KR20220109203A - System and method for self-calibration to improve performance of tomographic particle image velocimetry - Google Patents

Abstract

Various embodiments relate to a camera self-calibration method and system to improve the performance of tomographic particle image velocimetry (PIV), which may be configured to perform calibration on cameras for tomographic PIV, create a disparity map for particles using particle images obtained through cameras, and perform self-calibration for the cameras based on an error for each of the particles calculated from the disparity map.

Description

SYSTEM AND METHOD FOR SELF-CALIBRATION TO IMPROVE PERFORMANCE OF TOMOGRAPHIC PARTICLE IMAGE VELOCIMETRY

Various embodiments relate to a camera self-calibration method and system for improving the performance of tomographic flow field metrology.

Among the methods for measuring the velocity of a number of fluids, the particle image velocimetry (PIV) method is a non-contact quantitative measurement method, and has the advantage of quantitatively analyzing the velocity distribution of the entire flow field. applied in the field. In particular, the tomographic PIV (Tomographic PIV) method has the advantage of being able to precisely measure the three-dimensional flow of a target, and many studies are currently being conducted. In order for this tomographic PIV method to be widely used in the industrial field, it is required to develop an algorithm capable of minimizing errors along with the development of hardware.

Various embodiments provide a camera self-calibration method and system for minimizing 3D metrology errors to improve performance of tomographic flow field metrology.

A camera self-calibration method of a tomographic flow field measurement system according to various embodiments includes: performing calibration on cameras for tomographic flow field measurement, using a particle image obtained through the cameras, It may include producing a disparity map (disparity map), and based on the error of each of the particles calculated from the disparity map, performing self-calibration for the cameras.

A tomographic flow field measurement system for camera self-calibration according to various embodiments includes a plurality of cameras for tomographic flow field measurement, and at least one processor connected to the cameras, wherein the processor includes the camera Calibration is performed on the particles, and a disparity map for the particles is produced using the particle image obtained through the cameras, and based on the error for each of the particles calculated from the disparity map, the The cameras may be configured to perform self-calibration.

According to various embodiments, the tomographic flow field measurement system calculates a three-dimensional position using image coordinates obtained from multiple cameras in order to improve measurement performance in three-dimensional measurement of a flow field, and re-calculates the calculated three-dimensional position. By re-converting to image coordinates to evaluate and correct the errors of the particles, and to calibrate the cameras, the calibration error caused by the error occurring in the calibrator in 3D measurement or the error occurring in the image processing process to find the center of the particle is minimized. Measurement accuracy can be improved.

1 is a diagram illustrating a tomographic flow field metrology system in accordance with various embodiments.
2 is a diagram illustrating a method of a tomographic flow field metrology system in accordance with various embodiments.
FIG. 3 is a diagram showing the internal configuration of the camera self-calibration step of FIG. 2 .
FIG. 4 is a view for explaining in detail the camera self-calibration step of FIG. 2 .
5 is a diagram exemplarily illustrating a disparity map.

Hereinafter, various embodiments of the present document will be described with reference to the accompanying drawings.

Various embodiments provide a 3D particle initially calculated as a method of improving the accuracy of a camera observation equation in order to minimize a velocity measurement error generated in 3D particle image velocimetry (PIV) (hereinafter also referred to as PIV). We present the technology development for the self-calibration method that corrects the observation equation of the camera using the position information of the camera.

A method of measuring the velocity distribution of a flow field by PIV (Particle Image Velocity Meter) is a non-contact velocity measurement method of a flow field. Tracking particles moving like a fluid in the flow field are injected, the particles are irradiated with a laser, and the particles are The tomographic PIV (hereinafter, also referred to as tomo PIV) method for three-dimensional measurement is a velocimetry that measures the velocity by analyzing the image acquired by the camera with the reflected light reflected from the field. applied a lot.

Tomographic PIV is a method of measuring the three-dimensional velocity distribution of a flow field using multiple cameras. It determines the position of a three-dimensional particle using the time difference from three or more cameras, and reconstructs the particle in a three-dimensional virtual space. This is a method of producing a 3D voxel image and calculating a 3D velocity vector from these time series images.

In tomographic PIV, calibration is performed on cameras for three-dimensional measurement, and particles of a high-density particle image are reconstructed in a three-dimensional virtual voxel space using the camera parameters obtained therefrom, and finally these voxels The three-dimensional velocity vector distribution is obtained through cross-correlation calculation for the time series image of the image.

In order to improve the performance of tomographic PIV, accurate positioning of particles is very important. For this reason, various self-correction methods have been developed. In various embodiments, in order to improve the degree of calibration of the camera, the position of the particle from the particle image is calculated, the position of the particle is corrected by the error, and the camera error is minimized by re-calibrating the observation equation using 10 parameters. The content of the correction technique is presented.

1 is a diagram illustrating a tomographic flow field metrology system 100 in accordance with various embodiments.

Referring to FIG. 1 , a tomographic flow field measurement system 100 according to various embodiments may include a plurality of cameras 110 and a processor 120 . The processor 120 may perform tomographic flow field measurement using the cameras 110 . In this case, the processor 120 may perform calibration and self-calibration for the cameras 110 in order to improve the performance of tomographic flow field measurement. That is, the processor 120 may correct and self-correct the camera observation equation in order to minimize the 3D measurement error.

Specifically, the tomographic flow field measurement system 100 includes a first step of performing camera calibration using a plurality of cameras 110, a second step of producing a disparity map using a particle image, and You can proceed to the third step of calculating the error and performing self-correction.

In the first step, the tomographic flow field measurement system 100 installs a plurality of cameras 110 and a calibrator, and uses a calibration plate to acquire a plurality of images through z-axis movement at regular intervals. have. In addition, the tomographic flow field measurement system 100 may perform camera calibration using a three-dimensional calibrator coordinate value that matches the two-dimensional image coordinates of points on the calibration plate in the obtained image.

In the second step, the tomographic flow field measurement system 100 may acquire a particle image, and obtain a particle center from the acquired image. And, the tomographic flow field measurement system 100 determines the three-dimensional positions of all particles, and meshes the two-dimensional error values generated when the determined three-dimensional particles are converted back into a two-dimensional image image. A disparity map can be produced by a method of reconstructing a virtual computation lattice formed of .

In the third step, the tomographic flow field measurement system 100 calculates the central movement amount of the particle for each mesh point from the result obtained in the second step, and calculates the two-dimensional image coordinate value of the calibration point by the calculated error. can be converted Finally, 3D self-correction may be performed using 3D mesh coordinates matching the transformed 2D image coordinates.

As described above, the tomographic flow field measurement system 100 calculates a three-dimensional position using the image coordinates obtained from several cameras 110 in order to improve the measurement performance in the three-dimensional measurement of the flow field, and calculates the calculated By re-converting the three-dimensional position back into image coordinates to evaluate and correct the errors of the particles, and correct the cameras 110, errors occurring in the calibrator in three-dimensional measurement or the image processing process to find the center of the particles Measurement accuracy can be improved by minimizing calibration errors caused by errors.

2 is a diagram illustrating a method of a tomographic flow field metrology system 100 in accordance with various embodiments.

Referring to FIG. 2 , the method of the tomographic flow field measurement system 100 is a non-contact quantitative measurement method using a plurality of cameras 110 , and may measure the flow field. According to various embodiments, the tomographic flow field metrology system 100 performs calibration on the cameras 110 using a calibrator to minimize errors, and acquires a low-density particle image before acquiring a high-density particle image. Accordingly, the 3D measurement error can be minimized by producing a disparity map showing the error and performing self-calibration for the cameras 110 through image coordinate transformation.

First, the tomographic flow field measurement system 100 may acquire images of the calibrator through the cameras 110 in step 210 . The calibrator can provide three-dimensional information by arranging circular points in the form of a grid at regular intervals on a flat plate to determine (x, y) coordinate values and acquiring images while moving by a certain z interval. The processor 120 may obtain the center of the points through image processing for each acquired image. According to an embodiment, the processor 120 uses the Gaussian filtering (Gaussian filtering, logical multiplication premised on the standard distribution of the brightness distribution of particles as a starting point) given as shown in Equation 1 below). You can find the center coordinates.

는 입자의 크기로 교정기 입자의 크기에 해당할 수 있다. Here, F _Gaussian is a Gaussian filter, and the maximum brightness of the particles may correspond to twice the average value of the image. x and y are the distances from the center of the particle

is the size of the particle and may correspond to the size of the corrector particle.

The tomographic flow field measurement system 100 may perform calibration for each of the cameras 110 using a calibrator in step 220 . The processor 120 may perform calibration by obtaining values of variables of an observation equation using a least squares method using three-dimensional space coordinates that match the two-dimensional calibration point coordinates in the obtained image. Various camera observation equations exist, and according to an embodiment, the processor 120 may use an observation equation as shown in Equation 2 below.

Here, x and y may be an image coordinate system, and X, Y, and Z may be a corrector coordinate system. c _x , c _y represents the proportion of the image coordinate system and the object coordinate system, d is the distance between the camera center and the object coordinate system point O, m _x , m _y are the offset amounts of the central axis, and the subscript m means rotation. That is, (x _m , y _m , z _m ) may be a value obtained by 3D rotation of (X, Y, Z). may be the amount of refraction of the image considering up to the second term.

The tomographic flow field measurement system 100 may perform self-calibration for each of the cameras 110 based on the disparity map in operation 230 . This will be described in detail later with reference to FIGS. 3, 4 and 5 .

The tomographic flow field measurement system 100 may acquire a high-density tracking particle image through the cameras 110 in step 240 . The tomographic flow field measurement system 100 may inject tracer particles into the flow field and acquire a particle image. Here, it is classified into low-density PIV and high-density PIV according to the amount of injected particles, and tomographic flow field measurement is a method of measuring the flow field by cross-correlation using high-density particle distribution. can be obtained.

The tomographic flow field measurement system 100 may reconstruct the particle image in step 250 . Various methods exist for reconstructing particle images. According to an embodiment, when the processor 120 defines a voxel image (an image defined in a three-dimensional space like a pixel in a two-dimensional image) in a three-dimensional space and projects the voxel image into a two-dimensional image, The voxel image can be reconstructed using MART (Multiple Algebraic Reconstruction Technique), which reconstructs the voxel region to be as identical to the experimental image as possible.

The tomographic flow field measurement system 100 may perform velocity measurement by cross-correlation in operation 260 . The processor 120 may calculate the vector field as shown in Equation 3 below by using the cross-correlation by the 3D FFT.

,

는 복셀 영상을 의미할 수 있다. where X, Y, and Z are reference regions in a 3D voxel,

,

may mean a voxel image.

FIG. 3 is a diagram showing the internal configuration of the self-calibration step (step 230) of the camera 110 of FIG. 2 .

Referring to FIG. 3 , a self-calibration step (step 2320 ) of the camera 110 may be added to improve the performance of flow field measurement.

First, the tomographic flow field measurement system 100 may acquire a low-density tracking particle image in step 310 . When 3D reconstruction is performed with particles of a 2D image, real particles and ghost particles may coexist. These false particles increase exponentially as the number of particles increases. Accordingly, when the processor 120 obtains a low-density tracking particle image to obtain a three-dimensional position, the number of false particles can be minimized and three-dimensional position values with a high proportion of real particles can be obtained.

The tomographic flow field measurement system 100 may produce a disparity map in operation 320 . The processor 120 may produce a disparity map for each location from three-dimensional location values having a high proportion of actual particles.

The tomographic flow field measurement system 100 may perform self-calibration for each of the cameras 110 using the corrected coordinates based on the disparity map in operation 330 . The processor 120 may correct the coordinates through error analysis. In addition, the processor 120 may minimize an error by performing self-calibration for each of the cameras 110 using the corrected coordinates.

FIG. 4 is a diagram for explaining in detail the self-calibration step (step 230) of the camera 110 of FIG. 2 .

Referring to FIG. 4 , the tomographic flow field measurement system 100 may search for centroids of particles for all images in order to produce a disparity map. Although there are various methods, the processor 120 may determine the center of a particle by applying a Gaussian filtering method according to an embodiment. The tomographic flow field measurement system 100 may produce a disparity map as described below.

The tomographic flow field measurement system 100 may divide a 3D region to be calculated with respect to the particle image at regular intervals in step 410 . At this time, the tomographic flow field measurement system 100 may search for the positions of particles in all camera images. The tomographic flow field measurement system 100 may obtain the equation of a straight line passing through the particles of the central camera 1 (epipolar line) in step 420 . The tomographic flow field measurement system 100 may project the equation of the straight line to the camera 2 in step 430 and obtain particles existing on the straight line. The tomographic flow field measurement system 100 may determine a three-dimensional position using particles of two cameras, ie, camera 1 and camera 2, in step 440 . The tomographic flow field measurement system 100 may project the three-dimensional particle onto the image of the remaining cameras (eg, camera 3, camera 4, ...) in step 450 and determine whether or not the particle is present. The tomographic flow field measurement system 100 determines a new three-dimensional position for the points of all cameras if particles are present in all the remaining cameras (eg, camera 3, camera 4, ...) in step 460, and in step 470 You can create a list by adding particles to the corresponding voxel region. The tomographic flow field measurement system 100 calculates the error between the 2D position and the particle center when the 3D particle of each voxel area is projected to each camera in step 480, and draws the virtual particle in the area separated by the error. As a result, a disparity map corresponding to each region and camera can be created. The tomographic flow field measurement system 100 searches for the maximum value of the disparity map in step 490, calculates an error, and performs self-correction using the data pairing the point on the image corrected by the error and the three-dimensional position point. can be done

5 is a diagram exemplarily illustrating a disparity map.

Referring to FIG. 5 , a calculated disparity map at z=0 of a certain camera 110 is shown. If tens of thousands of particles are counted, an image gathered at one point as shown in FIG. 5 may be obtained, and the distance from the center may indicate an error in the two-dimensional image. The data can be corrected by correcting as much as this error.

As described above, the tomographic flow field measurement system 100 calculates a three-dimensional position using the image coordinates obtained from several cameras 110 in order to improve the measurement performance in the three-dimensional measurement of the flow field, and calculates the calculated By re-converting the three-dimensional position back into image coordinates to evaluate and correct the errors of the particles, and correct the cameras 110, errors occurring in the calibrator in three-dimensional measurement or the image processing process to find the center of the particles Measurement accuracy can be improved by minimizing calibration errors caused by errors.

The camera 110 self-calibration method of the tomographic flow field measurement system 100 according to various embodiments includes performing calibration on the cameras 110 for tomographic flow field measurement (step 220), the camera 110 ) using a particle image obtained through ) may include performing self-calibration for the (steps 230 and 330).

According to various embodiments, performing calibration on the cameras 110 (step 220) includes obtaining a plurality of images through the cameras 110 while moving the calibrator at a predetermined interval, and obtaining The method may include performing calibration on each of the cameras 110 by using a three-dimensional coordinate value that matches two-dimensional image coordinates for points on a corrector in each of the images.

According to various embodiments, the step of producing the disparity map (step 320) includes, through the cameras 110, obtaining a particle image (step 310), obtaining the center of the particles from the particle image, and the particle image Determining a three-dimensional position for each of the particles in the step (steps 410 to 470), projecting each of the particles to each of the cameras 110 to determine a two-dimensional position for each of the particles (step 480) ), and based on the two-dimensional error between the center and the two-dimensional position, producing a disparity map for each of the cameras 110 (operation 480).

According to various embodiments, the step (step 320) of producing a disparity map for each of the cameras 110 includes a two-dimensional error for each of the particles by using a mesh formed at a predetermined interval in the particle image. By reconstructing , a disparity map for each of the cameras 110 may be produced.

According to various embodiments, the step of producing a disparity map for each of the cameras 110 (step 320) includes drawing virtual particles corresponding to each of the particles based on a two-dimensional error, thereby making the camera 110 ) can produce a disparity map for each of them.

According to various embodiments, performing self-calibration on the cameras 110 (step 330) may include calculating an error for each of the particles in the particle image from the disparity map for each of the cameras 110 . and performing self-calibration for each of the cameras 110 using the calculated error and the three-dimensional position.

The tomographic flow field measurement system 100 for camera 110 self-calibration according to various embodiments includes a plurality of cameras 110 for tomographic flow field measurement, and at least one connected to the cameras 110 . It may include a processor 120 .

According to various embodiments, the processor 120 performs calibration on the cameras 110 and produces a disparity map for the particles by using the particle image obtained through the cameras 110 , Based on an error for each of the particles calculated from the disparity map, the camera 110 may be configured to perform self-calibration.

According to various embodiments, the processor 120 acquires a plurality of images through the cameras 110 while moving the corrector at a predetermined interval, and two-dimensional image coordinates for points on the corrector in each of the obtained images. It may be configured to perform calibration for each of the cameras 110 by using a three-dimensional coordinate value matching with .

According to various embodiments, the processor 120, through the cameras 110, obtains a particle image, obtains the center of the particles in the particle image, determines a three-dimensional position for each of the particles in the particle image, Projecting each of the particles to each of the cameras 110 to determine a two-dimensional position for each of the particles, and based on the two-dimensional error between the center and the two-dimensional position, disparity for each of the cameras 110 It may be configured to create a map.

According to various embodiments, the processor 120 reconstructs a two-dimensional error for each of the particles by using a mesh formed at a predetermined interval in the particle image, thereby providing a disparity map for each of the cameras 110 . may be configured to produce

According to various embodiments, the processor 120 may be configured to produce a disparity map for each of the cameras 110 by drawing a virtual particle corresponding to each of the particles based on a two-dimensional error. .

According to various embodiments, the processor 120 calculates an error for each of the particles in the particle image from the disparity map for each of the cameras 110, and uses the calculated error and the three-dimensional position, Each of the cameras 110 may be configured to perform self-calibration.

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, the devices and components described in the embodiments may include a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), and a programmable logic unit (PLU). It may be implemented using one or more general purpose or special purpose computers, such as a logic unit, microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that may include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be embodied in any type of machine, component, physical device, computer storage medium or device to be interpreted by or provide instructions or data to the processing device. have. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to various embodiments may be implemented in the form of program instructions that may be executed by various computer means and recorded in a computer-readable medium. In this case, the medium may continue to store the program executable by the computer, or may be temporarily stored for execution or download. In addition, the medium may be various recording means or storage means in the form of a single or several hardware combined, it is not limited to a medium directly connected to any computer system, and may exist distributed on a network. Examples of the medium include a hard disk, a magnetic medium such as a floppy disk and a magnetic tape, an optical recording medium such as CD-ROM and DVD, a magneto-optical medium such as a floppy disk, and those configured to store program instructions, including ROM, RAM, flash memory, and the like. In addition, examples of other media may include recording media or storage media managed by an app store that distributes applications, sites that supply or distribute other various software, and servers.

Various embodiments of this document and terms used therein are not intended to limit the technology described in this document to a specific embodiment, but it should be understood to include various modifications, equivalents, and/or substitutions of the embodiments. In connection with the description of the drawings, like reference numerals may be used for like components. The singular expression may include the plural expression unless the context clearly dictates otherwise. In this document, expressions such as “A or B”, “at least one of A and/or B”, “A, B or C” or “at least one of A, B and/or C” refer to all of the items listed together. Possible combinations may be included. Expressions such as "first", "second", "first" or "second" can modify the corresponding elements regardless of order or importance, and are only used to distinguish one element from another. It does not limit the corresponding components. When an (eg, first) component is referred to as being “(functionally or communicatively) connected” or “connected” to another (eg, second) component, that component is It may be directly connected to the component, or may be connected through another component (eg, a third component).

As used herein, the term “module” includes a unit composed of hardware, software, or firmware, and may be used interchangeably with terms such as, for example, logic, logic block, component, or circuit. A module may be an integrally formed part or a minimum unit or a part of one or more functions. For example, the module may be configured as an application-specific integrated circuit (ASIC).

According to various embodiments, each component (eg, a module or a program) of the described components may include a singular or a plurality of entities. According to various embodiments, one or more components or steps among the above-described corresponding components may be omitted, or one or more other components or steps may be added. Alternatively or additionally, a plurality of components (eg, a module or a program) may be integrated into one component. In this case, the integrated component may perform one or more functions of each component of the plurality of components identically or similarly to those performed by the corresponding component among the plurality of components prior to integration. According to various embodiments, steps performed by a module, program, or other component are executed sequentially, in parallel, iteratively, or heuristically, or one or more of the steps are executed in a different order, omitted, or , or one or more other steps may be added.

Claims (12)

A method for camera self-calibration of a tomographic particle image velocimetry (PIV) system, the method comprising:
performing calibration on cameras for tomographic flow field metrology;
preparing a disparity map for the particles by using the particle images obtained through the cameras; and
performing self-calibration on the cameras based on the error for each of the particles calculated from the disparity map
containing,
How to self-calibrate the camera.
The method of claim 1,
Performing the calibration on the cameras comprises:
acquiring a plurality of images through the cameras while moving the corrector at predetermined intervals; and
performing calibration for each of the cameras by using a three-dimensional coordinate value that matches two-dimensional image coordinates for points on the calibrator in each of the obtained images
containing,
How to self-calibrate the camera.
The method of claim 1,
The step of producing the disparity map,
acquiring the particle image through the cameras;
obtaining the centers of the particles from the particle image;
determining a three-dimensional position for each of the particles in the particle image;
projecting each of the particles onto each of the cameras to determine a two-dimensional position for each of the particles; and
producing a disparity map for each of the cameras based on a two-dimensional error between the center and the two-dimensional position
containing,
How to self-calibrate the camera.
4. The method of claim 3,
The step of producing a disparity map for each of the cameras,
By reconstructing the two-dimensional error for each of the particles using a mesh formed at a predetermined interval in the particle image, producing a disparity map for each of the cameras,
How to self-calibrate the camera.
5. The method of claim 4,
The step of producing a disparity map for each of the cameras,
Based on the two-dimensional error, by drawing a virtual particle corresponding to each of the particles, producing a disparity map for each of the cameras,
How to self-calibrate the camera.
4. The method of claim 3,
Performing the self-calibration for the cameras comprises:
calculating an error for each of the particles in the particle image from a disparity map for each of the cameras; and
performing self-calibration for each of the cameras by using the calculated error and the three-dimensional position
containing,
How to self-calibrate the camera.
A tomographic flow field measurement system for camera self-calibration, comprising:
a plurality of cameras for tomographic flow field metrology; and
at least one processor connected to the cameras
including,
The processor is
Calibration is performed on the cameras;
Using the particle image obtained through the cameras to produce a disparity map for the particles,
configured to perform self-calibration on the cameras based on the error for each of the particles calculated from the disparity map,
Tomographic flow field metrology system.
8. The method of claim 7,
The processor is
While moving the corrector at a predetermined interval, acquiring a plurality of images through the cameras,
configured to perform calibration for each of the cameras using a three-dimensional coordinate value that matches two-dimensional image coordinates for points on the calibrator in each of the obtained images,
Tomographic flow field metrology system.
8. The method of claim 7,
The processor is
Acquire the particle image through the cameras,
Finding the center of the particles in the particle image,
determining a three-dimensional position for each of the particles in the particle image,
projecting each of the particles onto each of the cameras to determine a two-dimensional position for each of the particles;
configured to produce a disparity map for each of the cameras based on a two-dimensional error between the center and the two-dimensional position.
Tomographic flow field metrology system.
10. The method of claim 9,
The processor is
configured to produce a disparity map for each of the cameras by reconstructing the two-dimensional error for each of the particles using a mesh formed at a predetermined interval in the particle image,
Tomographic flow field metrology system.
11. The method of claim 10,
The processor is
configured to produce a disparity map for each of the cameras by drawing a virtual particle corresponding to each of the particles based on the two-dimensional error,
Tomographic flow field metrology system.
10. The method of claim 9,
The processor is
calculating an error for each of the particles in the particle image from the disparity map for each of the cameras,
configured to perform self-calibration for each of the cameras using the calculated error and the three-dimensional position.
Tomographic flow field metrology system.

Images