CN107144703B - Built-in image collection and processing system and method based on particle image velocimetry - Google Patents
Built-in image collection and processing system and method based on particle image velocimetry Download PDFInfo
- Publication number
- CN107144703B CN107144703B CN201710319285.9A CN201710319285A CN107144703B CN 107144703 B CN107144703 B CN 107144703B CN 201710319285 A CN201710319285 A CN 201710319285A CN 107144703 B CN107144703 B CN 107144703B
- Authority
- CN
- China
- Prior art keywords
- image
- particle
- built
- time
- window
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
- G01P5/00—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft
- G01P5/18—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft by measuring the time taken to traverse a fixed distance
- G01P5/22—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft by measuring the time taken to traverse a fixed distance using auto-correlation or cross-correlation detection means
Abstract
The invention discloses a kind of built-in image collections based on particle image velocimetry and processing system and method, the system includes one-shot camera and built-in image collection processing module, two CCD camera different-time exposures in one-shot camera acquire particle picture, and particle picture is transferred in image acquisition and processing board；Built-in image collection processing module includes image preprocessing submodule and image acquisition time prediction submodule, wherein image preprocessing submodule completes image preprocessing, image acquisition time predicts that submodule is completed the cross correlation algorithm that particle flow field is tested the speed based on the improved cross correlation algorithm of FFT and calculated, and velocity field predict then to calculate image acquisition interval, and time interval is sent to the control for realizing the exposure time interval of described two CCD cameras in one-shot camera.The present invention can be improved the accuracy of cross-correlation peak value, improve computational accuracy, and performance is high, easy to carry, and the particle flow field of various occasions is suitble to test the speed.
Description
Technical field
The invention belongs to Image Acquisition and processing system field, more particularly, to a kind of based on particle image velocimetry
Built-in image collection and processing system and method.
Background technique
In the research of High Speed Flow Field, particle image velocimetry as a kind of means of testing there is high-precision, transient state, nothing to connect
The features such as touching, measurement of full field, it is widely used the measurement in High Speed Flow Field.General particle image speed-measuring system, is given birth to by particle
It is formed at projection system, laser illumination system, image capturing system and particle picture processing system, in particle picture processing system
In system, PC is mostly used to carry out image processing algorithm.Since continually changing characteristic, PC are locating the speed in ultrahigh speed flow field at any time
It manages on the image of flow field, speed is not fast enough, also cannot achieve processing in real time and control, and high performance PC machine carries not side
Just, a variety of detection occasions can not be suitable for.
At present in the image procossing of particle image velocimetry, optical flow method and cross correlation algorithm are mostly used.Cross correlation algorithm base
Present principles are the cross-correlation letters by two tiny image regions (query window) for calculating the corresponding position of adjacent image pair
Number, by the analysis to cross correlation results, obtains the average displacement of each particle in image querying window, because of adjacent two width picture
Time interval it is known that the actual speed vector of particle can be found out, using FFT, (fast Fourier becomes the cross correlation algorithm
Change), due to selected template when matching, it may cause the not high enough problem of precision；In addition, in current particle image velocimetry
The used camera exposure time is fixed on image capturing system, and is directed to the measurement of High Speed Flow Field, needs frame per second very high
Camera, higher cost, for the flow field of speed real-time change, the fixed time for exposure will affect the accuracy of flow field survey.
Based on drawbacks described above and deficiency, this field is needed to existing particle image velocimetry processing system and image processing method
Method makes further improvement, with obtain easy to detect, detection accuracy is high, accuracy is high particle image velocimetry processing system and
Processing method.
Summary of the invention
Aiming at the above defects or improvement requirements of the prior art, the present invention provides a kind of Image Acquisition and processing system and
Method accordingly devises the built-in image collection based on particle image velocimetry and place wherein particle picture is combined to acquire feature
System and method is managed, through modified hydrothermal process in conjunction with hardware platform, the accuracy of cross-correlation peak value detection is improved, improves
The precision of flow field velocity measurement, and the exposure of camera will be controlled by image acquisition interval, it efficiently solve at a high speed
Unsteady Flow, the problem of tachometric survey inaccuracy.
To achieve the above object, according to one aspect of the present invention, a kind of insertion based on particle image velocimetry is proposed
Formula Image Acquisition and processing system, which is characterized in that including one-shot camera and built-in image collection processing module, in which:
The one-shot camera includes the camera lens set gradually, optical filter, half reflection prism and two CCD cameras, the light splitting phase
Machine adjusts optical path by the half reflection prism, and light beam is divided into two-way and is used for the CCD camera, the CCD camera point
When exposure acquisition particle picture, and the particle picture of acquisition is sent in the built-in image collection processing module；
The built-in image collection processing module includes that image preprocessing submodule and image acquisition time predict submodule,
Wherein, described image pretreatment submodule is used to complete the pretreatment of the particle picture, and by pretreated particle picture
It is sent in described image acquisition time prediction submodule；Described image acquisition time prediction submodule is for carrying out based on FFT
Improved cross correlation algorithm calculate the velocity field at the pretreated particle picture current time, and to the speed of subsequent time
Degree field is predicted, then calculates the image acquisition interval of predetermined speed off field, and time interval signal is sent to
The control of described two CCD camera exposure time intervals is realized in the one-shot camera.
As it is further preferred that the improved cross correlation algorithm based on FFT calculates the pretreated particle
The velocity field at image current time specifically: the query window for the particle picture that current time and subsequent time are acquired respectively into
Row Fast Fourier Transform (FFT) obtains the frequency-domain function of two query windows, and calculate current time query window frequency-domain function with
The product of the conjugation of subsequent time query window frequency-domain function；2 times of up-samplings are done to the product by the way of zero padding matrix,
Fourier inversion is carried out again, obtains cross-correlation peak value position (x0,y0), and obtain window peak positionAgain with this
The picture in 1.5 × 1.5 neighborhoods in the particle picture of current time and subsequent time acquisition is obtained centered on window peak position
Plain window M；Fourier transformation is asked to the pixel window M in current time and subsequent time two images respectively, obtains two windows
The frequency-domain function of mouth M, and calculate the conjugation of the frequency-domain function of current time window M and the frequency-domain function of subsequent time window M
Product；K times is to the product by the way of zero padding to up-sample, then carries out Fourier inversion, obtains cross-correlation peak value position
Set (x1,y1)；According to (the x0,y0) and (x1,y1) acquire fine positioning position:According to described
Acquire current time velocity field in fine positioning position.
As it is further preferred that described image pretreatment submodule is specially fpga chip, described image acquisition time
Predict that submodule is specially dsp chip；Image is transmitted to embedded image by Cameralink interface and adopted by the CCD camera
Collect in processing module, realizes the transmission rate of highest 5.4Gbps；The built-in image collection processing module passes through GPIO interface
Realize the control to one-shot camera.
As it is further preferred that the fpga chip and dsp chip realize high-speed communication by SRIO interface, specifically
, realize the transmission rate of 4 road 5Gbps of highest.
As it is further preferred that utilizing Ethernet protocol by the calculating of image acquisition and processing module by Ethernet interface
As a result it is sent to PC；The pretreatment is gaussian filtering or median filtering.
It is another aspect of this invention to provide that providing a kind of built-in image collection based on particle image velocimetry and processing
Method, which comprises the steps of:
(1) particle picture is acquired using two CCD camera different-time exposures in one-shot camera, and by the particle picture of acquisition
It is sent in built-in image collection processing module；
(2) pretreatment of the particle picture is completed using image preprocessing submodule, and by pretreated particle figure
As being sent to image acquisition time prediction submodule, image acquisition time prediction submodule is for carrying out the improvement based on FFT
Cross correlation algorithm calculate the velocity field at the pretreated particle picture current time, and to the velocity field of subsequent time into
Row prediction calculates, and then calculates the image acquisition interval of predetermined speed off field, and time interval signal is sent to institute
State the control that the exposure time interval of described two CCD cameras is realized in one-shot camera.
As it is further preferred that the improved cross correlation algorithm based on FFT calculates the pretreated particle
The velocity field at image current time specifically: the query window for the particle picture that current time and subsequent time are acquired respectively into
Row Fast Fourier Transform (FFT) obtains the frequency-domain function of two query windows, and calculate current time query window frequency-domain function with
The product of the conjugation of subsequent time query window frequency-domain function；2 times of up-samplings are done to the product by the way of zero padding matrix,
Fourier inversion is carried out again, obtains cross-correlation peak value position (x0,y0), and obtain window peak positionAgain with this
The picture in 1.5 × 1.5 neighborhoods in the particle picture of current time and subsequent time acquisition is obtained centered on window peak position
Plain window M；Fourier transformation is asked to the pixel window M in current time and subsequent time two images respectively, obtains two windows
The frequency-domain function of mouth M, and calculate the conjugation of the frequency-domain function of current time window M and the frequency-domain function of subsequent time window M
Product；K times is to the product by the way of zero padding to up-sample, then carries out Fourier inversion, obtains cross-correlation peak value
(x1,y1)；According to (the x0,y0) and (x1,y1) acquire fine positioning position:According to the essence
Acquire current time velocity field in position location.
As it is further preferred that described image pretreatment submodule is specially fpga chip, described image acquisition time
Predict that submodule is specially dsp chip；The one-shot camera includes the camera lens set gradually, optical filter, half reflection prism and two
A CCD camera, the one-shot camera adjust optical path by the half reflection prism, and light beam is divided into two-way for the CCD camera
It uses.
As it is further preferred that image is transmitted to embedded image by Cameralink interface by the CCD camera
In acquisition processing module, the transmission rate of highest 5.4Gbps is realized；The built-in image collection processing module is connect by GPIO
The existing control to one-shot camera of cause for gossip；The fpga chip and dsp chip pass through SRIO interface and realize high-speed communication, specifically,
Realize the transmission rate of 4 road 5Gbps of highest.
As it is further preferred that utilizing Ethernet protocol by the calculating of image acquisition and processing module by Ethernet interface
As a result it is sent to PC；The pretreatment is gaussian filtering or median filtering.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, mainly have below
Technological merit:
1. cross correlation algorithm of the invention is to improve cross-correlation peak value detection based on the improved cross correlation algorithm of FFT
Accuracy improves the precision of flow field velocity measurement, accelerates the algorithm process time, and be directed to the non-of particle image velocimetry
The interval time of calculated acquisition two images is fed back to camera, camera exposure is controlled, to form closed loop control by steady flow field
System, using the time interval feedback control of closed loop, constantly amendment image acquisition interval, it is accurate that raising particle flow field is tested the speed
Property, efficiently solve high speed Unsteady Flow, the problem of tachometric survey inaccuracy.
2. the present invention utilizes the ability of FPGA parallel processing, simple algorithm is quickly handled, utilizes DSP exploitation complicated algorithm
Flexibility and real-time realize the cross correlation algorithm of higher precision；Embedded image acquisition and processing board system is small in size, property
Can be high, it is easy to carry, it is suitble to the particle flow field of various occasions to test the speed.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of the built-in image collection based on particle image velocimetry and processing system of the invention；
Fig. 2 is one-shot camera structural schematic diagram of the invention；
Fig. 3 is flow field algorithm flow chart of the invention；
Fig. 4 is working-flow figure of the invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
Not constituting a conflict with each other can be combined with each other.
As shown in Figure 1, a kind of built-in image collection and place based on particle image velocimetry provided in an embodiment of the present invention
Reason system comprising one-shot camera and built-in image collection processing module, wherein as shown in Fig. 2, one-shot camera includes successively
Camera lens, optical filter, half reflection prism and two CCD cameras of setting, the one-shot camera adjust optical path by half reflection prism, will
Light beam is divided into two-way and uses for CCD camera, and the equal length of two CCD camera optical paths, the CCD camera different-time exposure acquires grain
Subgraph, and the particle picture of acquisition is sent in built-in image collection processing module；Built-in image collection handles mould
Block includes image preprocessing submodule and image acquisition time prediction submodule, and image preprocessing submodule is specially FPGA (existing
Field programmable gate array) chip, image acquisition time prediction submodule is specially DSP (digital signal processor) chip, wherein
Fpga chip is used to complete the pretreatment of particle picture, and pretreated particle picture is sent to dsp chip, dsp chip
For completing improved cross correlation algorithm and prediction of speed calculating based on FFT, specifically, dsp chip is for carrying out based on FFT
Improved cross correlation algorithm calculating speed field, and prediction calculating is carried out to velocity field, then calculates predetermined speed off field
Image acquisition interval, and time interval signal is sent to the time for exposure that two CCD cameras are realized in one-shot camera
The control at interval.
Further, fpga chip uses the processing of the model 6VLX240TFF1156 (V6) of match Sentos (Xilinx) company
Device after FPGA receives image, if image is excessive, is just stored to what FPGA was used for realizing Image Pretreatment Algorithm
In DDR3, pretreatment operation, such as the gaussian filtering of removal noise are carried out to image later, median filtering etc., FPGA is good
Parallel processing computing function can greatly shorten the time of Preprocessing Algorithm；DSP use Texas Instrument's (TI) company model for
The processor of TMS320C6678 (C6678), the DSP are 8 core processors, and monokaryon can reach the dominant frequency of 1.25GHz, can be flexible
Complicated algorithm is effectively handled, and is operated with real-time.
More specifically, built-in image collection processing module can also include two Cameralink interfaces, Serial
RapidIO (SRIO) interface, Ethernet interface, GPIO (universal input/output) interface, clock module, power module and storage
Module, these above-mentioned components are installed in mechanical cover.Wherein, Cameralink interface is used to receive CCD camera shooting
Image, highest can realize 5.4Gbps transmission rate；SRIO interface is used for the high-speed communication of FPGA and DSP, is based on SRIO agreement,
Highest can be realized the transmission rate of 4 road 5Gbps；Ethernet interface by by Ethernet protocol by image processing module based on
It calculates result, that is, particle picture VELOCITY DISTRIBUTION and is sent to PC；GPIO interface is for realizing the calculating of built-in image collection processing module
Feedback control is carried out to one-shot camera after the result of flow field out；Clock module guarantees for providing stable clock for entire board
Each proper device operation；Power module provides stabilization for distributing different voltage to different device circuitries for entire board
Power supply；Memory module is made of the Flash of DDR3 and two piece of 64MB of 2 pieces of 1GB, and the storage for being respectively used to FPGA and DSP is visited
It asks.In the present embodiment, Ethernet interface is general gigabit Ethernet mouth, ethernet control chip 88E1111；Clock module is main
It is made of crystal oscillator and CDCE62005 chip.
Specifically, image acquisition and processing board can be configured with JTAG (joint test working group) interface, jtag interface is used
In user algorithm routine programming to FPGA and DSP to realize algorithm flow field and control function, convenient for user in FPGA and DSP
Secondary development is carried out to Image Pretreatment Algorithm and particle image velocimetry algorithm and flow field velocity prediction algorithm, is maintained easily.
In the present embodiment, selected JTAG is general 10Pin jtag interface.
As the preferred embodiment of the present invention, the improved cross correlation algorithm calculating speed field based on FFT, as shown in figure 3,
Specifically: respectively to adjacent moment t1,t2Cross-correlation query window in the particle picture of acquisition carries out discrete Fourier transform,
Calculate its cross-correlation function；2 times of up-samplings are done by the way of zero padding matrix, then carry out Fourier inversion, are obtained mutually
Close the position (x of peak value0,y0)；Due to having carried out 2 times of up-samplings, then the first step displacement of practical off position mutually is x0/2,y0/2；
Again with positionCentered on obtain 1.5 × 1.5 neighborhoods in pixel window M, respectively in adjacent moment two images
Pixel window M seeks Fourier transformation, obtains two frequency-domain functions, calculate current time window M frequency-domain function and it is latter when
The product for carving the conjugation of the frequency-domain function of window M, is k times to the product by the way of zero padding and up-samples, then carry out in Fu
Leaf inverse transformation obtains cross-correlation peak value x1,y1, i.e. peak position is (x1,y1)；According to (x0,y0) and (x1,y1) acquire fine positioning
Peak position: Calculating cross-correlation peak value maximum position can be in the hope of the speed of a width picture
Vector field pattern is spent, after specifically finding out fine positioning peak position, peak position shows cross-correlation coefficient maximum, that is, inquires
Window shifts arrived here, due to two adjacent width particle picture time intervals be it is known, be equal to according to displacement divided by the time
Speed, to obtain speed.Computing cross-correlation is carried out to mutual corresponding all query windows of two width particle pictures, can be obtained
To the velocity field of entire image.
Specifically, the improved cross correlation algorithm based on FFT mainly comprises the steps that
(1) two-dimensional fast fourier transform is carried out to the cross-correlation query window of two width particle pictures of adjacent moment acquisition
(2D-FFT), and seek cross-correlation function, wherein t1Query window image grayscale function in moment particle picture is f (x, y), window
Mouth is having a size of M × M, t2Query window image grayscale function in moment particle picture is g (x, y), and window size is M × M, M
Size generally take 16,32 or 64.
DFT transform is carried out to the function f (x, y) (x, y represent the coordinate in spatial domain) that window size is M × M, can be obtained
It is F (u, v) to its transforming function transformation function:
Wherein: u, v are the coordinates of frequency domain, and j is imaginary unit.
DFT inverse transformation is carried out to function F (u, v), its available transforming function transformation function is f (x, y):
For t1Query window image grayscale function f (x, y) in moment particle picture and for t2In moment particle picture
Query window image grayscale function g (x, y) between cross-correlation function rfg(x, y) are as follows:
Wherein, m and n is parameter to be asked, and enables F (u, v) and G (u, v) is the Fourier transformation of f (x, y) and g (x, y) respectively,
Then:
From the above equation, we can see that cross-correlation function rfg(x, y) is F (u, v) G* (u, v) inversefouriertransform, and * indicates complex conjugate；
(2) 2 times of up-samplings are carried out to F (u, v) G* (u, v) using the method for zero padding, to the cross-correlation function of up-sampling
Fourier inversion is carried out, the peak value of the cross-correlation function after detecting Fourier inversion obtains (x0,y0), then obtain window
Peak position
(3) further fine positioning, with peak position in (2)On the basis of obtain 1.5 × 1.5 neighborhoods in pixel
The function of window matrix M, M in two images is respectively f'(x, y) and g'(x, y), to f'(x, y) and g'(x, y) ask in Fu
Leaf transformation obtains two frequency-domain functions F'(u, v) and G'(u, v),
G*'(u, v) be G'(u, v) conjugation, calculate the frequency-domain function F'(u, v of current time window M) and later moment in time
The conjugation G*'(u, v of the frequency-domain function of window M) product F'(u, v) G*'(u, v), to F'(u, v by the way of zero padding)
G*'(u, v) it is k times and up-samples, then Fourier inversion is carried out, obtain cross-correlation peak value (x1,y1)。
(4) according to peak position (x0,y0), (x1,y1) the peak position x of query window 1 and query window 2 is calculated,
Y are as follows:
Cross correlation algorithm is carried out by the step, can be improved peak detection accuracy, realizes that calculating speed is big on DSP
Width is promoted, and brings beneficial effect for particle image velocimetry.
Below to the specific mistake for carrying out Image Acquisition and processing using built-in image collection and processing system of the invention
Journey is described in detail, and includes the following steps:
(1) particle picture is acquired using two CCD camera different-time exposures in one-shot camera, and by the particle picture of acquisition
It is sent in built-in image collection processing module；
(2) pretreatment of the particle picture is completed using fpga chip, and pretreated particle picture is sent to
Dsp chip, the dsp chip carry out velocity field pre- for carrying out the improved cross correlation algorithm calculating speed field based on FFT
It surveys and calculates, then calculate the image acquisition interval of predetermined speed off field, and time interval signal is sent to described point
The control of the exposure time interval of described two CCD cameras is realized in light camera.
Specifically, as shown in figure 4, two independent CCD cameras will be acquired by Cameralink interface in one-shot camera
Image transmitting to FPGA；After FPGA receives image, if image is excessive, just it is stored in the DDR3 that FPGA is used, later
Pretreatment operation, such as the gaussian filtering of removal noise, median filtering etc. are carried out to image；After image preprocessing, FPGA is logical
SRIO is crossed by the DDR3 of image transmitting to DSP, SRIO maximum can support the transmission rate of 20Gbps；DSP visits its DDR3
Image is asked and is taken out, the improved cross correlation algorithm operation based on FFT is carried out, calculates cross-correlation peak value maximum position, so that it may
In the hope of the velocity vector field pattern of a width picture；After calculating speed image field, DSP is predicted by Kalman filtering algorithm
The velocity field of subsequent time can find out speed scalar maximum value Vw according to the velocity field of prediction, by the maximum value of piv speed scalar
The principle of side length pixel quantity no more than 1/4 query window M calculates image required for subsequent time two images and adopts
The size for collecting time interval Xw/Vw, wherein Xw=1/4 × window M side length, such as window M is 16 × 16, then Xw=1/4 ×
16=4；Then the time interval obtained will be calculated, one-shot camera fed back to by the GPIO interface in image acquisition and processing module,
And the different-time exposure of two CCD cameras is controlled, so that exposure time interval is as much as possible equal to the Image Acquisition interval predicted
Time completes closed-loop control with this.In addition, between the calculating and speed image field and acquisition time of DSP progress speed image field
Every prediction can be transmitted to PC by the Ethernet interface of image acquisition and processing board, be convenient for human-computer interaction.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should all include
Within protection scope of the present invention.
Claims (8)
1. a kind of built-in image collection and processing system based on particle image velocimetry, which is characterized in that including one-shot camera
With built-in image collection processing module, in which:
The one-shot camera includes the camera lens set gradually, optical filter, half reflection prism and two CCD cameras, and the one-shot camera is logical
It crosses the half reflection prism and adjusts optical path, light beam is divided into two-way and is used for the CCD camera, the CCD camera timesharing exposes
Light collection particle picture, and the particle picture of acquisition is sent in the built-in image collection processing module；
The built-in image collection processing module includes that image preprocessing submodule and image acquisition time predict submodule,
In, described image pretreatment submodule is used to complete the pretreatment of the particle picture, and pretreated particle picture is sent out
It send into described image acquisition time prediction submodule；Described image acquisition time prediction submodule is for carrying out based on FFT's
Improved cross correlation algorithm calculates the velocity field at the pretreated particle picture current time, and to the speed of subsequent time
Field is predicted, then calculates the image acquisition interval of predetermined speed off field, and time interval signal is sent to institute
State the control that described two CCD camera exposure time intervals are realized in one-shot camera；The improved cross-correlation based on FFT
Algorithm calculates the velocity field at the pretreated particle picture current time specifically: respectively to current time and subsequent time
The query window of the particle picture of acquisition carries out Fast Fourier Transform (FFT), obtains the frequency-domain function of two query windows, and calculate
The product of the conjugation of current time query window frequency-domain function and subsequent time query window frequency-domain function；Using zero padding matrix
Mode 2 times of up-samplings are done to the product, then carry out Fourier inversion, obtain cross-correlation peak value position (x0,y0), and obtain
Window peak positionThe particle at current time and subsequent time acquisition is obtained centered on the window peak position again
Pixel window M in 1.5 × 1.5 neighborhoods in image；Respectively to the pixel window in current time and subsequent time two images
Mouthful M seeks Fourier transformation, obtains the frequency-domain function of two window M, and calculate current time window M frequency-domain function and lower a period of time
Carve the product of the conjugation of the frequency-domain function of window M；K times is to the product by the way of zero padding to up-sample, then is carried out in Fu
Leaf inverse transformation obtains cross-correlation peak value position (x1,y1)；According to (the x0,y0) and (x1,y1) acquire fine positioning position:Current time velocity field is acquired according to the fine positioning position.
2. built-in image collection and processing system based on particle image velocimetry as described in claim 1, which is characterized in that
Described image pretreatment submodule is specially fpga chip, and described image acquisition time predicts that submodule is specially dsp chip；Institute
It states CCD camera and image is transmitted in built-in image collection processing module by Cameralink interface, realize highest
The transmission rate of 5.4Gbps；The built-in image collection processing module realizes the control to one-shot camera by GPIO interface.
3. built-in image collection and processing system based on particle image velocimetry as claimed in claim 2, which is characterized in that
The fpga chip and dsp chip are by the realization high-speed communication of SRIO interface, specifically, realizing the transmission speed of 4 road 5Gbps of highest
Rate.
4. the built-in image collection and processing system as described in any one of claims 1-3 based on particle image velocimetry,
It is characterized in that, PC is sent for the calculated result of image acquisition and processing module using Ethernet protocol by Ethernet interface；Institute
Stating pretreatment is gaussian filtering or median filtering.
5. a kind of built-in image collection and processing method based on particle image velocimetry, which comprises the steps of:
(1) particle picture is acquired using two CCD camera different-time exposures in one-shot camera, and the particle picture of acquisition is transmitted
Into built-in image collection processing module；
(2) pretreatment of the particle picture is completed using image preprocessing submodule, and pretreated particle picture is sent out
It send to image acquisition time and predicts that submodule, image acquisition time prediction submodule are used to carry out improved mutual based on FFT
Related algorithm calculates the velocity field at the pretreated particle picture current time, and carries out to the velocity field of subsequent time pre-
It surveys and calculates, then calculate the image acquisition interval of predetermined speed off field, and time interval signal is sent to described point
The control of the exposure time interval of described two CCD cameras is realized in light camera；The improved cross-correlation based on FFT is calculated
Method calculates the velocity field at the pretreated particle picture current time specifically: adopts respectively to current time and subsequent time
The query window of the particle picture of collection carries out Fast Fourier Transform (FFT), obtains the frequency-domain function of two query windows, and calculate and work as
The product of the conjugation of preceding moment query window frequency-domain function and subsequent time query window frequency-domain function；Using zero padding matrix
Mode does 2 times of up-samplings to the product, then carries out Fourier inversion, obtains cross-correlation peak value position (x0,y0), and obtain window
Mouth peak positionThe particle figure at current time and subsequent time acquisition is obtained centered on the window peak position again
The pixel window M in 1.5 × 1.5 neighborhoods as in；Respectively to the pixel window M in current time and subsequent time two images
Fourier transformation is sought, the frequency-domain function of two window M is obtained, and calculates the frequency-domain function and subsequent time of current time window M
The product of the conjugation of the frequency-domain function of window M；K times is to the product by the way of zero padding to up-sample, then carries out Fourier
Inverse transformation obtains cross-correlation peak value (x1,y1)；According to (the x0,y0) and (x1,y1) acquire fine positioning position: Current time velocity field is acquired according to the fine positioning position.
6. built-in image collection and processing method based on particle image velocimetry as claimed in claim 5, which is characterized in that
Described image pretreatment submodule is specially fpga chip, and described image acquisition time predicts that submodule is specially dsp chip；Institute
Stating one-shot camera includes that the camera lens set gradually, optical filter, half reflection prism and two CCD cameras, the one-shot camera pass through institute
It states half reflection prism and adjusts optical path, light beam is divided into two-way and is used for the CCD camera.
7. built-in image collection and processing method based on particle image velocimetry as claimed in claim 5, which is characterized in that
Image is transmitted in built-in image collection processing module by the CCD camera by Cameralink interface, realizes highest
The transmission rate of 5.4Gbps；The built-in image collection processing module realizes the control to one-shot camera by GPIO interface；
The fpga chip and dsp chip are by the realization high-speed communication of SRIO interface, specifically, realizing the transmission speed of 4 road 5Gbps of highest
Rate.
8. such as the described in any item built-in image collections and processing method based on particle image velocimetry of claim 5-7,
It is characterized in that, PC is sent for the calculated result of image acquisition and processing module using Ethernet protocol by Ethernet interface；Institute
Stating pretreatment is gaussian filtering or median filtering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710319285.9A CN107144703B (en) | 2017-05-09 | 2017-05-09 | Built-in image collection and processing system and method based on particle image velocimetry |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710319285.9A CN107144703B (en) | 2017-05-09 | 2017-05-09 | Built-in image collection and processing system and method based on particle image velocimetry |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107144703A CN107144703A (en) | 2017-09-08 |
| CN107144703B true CN107144703B (en) | 2019-06-18 |
Family
ID=59777279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710319285.9A Active CN107144703B (en) | 2017-05-09 | 2017-05-09 | Built-in image collection and processing system and method based on particle image velocimetry |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107144703B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108398572B (en) * | 2018-02-08 | 2020-07-17 | 上海理工大学 | Particle image speed measurement method and device based on Voronoi diagram |
| CN109035324A (en) * | 2018-07-20 | 2018-12-18 | 润电能源科学技术有限公司 | A kind of coal dust flow-speed measurement method and device based on image recognition |
| CN110751600B (en) * | 2019-09-03 | 2021-04-30 | 哈尔滨新光光电科技股份有限公司 | Miniaturized camera alink digital acquisition device and system |
| CN110687315B (en) * | 2019-10-31 | 2020-12-29 | 华中科技大学 | Flow field velocity measuring system capable of adaptively adjusting time interval |
| CN111693729B (en) * | 2020-06-28 | 2021-07-23 | 中国科学院力学研究所 | Particle image velocity measurement method and device based on global optimization |
| CN112730884A (en) * | 2020-12-28 | 2021-04-30 | 同济大学建筑设计研究院(集团)有限公司 | Method and system for determining high-altitude wind speed by adopting imaging means |
| CN114755449B (en) * | 2022-06-14 | 2022-08-26 | 中国空气动力研究与发展中心设备设计与测试技术研究所 | Particle image speed measurement distortion correction device and method |
| CN117825743A (en) * | 2024-03-04 | 2024-04-05 | 浙江大学 | PIV speed measurement method and device based on Fourier feature enhancement and global matching |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07120489A (en) * | 1993-10-26 | 1995-05-12 | Kowa Co | Fluid speed meausring method |
| JP2000283993A (en) * | 1999-03-29 | 2000-10-13 | Hitachi Ltd | Method and device for measuring flow velocity |
| CN101464471A (en) * | 2009-01-15 | 2009-06-24 | 清华大学 | Real-time measurement system for solid model surface flow field and its reprocessing method |
| CN102564508A (en) * | 2011-12-14 | 2012-07-11 | 河海大学 | Method for implementing online tests of stream flow based on video images |
| CN103604947A (en) * | 2013-11-28 | 2014-02-26 | 华中科技大学 | Flow field state measuring method with adaptive adjusted time resolution |
| CN103698554A (en) * | 2013-12-17 | 2014-04-02 | 华中科技大学 | Flow field real-time precise measuring system and method |
| CN105915813A (en) * | 2016-05-11 | 2016-08-31 | 华中科技大学 | Image acquisition frame rate adaptive adjustment device |
-
2017
- 2017-05-09 CN CN201710319285.9A patent/CN107144703B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07120489A (en) * | 1993-10-26 | 1995-05-12 | Kowa Co | Fluid speed meausring method |
| JP2000283993A (en) * | 1999-03-29 | 2000-10-13 | Hitachi Ltd | Method and device for measuring flow velocity |
| CN101464471A (en) * | 2009-01-15 | 2009-06-24 | 清华大学 | Real-time measurement system for solid model surface flow field and its reprocessing method |
| CN102564508A (en) * | 2011-12-14 | 2012-07-11 | 河海大学 | Method for implementing online tests of stream flow based on video images |
| CN103604947A (en) * | 2013-11-28 | 2014-02-26 | 华中科技大学 | Flow field state measuring method with adaptive adjusted time resolution |
| CN103698554A (en) * | 2013-12-17 | 2014-04-02 | 华中科技大学 | Flow field real-time precise measuring system and method |
| CN105915813A (en) * | 2016-05-11 | 2016-08-31 | 华中科技大学 | Image acquisition frame rate adaptive adjustment device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107144703A (en) | 2017-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107144703B (en) | Built-in image collection and processing system and method based on particle image velocimetry | |
| Jin et al. | A semi-direct approach to structure from motion | |
| CN110095968B (en) | Off-axis digital holographic image reconstruction device and method and microscopic imaging system | |
| CN103026171B (en) | Image processing apparatus and image processing method | |
| Florack et al. | The intrinsic structure of optic flow incorporating measurement duality | |
| CN106066987B (en) | Parameter on-line calibration and compensation in TOF imaging | |
| CN103959089A (en) | Depth imaging method and apparatus with adaptive illumination of an object of interest | |
| CN101243458A (en) | Image acquisition system for creating a depth map | |
| CN102074045A (en) | System and method for projection reconstruction | |
| CN106885622A (en) | A kind of big visual field multiple spot three-dimensional vibrating measuring method | |
| CN105915813A (en) | Image acquisition frame rate adaptive adjustment device | |
| CN110235431A (en) | Electronic stability augmentation method, image capture device, moveable platform | |
| CN106443691A (en) | Three-dimensional imaging system based on digital micromirror device (DMD) and imaging method | |
| McLeod et al. | A joint analysis for cosmology and photometric redshift calibration using cross-correlations | |
| CN105787464B (en) | A kind of viewpoint scaling method of a large amount of pictures in three-dimensional scenic | |
| CN109635870A (en) | Data processing method and device | |
| CN103747186A (en) | Time-division three-path image acquisition device and calibration method for same | |
| Wernet | Real-time background oriented schlieren: catching up with knife edge schlieren | |
| Vermaak et al. | Extended object tracking using particle techniques | |
| Huber et al. | Application of time-delay estimation to mixed reality multisensor tracking | |
| CN108964998A (en) | A kind of network entity behavior Singularity Detection and device | |
| CN104657580B (en) | The method and system that 3-D photography for catching Global Lighting equipment illuminates | |
| Mahapatra et al. | Comparison of artificial immune system and particle swarm optimization techniques for error optimization of machine vision based tool movements | |
| Mihaldinec et al. | Method for joint flexion angle estimation using UWB ranging with clock model compensation | |
| Ju et al. | Direct variational perspective shape from shading with Cartesian depth parametrisation |
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 |