CN104867117B - A kind of flow field image pre-processing method and its system - Google Patents
A kind of flow field image pre-processing method and its system Download PDFInfo
- Publication number
- CN104867117B CN104867117B CN201510240957.8A CN201510240957A CN104867117B CN 104867117 B CN104867117 B CN 104867117B CN 201510240957 A CN201510240957 A CN 201510240957A CN 104867117 B CN104867117 B CN 104867117B
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- image
- pixel
- flow field
- 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.)
- Active
Links
- 238000007781 pre-processing Methods 0.000 title claims abstract description 53
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000003595 mist Substances 0.000 claims description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 9
- 238000012935 Averaging Methods 0.000 claims description 7
- 238000005286 illumination Methods 0.000 claims description 7
- 239000007787 solid Substances 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 15
- 230000001360 synchronised effect Effects 0.000 abstract description 5
- 238000010276 construction Methods 0.000 abstract 1
- 239000002245 particle Substances 0.000 description 10
- 238000005259 measurement Methods 0.000 description 8
- 230000005540 biological transmission Effects 0.000 description 5
- 230000033001 locomotion Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000000873 masking effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000011410 subtraction method Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
A kind of flow field image pre-processing method and its system, belong to image pre-processing method and device, for pre-processing the problem of Occupation time is longer in current flow field survey, improve flow field image preprocessing efficiency.The flow field image pre-processing method of the present invention includes equalization step, image stretch step, noise judgment step, removal steady noise step, construction binaryzation average image step and mask operation step.The flow field figure of the present invention is as pretreatment system, including input Ethernet interface, Ethernet input control chip, digital signal processor, synchronous DRAM, dynamic memory chip, band EEPROM, output Ethernet interface and Ethernet output control chip.Volume and power consumption of the present invention are smaller, can be embedded in existing flow field survey instrument, carry out flow field survey image preprocessing, and the less flow field survey picture of output noise improves the efficiency and processing speed of whole back-end processor.
Description
Technical Field
The invention belongs to an image preprocessing method and device, and particularly relates to a flow field image preprocessing method and a system thereof, which are used for Ethernet communication.
Background
The research of high-speed flow fields has become more and more widely applied to engineering practices (such as maneuvering flight of an aircraft, complex-shape flow of a space plane, and the like), and there are many questions in the current research, mainly that the environment of the flow fields is special, the movement speed of particles inside the flow fields is high, the particles are not uniformly distributed, the illumination of the whole flow fields is not uniform in the laser environment, and the interference of other external factors (bubbles may appear when the particles move, water mist may appear on observation walls, and the like) makes it difficult to obtain stably moving particle images, the particle movement speed is high, and the acquisition performance of a camera is required to be higher. The current common method is to use a high-speed camera to collect the current field picture and transmit the current field picture to a PC for processing. Because a great deal of noise (uneven illumination, bubbles, water mist and the like) exists in an original particle picture, a certain time is needed for image preprocessing at the very beginning, and because the data volume is very large (the frame rate of a camera used for observing the current flow field is more than 100fps, and the resolution is more than 100w pixels), the time spent on image preprocessing is very high in the field of flow field measurement emphasizing high real-time processing performance, so that the whole processing speed is seriously influenced, and for traditional flow field measurement image preprocessing, the time spent on preprocessing usually occupies 30% -40% of the whole processing time, and the improvement of the processing speed is seriously influenced.
Disclosure of Invention
The invention provides a flow field image preprocessing method and an image preprocessing system thereof, aiming at solving the problem of longer preprocessing occupation time in the current flow field measurement and improving the flow field image preprocessing efficiency.
The invention provides a flow field image preprocessing method, which is characterized in that for a flow field to be observed, N flow field images are obtained by utilizing an Ethernet camera, and the method comprises the following steps of:
(1) equalizing step:
for the k flow field image IkSliding the image window line by line for all the pixels, and calculating the mean value N of each pixel x in the jth image windowk j(x)：
k＝1、…、N，N＝1000～50000，j≥1，
0≤Nk j(x)≤1；
Wherein, Ik j(x) Is the pixel value, I, of the pixel x in the jth window of the kth flow field imagej maxAnd Ij minMaximum pixel value and minimum pixel value in jth image window respectively:
Ω is a domain where an image window is located, and the size of the image window is M × M, where M is 8, 16, or 32 pixels;
(2) an image stretching step comprising the sub-steps of:
(2.1) for the jth image window of the kth image, calculating the stretching value S of each pixel xj k(x)：
WhereinRespectively a minimum pixel value and a maximum pixel value in the kth image;
(2.2) for the k-th image, the stretching value S of each pixel xj k(x) According to the position of its corresponding pixel xInto the k-th stretched image Sk；
An image equalization step and an image stretching step, which are carried out in each window in an image;
in the flow field measurement, because the particles are very small and the flow speed is very high, the particles in an area to be observed are generally illuminated by laser, and because of the problem of an image acquisition angle, the phenomenon of uneven illumination exists when a camera acquires a particle picture, which shows that the brightness of some places is very high and some places are dark, and the phenomenon of uneven illumination is removed by performing image stretching operation, so that the brightness of an image is more uniform, and the later-stage processing is convenient;
(3) a noise judgment step including the substeps of:
(3.1) visually judging the kth flow field image IkIf the water mist exists, performing the step (4); otherwise, performing the substep (3.2);
(3.2) judging the kth flow field image IkIf there is a bubble, if yes, go to step (5), otherwise output SkAnd ending;
(4) a fixed noise removing step, comprising the following substeps:
(4.1) first, a template image B is obtainedref：
The meaning of the above formula is that the pixel values corresponding to N stretched images are added, and the addition result is divided by N to form the template image B according to the position of the corresponding pixel xref；
(4.2) the kth stretched image SkAnd template image BrefCarrying out image subtraction to obtain a de-noised image
The image subtraction means that pixel values corresponding to two images are subtracted, and the subtraction result forms a de-noised image according to the position of a corresponding pixel x;
the image subtraction operation is to remove fixed noise in a flow field, and in a flow field measurement experiment, water mist on a chamber wall is mainly observed, the water mist is static in an image, and particles move, so that the water mist can be removed by a background subtraction method;
(4.3) judging the kth flow field image IkIf there is a bubble, if yes, go to step (5), otherwise outputFinishing;
(5) and constructing a binary mean image, which comprises the following substeps:
(5.1) selecting a flow field image containing no air bubbles as a background image InomalThe k-th flow field image IkWith background image InomalCarrying out image subtraction to obtain the kth difference image
(5.2) carrying out averaging operation on all pixels of the kth difference image to obtain a difference averaging value G of each pixel xk(x)：
0≤Gk(x)≤1；
Wherein,representing the k-th difference imageIs determined by the minimum pixel value of (c),representing the k-th difference imageA maximum pixel value of;
(5.3) averaging value G of the difference value of each pixel xk(x) Forming a k-th mean image G according to the position of the corresponding pixel xk；
(5.4) for GkPerforming binarization operation to form a k-th binarized imageWherein the pixel value of each pixel x
Wherein, the binary threshold G is not less than 0kthLess than or equal to 1; the larger the laser energy irradiated to the flow field is, the larger GkthCloser to 1, otherwise, GkthThe closer to 0;
the pixel value of the picture is not 0 or 1, and because the actual flow field is very dark, bubbles in the bubble place can obviously reflect illumination in the actual laser lighting process, and the brightness is higher in the actual laser lighting processThe middle pixel value is 1;
(6) a mask operating step comprising the substeps of:
(6.1) inAll pixel points with the pixel value of 1 are judged and processed to form a mask image
(6.2) stretching the k-th stretched image SkOr de-noising an imageAnd mask imagePerforming image subtraction to obtain a preprocessed result image, outputting the result image, and ending; the resulting image eliminates image noise such as bubbles, water mist, uneven lighting, etc.
The substep (3.2) and the step (4.3) judge the kth flow field image IkWhether or not the air bubbles exist, comprising the following processes:
A. visually selecting a flow field image without air bubbles as a background image InomalThe k-th flow field image IkWith background image InomalCarrying out image subtraction to obtain the kth difference image
B. For difference imagesSliding the difference image window line by line for all the pixels, and comparing each pixel value contained in the difference image window with the bubble threshold valueIs greater than the difference image windowIf the pixel value exceeds (Q/2+0.5), judging that the difference image window contains bubbles, otherwise, judging that the difference image window does not contain bubbles;
the difference image window size is Q × Q, Q ═ 9, 11, or 13 pixels;
selecting suitable ones according to different lighting conditionsWhen the light is strong, the light source is in a dark state,values close to 255, when the light is weak,values close to 0, in generalThe value is 100;
C. difference imageIf any one of the difference image windows contains bubbles, the flow field image I is considered to bekContains bubbles; only difference imagesAll the difference image windows are notThe bubble is included, and the flow field image I is consideredkNo air bubbles are contained.
Said substep (6.1) forming a mask imageWhen is atSequentially performing the following operations on all pixels with the pixel values of 1 row by row and column by column until all pixels with the pixel values of 1 are traversed:
(6.1.1) with the pixel with the current pixel value of 1 as the center point, judging whether the pixel with the pixel value of 1 exists in the surrounding 3 × 3 pixel area, if so, carrying out the process (6.1.2), otherwise, setting the pixel value of the center point as 0, and moving to the center pointTurning to the pixel with the next pixel value of 1 (6.1.1);
(6.1.2) with the pixel with the current pixel value of 1 as the center point, judging whether a new pixel with the pixel value of 1 exists in the surrounding 5 × 5 pixel area, if so, performing the process (6.1.3), otherwise, setting all the pixel values of all the pixels in the 5 × 5 pixel area to be 0, and moving to the pixel area with the pixel value of 1Turning to the pixel with the next pixel value of 1 (6.1.1);
(6.1.3) with the pixel with the current pixel value of 1 as the center point, judging whether a new pixel with the pixel value of 1 exists in the surrounding 7 × 7 pixel area, if so, performing the process (6.1.4), otherwise, setting all the pixel values of all the pixels in the 7 × 7 pixel area to be 0, and moving to the 7 × pixel areaTurning to the pixel with the next pixel value of 1 (6.1.1);
(6.1.4) with the pixel with the current pixel value of 1 as the center point, judging whether a new pixel with the pixel value of 1 exists in the surrounding 9 multiplied by 9 pixel area, if so, performing the process (6.1.6), otherwise, performing the process (6.1.5);
(6.1.5) making a circle by taking the center point as the center of all pixels with the pixel values of 1 in the 7 x 7 area and taking the distance from the pixel with the pixel value of 1 farthest from the center point in the 5 x 5 pixel area around the pixel to the center point as the radius, setting the pixel values of all pixels on the circumference as 1, forming a closed area, and turning to the process (6.1.1);
(6.1.6) taking the pixel with the current pixel value of 1 as a center point, judging whether a new pixel with the pixel value of 1 exists in a surrounding 11 × pixel area, if so, judging whether a new pixel with the pixel value of 1 exists in a surrounding 13 × pixel area, otherwise, taking the center point as a center point of all pixels with the pixel value of 1 in a 9 × area, taking the distance from the pixel with the farthest distance from the center point to the center point as a radius to make a circle, taking the pixel with the pixel value of 1 in a surrounding 7 × pixel area as a radius, setting the pixel values of all pixels on the circumference as 1 to form a closed area, repeating the step … until the pixel with the current pixel value of 1 is taken as the center point, and the pixel with the farthest distance from the center point to the center point of (2P +1) × (2P +1) pixel area (2P +1) no longer has the pixel value of 1, setting the distance from the center point of all pixels with the pixel value of 2P-1) in the surrounding × (2P-1) area as a center point, and taking the distance from the center point to the center point of the center point as a circle, setting the distance from the center point to the center point as a circle, and taking the radius of all pixels (3-3) as a circle, wherein the distance from the center point to the center point is 3.78.78 to form a circle, and the process of all pixelsA positive integer of half the size;
(6.1.7) inAfter the above operation is completed for all the pixels having the pixel value of 1, a mask intermediate image is formedIn thatIn which the pixel values of all pixels within each circle are set to 1 to form a mask imageA plurality of "solid circles" having a pixel value of 1; the position and size of the bubble in the image are determined in their entirety.
The invention provides a flow field image preprocessing system, which comprises an input Ethernet port, an Ethernet input control chip, a digital signal processor, a synchronous dynamic random access memory, a dynamic storage chip, a charged erasable programmable read-only memory, an output Ethernet port and an Ethernet output control chip, and is characterized in that:
the Ethernet input control chip controls an input Ethernet port to communicate with an Ethernet camera, acquires image data of a flow field to be measured, and sends the image data to a synchronous dynamic random access memory for temporary storage through a digital signal processor, the dynamic storage chip is used for storing an image preprocessing algorithm, the electrified erasable programmable read only memory is used for storing configuration parameters, and the configuration parameters comprise a flow field image amplitude N and an image window size parameter M; when the device works, the Ethernet camera collects original image data of a high-speed motion flow field, the original image data are transmitted to an input Ethernet port through a gigabit Ethernet, the digital signal processor reads parameters in the EEPROM, then an image preprocessing algorithm stored by the FLASH is loaded to the digital signal processor, then the image data are read from the SDRAM to be subjected to image preprocessing, and the result of the image preprocessing is transmitted to the PC through a network cable from an output Ethernet port by the Ethernet output control chip.
The flow field image preprocessing system is further characterized in that:
the digital signal processor is provided with a joint test port, so that a user can conveniently develop an image preprocessing algorithm for the second time in the digital signal processor.
The invention mainly aims at the requirement of pretreatment in high-speed flow field measurement. The image preprocessing system can be powered by a small 12V direct current switching power supply, the size is not more than 13cm multiplied by 10cm, the weight is not more than 400g (without a power supply), the size is small, the weight is light, and the carrying is convenient; the gigabit Ethernet is used for communication, the data transmission speed can reach very high, the gigabit Ethernet can be matched with an Ethernet camera with very high pixel and frame rate to work cooperatively, and the gigabit Ethernet is more suitable for flow field measurement occasions with large data transmission quantity, the selected DSP is an 8-core DSP processor, the highest processing speed of each core is 1.25GHz, 8-core parallel processing is utilized, the flow field image preprocessing function with large data quantity can be realized under the platform, and the gigabit Ethernet has better processing effects on the phenomena of uneven illumination, bubbles, water mist and the like in flow field shooting; the selected FLASH has the capacity of 32MB, can store a plurality of image preprocessing algorithms, can realize the selection of different image preprocessing algorithms according to the actual situation of a flow field by matching with the EEPROM, and greatly improves the adaptability; the processing result is transmitted by using a gigabit Ethernet, the transmission speed is high, the transmission data volume is large, not only the preprocessed flow field image but also the original flow field image can be transmitted, and the long-distance transmission can be carried out, so that the problem that the image preprocessing in the conventional PIV instrument takes longer time can be mainly solved, and the image preprocessing speed is improved.
The flow field image preprocessing system and the Ethernet camera are matched for use, the flow field image acquisition and image preprocessing are placed on the flow field image preprocessing system for processing, finally, a flow field measurement image with low noise is output, the efficiency and the processing speed of a whole back-end processor are improved, meanwhile, the flow field image preprocessing system is a micro embedded processing system, the size and the power consumption are small, the flow field image preprocessing system can be well embedded into an existing flow field measuring instrument, and the image preprocessing function of the existing flow field measuring instrument is improved.
Drawings
FIG. 1 is a flow chart of a method for preprocessing a flow field image according to the present invention;
FIG. 2 is a schematic view of forming a mask imageA flow diagram of (a);
fig. 3 is a schematic structural diagram of the flow field image preprocessing system of the present invention.
Detailed Description
The invention is further illustrated below with reference to the figures and examples.
As shown in fig. 1, the method for preprocessing a flow field image of the present invention includes: (1) the method comprises the following steps of (1) an averaging step, (2) an image stretching step, (3) a noise judging step, (4) a fixed noise removing step, (5) a binarization mean image constructing step and (6) a mask operating step.
As an example, in the equalization step (1), N is 3000, and M is 8 pixels; in the step (5) of constructing a binary mean image, a binary threshold value Gkth＝0.4。
The substep (3.2) of the noise judging step and the substep (4.3) of the fixed noise removing step judge the kth flow field image IkIn the present embodiment, the window size of the difference image is Q × Q, Q is 11 pixels, and the bubble threshold is set as the bubble threshold
FIG. 2 illustrates the formation of a mask image during a masking operation stepA flow diagram of (a);
as shown in fig. 3, the system for preprocessing streaming image of the present invention includes an input ethernet port, an ethernet input control chip, a digital signal processor DSP, a synchronous dynamic random access memory SDRAM, a dynamic storage chip FLASH, a charged erasable programmable read only memory EEPROM, an output ethernet port, and an ethernet output control chip;
the Ethernet input control chip controls an input Ethernet port to communicate with an Ethernet camera, acquires image data of a flow field to be measured, and sends the image data to a synchronous dynamic random access memory SDRAM for temporary storage through a digital signal processor DSP, a dynamic storage chip FLASH is used for storing an image preprocessing algorithm, an electrically erasable programmable read-only memory EEPROM is used for storing configuration parameters, and the configuration parameters comprise a flow field image amplitude N and an image window size parameter M; when the device works, the Ethernet camera collects original image data of a high-speed motion flow field, the original image data are transmitted to an input Ethernet port through a gigabit Ethernet, the digital signal processor DSP reads parameters in the EEPROM, then an image preprocessing algorithm stored in the FLASH is loaded to the digital signal processor DSP, then the image data are read from the SDRAM to be subjected to image preprocessing, and the result of the image preprocessing is transmitted to the PC through a network cable from an output Ethernet port by the Ethernet output control chip.
The DSP can be configured with a joint test port JTAG, which is convenient for users to carry out secondary development on the image preprocessing algorithm in the DSP.
The developed image preprocessing algorithm can be burnt into FLASH through a JTAG port, then corresponding parameter adjustment is carried out in EEPROM, and the image preprocessing algorithm developed by the user is selected for image processing when the system is started next time.
Because the DSP processor has high requirements on the power supply, the power supply module not only ensures that the ripple of the voltage is small, but also ensures that the power-on time sequence of the DSP processor is strictly carried out according to a manual.
In this embodiment, the selected DSP processor is TMS320C6678, FLASH is NAND512R3A2D, EEPROM is M24M01, JTAG is a general purpose 10P JTAG interface, the ethernet port is a general purpose gigabit ethernet port, the ethernet control chip is 88E1111, and SDRAM is K4B2G 1646C.
Claims (3)
1. A flow field image preprocessing method, for a flow field to be observed, using an Ethernet camera to obtain N flow field images, is characterized in that, for any k-th flow field image, the method comprises the following steps:
(1) equalizing step:
for the k flow field image IkSliding the image window line by line for all the pixels, and calculating the mean value N of each pixel x in the jth image windowk j(x)：
<mfenced open = "" close = ""> <mtable> <mtr> <mtd> <mrow> <msup> <msub> <mi>N</mi> <mi>k</mi> </msub> <mi>j</mi> </msup> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msup> <msub> <mi>I</mi> <mi>k</mi> </msub> <mi>j</mi> </msup> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>-</mo> <msub> <msup> <mi>I</mi> <mi>j</mi> </msup> <mi>min</mi> </msub> </mrow> <mrow> <msub> <msup> <mi>I</mi> <mi>j</mi> </msup> <mi>max</mi> </msub> <mo>-</mo> <msub> <msup> <mi>I</mi> <mi>j</mi> </msup> <mi>min</mi> </msub> </mrow> </mfrac> <mo>,</mo> <mi>k</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mn>...</mn> <mo>,</mo> <mi>N</mi> <mo>,</mo> <mi>N</mi> <mo>=</mo> <mn>1000</mn> <mo>~</mo> <mn>50000</mn> <mo>,</mo> <mi>j</mi> <mo>&GreaterEqual;</mo> <mn>1</mn> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mn>0</mn> <mo>&le;</mo> <msup> <msub> <mi>N</mi> <mi>k</mi> </msub> <mi>j</mi> </msup> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>&le;</mo> <mn>1</mn> <mo>;</mo> </mrow> </mtd> </mtr> </mtable> </mfenced>
Wherein, Ik j(x) Is the pixel value, I, of the pixel x in the jth window of the kth flow field imagej maxAnd Ij minMaximum pixel value and minimum pixel value in jth image window respectively:
<mrow> <msub> <msup> <mi>I</mi> <mi>j</mi> </msup> <mi>max</mi> </msub> <mo>=</mo> <munder> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> <mrow> <mi>x</mi> <mo>&Element;</mo> <mi>&Omega;</mi> </mrow> </munder> <mrow> <mo>(</mo> <msup> <msub> <mi>I</mi> <mi>k</mi> </msub> <mi>j</mi> </msup> <mo>(</mo> <mi>x</mi> <mo>)</mo> <mo>)</mo> </mrow> <mo>,</mo> <msub> <msup> <mi>I</mi> <mi>j</mi> </msup> <mi>min</mi> </msub> <mo>=</mo> <munder> <mrow> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mi>x</mi> <mo>&Element;</mo> <mi>&Omega;</mi> </mrow> </munder> <mrow> <mo>(</mo> <msub> <msup> <mi>I</mi> <mi>j</mi> </msup> <mi>k</mi> </msub> <mo>(</mo> <mi>x</mi> <mo>)</mo> <mo>)</mo> </mrow> <mo>,</mo> </mrow>
Ω is a domain where an image window is located, and the size of the image window is M × M, where M is 8, 16, or 32 pixels;
(2) an image stretching step comprising the sub-steps of:
(2.1) for the jth image window of the kth image, calculating the stretching value S of each pixel xj k(x)：
<mrow> <msub> <msup> <mi>S</mi> <mi>j</mi> </msup> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <msup> <mi>N</mi> <mi>j</mi> </msup> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>-</mo> <mover> <msub> <mi>I</mi> <mrow> <mi>k</mi> <mi>min</mi> </mrow> </msub> <mo>&OverBar;</mo> </mover> </mrow> <mrow> <mover> <msub> <mi>I</mi> <mrow> <mi>k</mi> <mi>max</mi> </mrow> </msub> <mo>&OverBar;</mo> </mover> <mo>-</mo> <mover> <msub> <mi>I</mi> <mrow> <mi>k</mi> <mi>min</mi> </mrow> </msub> <mo>&OverBar;</mo> </mover> </mrow> </mfrac> <mo>,</mo> </mrow>
WhereinRespectively a minimum pixel value and a maximum pixel value in the kth image;
(2.2) for the k-th image, the stretching value S of each pixel xj k(x) Forming a k-th stretched image S according to the position of the corresponding pixel xk；
An image equalization step and an image stretching step, which are carried out in each window in an image;
(3) a noise judgment step including the substeps of:
(3.1) visually judging the kth flow field image IkIf the water mist exists, performing the step (4); otherwise, performing the substep (3.2);
(3.2) judging the kth flow field image IkIf there is a bubble, if yes, go to step (5), otherwise output SkAnd ending;
(4) a fixed noise removing step, comprising the following substeps:
(4.1) first, a template image B is obtainedref：
<mrow> <msub> <mi>B</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <mi>N</mi> </mfrac> <munderover> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>S</mi> <mi>k</mi> </msub> <mo>;</mo> </mrow>
The meaning of the above formula is that the pixel values corresponding to N stretched images are added, and the addition result is divided by N to form the template image B according to the position of the corresponding pixel xref；
(4.2) the kth stretched image SkAnd template image BrefCarrying out image subtraction to obtain a de-noised image
<mrow> <msub> <mover> <mi>S</mi> <mo>&OverBar;</mo> </mover> <mi>k</mi> </msub> <mo>=</mo> <msub> <mi>S</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>B</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msub> <mo>;</mo> </mrow>
The image subtraction means that pixel values corresponding to two images are subtracted, and the subtraction result forms a de-noised image according to the position of a corresponding pixel x;
(4.3) judging the kth flow field image IkIf there is a bubble, if yes, go to step (5), otherwise outputFinishing;
(5) and constructing a binary mean image, which comprises the following substeps:
(5.1) selecting a flow field image containing no air bubbles as a background image InomalThe k-th flow field image IkWith background image InomalCarrying out image subtraction to obtain the kth difference image
<mrow> <mover> <msub> <mi>I</mi> <mi>k</mi> </msub> <mo>-</mo> </mover> <mo>=</mo> <msub> <mi>I</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>I</mi> <mrow> <mi>n</mi> <mi>o</mi> <mi>m</mi> <mi>a</mi> <mi>l</mi> </mrow> </msub> <mo>,</mo> </mrow>
(5.2) carrying out averaging operation on all pixels of the kth difference image to obtain a difference averaging value G of each pixel xk(x)：
<mrow> <msub> <mi>G</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <mover> <msub> <mi>I</mi> <mi>k</mi> </msub> <mo>-</mo> </mover> <mo>-</mo> <mover> <msub> <mi>I</mi> <mrow> <mi>k</mi> <mi>min</mi> </mrow> </msub> <mo>-</mo> </mover> </mrow> <mrow> <msub> <mover> <mi>I</mi> <mo>&OverBar;</mo> </mover> <mrow> <mi>k</mi> <mi>max</mi> </mrow> </msub> <mo>-</mo> <mover> <msub> <mi>I</mi> <mrow> <mi>k</mi> <mi>min</mi> </mrow> </msub> <mo>-</mo> </mover> </mrow> </mfrac> <mo>,</mo> <mn>0</mn> <mo>&le;</mo> <msub> <mi>G</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>&le;</mo> <mn>1</mn> <mo>;</mo> </mrow>
Wherein,representing the k-th difference imageIs determined by the minimum pixel value of (c),representing the k-th difference imageA maximum pixel value of;
(5.3) averaging value G of the difference value of each pixel xk(x) Forming a k-th mean image G according to the position of the corresponding pixel xk；
(5.4) for GkPerforming binarization operation to form a k-th binarized imageWherein the pixel value of each pixel x
<mrow> <msub> <mover> <mi>G</mi> <mo>&OverBar;</mo> </mover> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mn>0</mn> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mo>(</mo> <msub> <mi>G</mi> <mi>k</mi> </msub> <mo>(</mo> <mi>x</mi> <mo>)</mo> <mo>&le;</mo> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mi>t</mi> <mi>h</mi> </mrow> </msub> <mo>)</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mn>1</mn> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mo>(</mo> <msub> <mi>G</mi> <mi>k</mi> </msub> <mo>(</mo> <mi>x</mi> <mo>)</mo> <mo>></mo> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mi>t</mi> <mi>h</mi> </mrow> </msub> <mo>)</mo> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>;</mo> </mrow>
Wherein, the binary threshold G is not less than 0kthLess than or equal to 1; the larger the laser energy irradiated to the flow field is, the larger GkthCloser to 1, otherwise, GkthThe closer to 0;
(6) a mask operating step comprising the substeps of:
(6.1) inAll pixel points with the pixel value of 1 are judged and processed to form a mask image
(6.2) stretching the k-th stretched image SkOr de-noising an imageAnd mask imageAnd (5) carrying out image subtraction to obtain a result image after preprocessing, outputting the result image, and ending.
2. The flow field image preprocessing method of claim 1, characterized by:
the substep (3.2) and the substep (4.3) judge the kth flow field image IkWhether or not the air bubbles exist, comprising the following processes:
A. visually selecting a flow field image without air bubbles as a background image InomalThe k-th flow field image IkWith background image InomalCarrying out image subtraction to obtain the kth difference image
<mrow> <mover> <msub> <mi>I</mi> <mi>k</mi> </msub> <mo>-</mo> </mover> <mo>=</mo> <msub> <mi>I</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>I</mi> <mrow> <mi>n</mi> <mi>o</mi> <mi>m</mi> <mi>a</mi> <mi>l</mi> </mrow> </msub> <mo>,</mo> </mrow>
B. For difference imagesSliding the difference image window line by line for all the pixels, and comparing each pixel value contained in the difference image window with the bubble threshold valueIs greater than the difference image windowIf the pixel value exceeds (Q/2+0.5), judging that the difference image window contains bubbles, otherwise, judging that the difference image window does not contain bubbles;
the difference image window size is Q × Q, Q ═ 9, 11, or 13 pixels;
selecting suitable ones according to different lighting conditionsWhen the light is strong, the light source is in a dark state,values close to 255, when the light is weak,the value is close to 0, when the illumination is in a common state,the value is 100;
C. difference imageIf any one of the difference image windows contains bubbles, the flow field image I is considered to bekContains bubbles; only difference imagesAll the difference image windows do not contain bubbles, and the flow field image I is consideredkNo air bubbles are contained.
3. The flow field image preprocessing method of claim 1, characterized by:
said substep (6.1) forming a mask imageWhen is atSequentially performing the following operations on all pixels with the pixel values of 1 row by row and column by column until all pixels with the pixel values of 1 are traversed:
(6.1.1) with the pixel with the current pixel value of 1 as the center point, judging whether the pixel with the pixel value of 1 exists in the surrounding 3 × 3 pixel area, if so, carrying out the process (6.1.2), otherwise, setting the pixel value of the center point as 0, and moving to the center pointTurning to the pixel with the next pixel value of 1 (6.1.1);
(6.1.2) with the pixel whose current pixel value is 1 as the center point, judging the surrounding 5 × 5 imageIf there is a new pixel with a pixel value of 1 in the pixel area, the process is performed (6.1.3), otherwise, all the pixel values of all the pixels in the 5 × 5 pixel area are set to 0, and the process moves toTurning to the pixel with the next pixel value of 1 (6.1.1);
(6.1.3) with the pixel with the current pixel value of 1 as the center point, judging whether a new pixel with the pixel value of 1 exists in the surrounding 7 × 7 pixel area, if so, performing the process (6.1.4), otherwise, setting all the pixel values of all the pixels in the 7 × 7 pixel area to be 0, and moving to the 7 × pixel areaTurning to the pixel with the next pixel value of 1 (6.1.1);
(6.1.4) with the pixel with the current pixel value of 1 as the center point, judging whether a new pixel with the pixel value of 1 exists in the surrounding 9 multiplied by 9 pixel area, if so, performing the process (6.1.6), otherwise, performing the process (6.1.5);
(6.1.5) making a circle by taking the center point as the center of all pixels with the pixel values of 1 in the 7 x 7 area and taking the distance from the pixel with the pixel value of 1 farthest from the center point in the 5 x 5 pixel area around the pixel to the center point as the radius, setting the pixel values of all pixels on the circumference as 1, forming a closed area, and turning to the process (6.1.1);
(6.1.6) with the pixel with the current pixel value of 1 as the center point, judging whether a new pixel with the pixel value of 1 exists in the surrounding 11 × pixel area, if so, judging whether a new pixel with the pixel value of 1 exists in the surrounding 13 × 13 pixel area, otherwise, making a circle with the center point as the center point and the distance from the pixel with the pixel value of 1 farthest from the center point in the surrounding 7 × pixel area as the radius, setting the pixel values of all the pixels on the circumference as 1 to form a closed area, and repeating … until no new pixel with the current pixel value of 1 exists in the surrounding (2P +1) × (2P +1) pixel area, and then, setting the pixel with the pixel value of 1 as the center point, and setting the pixel values of 1 as the pixel of the surrounding (2P +1) × (2P +1) pixel areaAll pixels with pixel values of 1 in the (2P-1) × (2P-1) area are rounded by taking the central point as the center of a circle and taking the distance from the pixel with the pixel value of 1 farthest from the central point in the (2P-3) × (2P-3) pixel area around the pixel to the central point as the radius, the pixel values of all pixels on the circumference are set to be 1, a closed area is formed, and the process is repeated (6.1.1), wherein P is less than or equal to PA positive integer of half the size;
(6.1.7) inAfter the above operation is completed for all the pixels having the pixel value of 1, a mask intermediate image is formedIn thatIn which the pixel values of all pixels within each circle are set to 1 to form a mask imageA plurality of "solid circles" having a pixel value of 1; the position and size of the bubble in the image are determined in their entirety.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510240957.8A CN104867117B (en) | 2015-05-13 | 2015-05-13 | A kind of flow field image pre-processing method and its system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510240957.8A CN104867117B (en) | 2015-05-13 | 2015-05-13 | A kind of flow field image pre-processing method and its system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104867117A CN104867117A (en) | 2015-08-26 |
| CN104867117B true CN104867117B (en) | 2017-10-27 |
Family
ID=53912934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510240957.8A Active CN104867117B (en) | 2015-05-13 | 2015-05-13 | A kind of flow field image pre-processing method and its system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104867117B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778722A (en) * | 2016-11-25 | 2017-05-31 | 天方创新（北京）信息技术有限公司 | Image processing method and device |
| CN107747549B (en) * | 2017-11-27 | 2019-05-31 | 江苏大学 | A kind of centrifugal pump interior flow field observation experiment platform |
| CN109919853B (en) * | 2019-01-07 | 2023-04-25 | 南京航空航天大学 | PIV image preprocessing method based on particle feature retrieval |
| CN112816187B (en) * | 2021-01-06 | 2022-12-13 | 北京工业大学 | Quality judgment method for laser spots |
| CN113610740A (en) * | 2021-08-13 | 2021-11-05 | 江苏富恩日化科技有限公司 | Bubble feature enhancement method and system in air tightness detection based on artificial intelligence |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102055894A (en) * | 2010-11-23 | 2011-05-11 | 无锡市博阳精密机械制造有限公司 | Modularized CCD industrial camera |
| CN102980527A (en) * | 2012-10-18 | 2013-03-20 | 北京航空航天大学 | Three-dimensional sensor true-time high speed data processing system based on complementary metal-oxide-semiconductor (CMOS) |
| CN103793924A (en) * | 2014-01-28 | 2014-05-14 | 河海大学 | Flow field image self-adaption motion vector estimating method based on FHT-CC |
| CN104156920A (en) * | 2014-08-07 | 2014-11-19 | 江苏大学 | Method for processing impeller flow field images of centrifugal pump |
| CN104284065A (en) * | 2014-09-30 | 2015-01-14 | 华中科技大学 | Industrial intelligent camera based on FPGA |
-
2015
- 2015-05-13 CN CN201510240957.8A patent/CN104867117B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102055894A (en) * | 2010-11-23 | 2011-05-11 | 无锡市博阳精密机械制造有限公司 | Modularized CCD industrial camera |
| CN102980527A (en) * | 2012-10-18 | 2013-03-20 | 北京航空航天大学 | Three-dimensional sensor true-time high speed data processing system based on complementary metal-oxide-semiconductor (CMOS) |
| CN103793924A (en) * | 2014-01-28 | 2014-05-14 | 河海大学 | Flow field image self-adaption motion vector estimating method based on FHT-CC |
| CN104156920A (en) * | 2014-08-07 | 2014-11-19 | 江苏大学 | Method for processing impeller flow field images of centrifugal pump |
| CN104284065A (en) * | 2014-09-30 | 2015-01-14 | 华中科技大学 | Industrial intelligent camera based on FPGA |
Non-Patent Citations (1)
| Title |
|---|
| 基于低密度粒子图像叠加的Micro-PIV速度场测量;徐明 等;《实验流体力学》;20130430;第27卷(第2期);106-112 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104867117A (en) | 2015-08-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104867117B (en) | A kind of flow field image pre-processing method and its system | |
| CN107393542B (en) | Bird species identification method based on two-channel neural network | |
| Shen et al. | Denoising gravitational waves with enhanced deep recurrent denoising auto-encoders | |
| CN102054270B (en) | Method and device for extracting foreground from video image | |
| CN106127702A (en) | A kind of image mist elimination algorithm based on degree of depth study | |
| CN103258332B (en) | A kind of detection method of the moving target of resisting illumination variation | |
| CN106846261A (en) | Underwater picture processing method based on convolutional neural networks | |
| CN105678248A (en) | Face key point alignment algorithm based on deep learning | |
| CN105551010A (en) | Multi-focus image fusion method based on NSCT (Non-Subsampled Contourlet Transform) and depth information incentive PCNN (Pulse Coupled Neural Network) | |
| CN106803063A (en) | A kind of metric learning method that pedestrian recognizes again | |
| CN103400368B (en) | Based on graph theory and the parallel rapid SAR image segmentation method of super-pixel | |
| CN103761717B (en) | A kind of urban water-body extracting method based on panchromatic remote sensing image | |
| CN106295645B (en) | A kind of license plate character recognition method and device | |
| CN111652822B (en) | Single image shadow removing method and system based on generation countermeasure network | |
| CN104599290A (en) | Video sensing node-oriented target detection method | |
| CN104992452A (en) | Flight object automatic tracking method based on thermal imaging video | |
| CN114972323B (en) | Defect picture generation method and system for model training | |
| CN104217406A (en) | SAR image noise reduction method based on shear wave coefficient processing | |
| CN106447640A (en) | Multi-focus image fusion method based on dictionary learning and rotating guided filtering and multi-focus image fusion device thereof | |
| CN111860175A (en) | Unmanned aerial vehicle image vehicle detection method and device based on lightweight network | |
| CN114529484A (en) | Deep learning sample enhancement method for direct current component change in imaging | |
| CN116645595A (en) | Remote sensing image building roof contour recognition method, device, equipment and medium | |
| CN116385758A (en) | Detection method for damage to surface of conveyor belt based on YOLOv5 network | |
| Di Frischia et al. | Enhanced data augmentation using gans for Raman spectra classification | |
| CN105205834A (en) | Target detection and extraction method based on Gaussian mixture and shade detection model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |