 8, and preserve the form that the normal image can be shown and be stored for computer.
3. high dynamic range images imaging method according to claim 1, it is characterised in that in step 1, described original
The acquisition process of high dynamic multi-focus sequence image includes：
In step 1, first, by the height of moving stage, change between object to be observed and microscopical object lens away from
From realizing monocular microscope difference focussing plane image sequence；Secondly, wanting for each width focussing plane picture altitude information is recorded
Ask；Again, it is focused detection for each width focussing plane image, and records in each width focussing plane image and have
The pixel of maximum focus degree, rebuilds for follow-up three-dimensional shape.
4. high dynamic range images imaging method according to claim 1, it is characterised in that in step 2, the phase
The process that matching process carries out registration to the original high dynamic multi-focus sequence image of gained includes：
First, in the original poly- sequence image of high dynamic poly, for each image pair being connected before and after every two width, by image
Each image of centering is converted to gray level image, so as to obtain gray level image pair；
Secondly, the gray level image centering using complex bandpass filters from after conversion extracts the phase information of each frequency range；
Again, using the phase information extracted, realize the gray level image in super-pixel level by Fourier transformation
On movement, with the uniformity of the position of connected two images before and after ensureing；
Finally, for each group of image pair in original high dynamic multi-focus sequence image, the process is repeated, until high dynamic is more
The zoom scale of all images and displacement are consistent in focusedimage sequence.
5. high dynamic range images imaging method according to claim 1, it is characterised in that four are used in the step 4
The process in fork tree automatic Segmentation observation sample region includes：
First, it is input to original high dynamic multi-focus sequence image as one layer of quaternary tree root in quaternary tree；
Secondly, picture breakdown condition is set, and whether each tomographic image in quaternary tree meets decomposition condition and is processed：
If meeting described picture breakdown condition for a tomographic image, four fork decomposition are carried out to this tomographic image, and be input to
Next layer of quaternary tree；The like, until the be decomposed minimum image block of gained of image sequence is all unsatisfactory for described image
Decomposition condition, then terminate quadtree decomposition process；Wherein, the picture breakdown condition for setting as：
Its (ionospheric) focussing factor maximum different value is calculated respectively for the image block that each layer in image sequence in quaternary tree is decomposed
MDFM and gradient disparities value SMDG；Wherein, the computing formula of (ionospheric) focussing factor maximum different value MDFM and gradient disparities value SMDG point
It is not as follows：
MDFM=FMmax-FMmin；
Wherein, FMmaxRepresent the maximum of focometry, FMminRepresent the minimum value of focometry；gradmax(x, y) is represented most
Big Grad, gradmin(x, y) represents minimal gradient value；
A tomographic image block in for quaternary tree, if meeting MDFM >=0.98 × SMDG, exists in the upper layer images sequence of surface
The image block being fully focused, then the tomographic image block will not continue downwards decomposition；Conversely, under the tomographic image block continues to decompose
Go, until all images are all broken down into the subimage block that cannot be decomposed in quaternary tree.
6. high dynamic range images imaging method according to claim 5, it is characterised in that the maximum of the focometry
Value FMmax, focometry minimum value FMminAcquisition process be：
First, the gradient matrix of each pixel in a tomographic image of quaternary tree root is calculated, computing formula is：
GMi=gradient (Ii), i=1,2 ..., n；
Wherein, IiIt is i-th original high dynamic multiple focussing image, GMiIt is and IiCorresponding gradient matrix；N is original high dynamic
Image total number in multi-focus sequence image；
Secondly, the gradient matrix of maximum in all gradient matrixs of this tomographic image every bit and the gradient square of minimum are found
Battle array, formula is as follows：
GMmax=max (GMi(x, y)), i=1,2 ..., n；
GMmin=min (GMi(x, y)), i=1,2 ..., n；
Again, calculate this tomographic image gradient matrix sum a little, computing formula is as follows：
FMi=ΣxΣygradi(x, y), i=1,2 ..., n；
Finally, the maximum and minimum value of above-mentioned gradient matrix sum are found respectively, and computing formula is as follows：
FMmax=max { FMi, i=1,2 ..., n；
FMmin=min { FMi, i=1,2 ..., n.
7. high dynamic range images imaging method according to claim 5, it is characterised in that for each in the step 5
The clear process for partly being merged of width image includes：For all of clear part of gained as clearly subimage block,
Its elevation information is recorded respectively, and all of clearly subimage block is fused into the 3 D stereo of the complete observation sample of a width
Image.
8. high dynamic range images imaging method according to claim 1, it is characterised in that also include in the step 5：
The three-dimensional shape for generating is filtered using median filter method, to eliminate three-dimensional shape because sample frequency is not enough
And the sawtooth effect for causing, so that the 3 D stereo of generation is formed more smoothing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710057799.1A CN106846383B (en) | 2017-01-23 | 2017-01-23 | High dynamic range image imaging method based on 3D digital microscopic imaging system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710057799.1A CN106846383B (en) | 2017-01-23 | 2017-01-23 | High dynamic range image imaging method based on 3D digital microscopic imaging system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106846383A true CN106846383A (en) | 2017-06-13 |
| CN106846383B CN106846383B (en) | 2020-04-17 |
Family
ID=59121732
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710057799.1A Expired - Fee Related CN106846383B (en) | 2017-01-23 | 2017-01-23 | High dynamic range image imaging method based on 3D digital microscopic imaging system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106846383B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392946A (en) * | 2017-07-18 | 2017-11-24 | 宁波永新光学股份有限公司 | A kind of micro- multiple focal length images series processing method rebuild towards 3D shape |
| CN107680152A (en) * | 2017-08-31 | 2018-02-09 | 太原理工大学 | Target surface topography measurement method and apparatus based on image procossing |
| CN108470149A (en) * | 2018-02-14 | 2018-08-31 | 天目爱视（北京）科技有限公司 | A kind of 3D 4 D datas acquisition method and device based on light-field camera |
| CN109360163A (en) * | 2018-09-26 | 2019-02-19 | 深圳积木易搭科技技术有限公司 | A kind of fusion method and emerging system of high dynamic range images |
| CN110197463A (en) * | 2019-04-25 | 2019-09-03 | 深圳大学 | High dynamic range image tone mapping method and its system based on deep learning |
| CN112489196A (en) * | 2020-11-30 | 2021-03-12 | 太原理工大学 | Particle three-dimensional shape reconstruction method based on multi-scale three-dimensional frequency domain transformation |
| CN110849266B (en) * | 2019-11-28 | 2021-05-25 | 江西瑞普德测量设备有限公司 | Telecentric lens telecentricity debugging method of image measuring instrument |
| CN117784388A (en) * | 2024-02-28 | 2024-03-29 | 宁波永新光学股份有限公司 | High dynamic range metallographic image generation method based on camera response curve |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100194902A1 (en) * | 2009-02-05 | 2010-08-05 | National Chung Cheng University | Method for high dynamic range imaging |
| CN103946732A (en) * | 2011-09-26 | 2014-07-23 | 微软公司 | Video display modification based on sensor input for a see-through near-to-eye display |
| CN104224127A (en) * | 2014-09-17 | 2014-12-24 | 西安电子科技大学 | Optical projection tomography device and method based on camera array |
| US20160125630A1 (en) * | 2014-10-30 | 2016-05-05 | PathPartner Technology Consulting Pvt. Ltd. | System and Method to Align and Merge Differently Exposed Digital Images to Create a HDR (High Dynamic Range) Image |
-
2017
- 2017-01-23 CN CN201710057799.1A patent/CN106846383B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100194902A1 (en) * | 2009-02-05 | 2010-08-05 | National Chung Cheng University | Method for high dynamic range imaging |
| CN103946732A (en) * | 2011-09-26 | 2014-07-23 | 微软公司 | Video display modification based on sensor input for a see-through near-to-eye display |
| CN104224127A (en) * | 2014-09-17 | 2014-12-24 | 西安电子科技大学 | Optical projection tomography device and method based on camera array |
| US20160125630A1 (en) * | 2014-10-30 | 2016-05-05 | PathPartner Technology Consulting Pvt. Ltd. | System and Method to Align and Merge Differently Exposed Digital Images to Create a HDR (High Dynamic Range) Image |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392946A (en) * | 2017-07-18 | 2017-11-24 | 宁波永新光学股份有限公司 | A kind of micro- multiple focal length images series processing method rebuild towards 3D shape |
| CN107392946B (en) * | 2017-07-18 | 2020-06-16 | 宁波永新光学股份有限公司 | Microscopic multi-focus image sequence processing method for three-dimensional shape reconstruction |
| CN107680152A (en) * | 2017-08-31 | 2018-02-09 | 太原理工大学 | Target surface topography measurement method and apparatus based on image procossing |
| CN108470149A (en) * | 2018-02-14 | 2018-08-31 | 天目爱视（北京）科技有限公司 | A kind of 3D 4 D datas acquisition method and device based on light-field camera |
| CN109360163A (en) * | 2018-09-26 | 2019-02-19 | 深圳积木易搭科技技术有限公司 | A kind of fusion method and emerging system of high dynamic range images |
| CN110197463A (en) * | 2019-04-25 | 2019-09-03 | 深圳大学 | High dynamic range image tone mapping method and its system based on deep learning |
| CN110197463B (en) * | 2019-04-25 | 2023-01-03 | 深圳大学 | High dynamic range image tone mapping method and system based on deep learning |
| CN110849266B (en) * | 2019-11-28 | 2021-05-25 | 江西瑞普德测量设备有限公司 | Telecentric lens telecentricity debugging method of image measuring instrument |
| CN112489196A (en) * | 2020-11-30 | 2021-03-12 | 太原理工大学 | Particle three-dimensional shape reconstruction method based on multi-scale three-dimensional frequency domain transformation |
| CN112489196B (en) * | 2020-11-30 | 2022-08-02 | 太原理工大学 | Particle three-dimensional shape reconstruction method based on multi-scale three-dimensional frequency domain transformation |
| CN117784388A (en) * | 2024-02-28 | 2024-03-29 | 宁波永新光学股份有限公司 | High dynamic range metallographic image generation method based on camera response curve |
| CN117784388B (en) * | 2024-02-28 | 2024-05-07 | 宁波永新光学股份有限公司 | High dynamic range metallographic image generation method based on camera response curve |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106846383B (en) | 2020-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106846383A (en) | High dynamic range images imaging method based on 3D digital micro-analysis imaging systems | |
| CN109360235B (en) | Hybrid depth estimation method based on light field data | |
| CN105931190B (en) | High angular resolution light filed acquisition device and image generating method | |
| US9224193B2 (en) | Focus stacking image processing apparatus, imaging system, and image processing system | |
| EP2757789A1 (en) | Image processing system, image processing method, and image processing program | |
| DE102015005267A1 (en) | Information processing apparatus, method therefor and measuring apparatus | |
| CN112019719B (en) | High-resolution light field system and imaging method based on optical framing light field camera | |
| CN107845145B (en) | Three-dimensional reconstruction system and method under electron microscopic scene | |
| KR101600681B1 (en) | Depth convertion method of 3D images interal imaging system | |
| JP7479729B2 (en) | Three-dimensional representation method and device | |
| JPWO2014192487A1 (en) | Multi-view imaging system, acquired image composition processing method, and program | |
| CN106023189A (en) | Light field data depth reconstruction method based on matching optimization | |
| JP6285686B2 (en) | Parallax image generation device | |
| CN115578296A (en) | Stereo video processing method | |
| KR102253320B1 (en) | Method for displaying 3 dimension image in integral imaging microscope system, and integral imaging microscope system implementing the same | |
| CN104104911B (en) | Timestamp in panoramic picture generating process is eliminated and remapping method and system | |
| Deshpande et al. | 3d image generation from single image using color filtered aperture and 2.1 d sketch-a computational 3d imaging system and qualitative analysis | |
| KR101841750B1 (en) | Apparatus and Method for correcting 3D contents by using matching information among images | |
| CN101686407A (en) | Method and device for acquiring sampling point information | |
| CN107103620B (en) | Depth extraction method of multi-optical coding camera based on spatial sampling under independent camera view angle | |
| CN113204107B (en) | Three-dimensional scanning microscope with double objective lenses and three-dimensional scanning method | |
| CN112102347B (en) | Step detection and single-stage step height estimation method based on binocular vision | |
| Yu et al. | Dynamic depth of field on live video streams: A stereo solution | |
| CN109089100B (en) | Method for synthesizing binocular stereo video | |
| CN111985535A (en) | Method and device for optimizing human body depth map through neural network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200417 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |