CN114608451A - Neural network-based off-plane displacement measurement method and device - Google Patents
Neural network-based off-plane displacement measurement method and device Download PDFInfo
- Publication number
- CN114608451A CN114608451A CN202210277755.0A CN202210277755A CN114608451A CN 114608451 A CN114608451 A CN 114608451A CN 202210277755 A CN202210277755 A CN 202210277755A CN 114608451 A CN114608451 A CN 114608451A
- Authority
- CN
- China
- Prior art keywords
- displacement
- neural network
- plane displacement
- layer
- reflecting surface
- 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.)
- Pending
Links
- 238000006073 displacement reaction Methods 0.000 title claims abstract description 194
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 53
- 238000000691 measurement method Methods 0.000 title claims description 14
- 238000005259 measurement Methods 0.000 claims abstract description 57
- 238000002474 experimental method Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 36
- 239000011159 matrix material Substances 0.000 claims description 33
- 238000013519 translation Methods 0.000 claims description 28
- 210000002569 neuron Anatomy 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 18
- 238000011176 pooling Methods 0.000 claims description 16
- 238000013527 convolutional neural network Methods 0.000 claims description 15
- 238000013507 mapping Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 239000000919 ceramic Substances 0.000 claims description 4
- 238000012545 processing Methods 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 description 11
- 230000000875 corresponding effect Effects 0.000 description 11
- 230000003287 optical effect Effects 0.000 description 10
- 230000008859 change Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000033001 locomotion Effects 0.000 description 6
- 239000005337 ground glass Substances 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 230000005684 electric field Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- CPBQJMYROZQQJC-UHFFFAOYSA-N helium neon Chemical compound [He].[Ne] CPBQJMYROZQQJC-UHFFFAOYSA-N 0.000 description 3
- 230000001960 triggered effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000005305 interferometry Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013401 experimental design Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000004556 laser interferometry Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000005428 wave function Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Instruments For Measurement Of Length By Optical Means (AREA)
Abstract
The invention discloses an out-of-plane displacement measuring method and device based on a neural network, wherein the method comprises the following steps: acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes; constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflection surface interference fringes and the diffuse reflection interference fringes; performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and translating a displacement end to obtain a fringe image; according to the stripe image, identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network; and determining the out-of-plane displacement direction by adopting a convolution neural network pair according to the fringe image. The invention can improve the processing efficiency and reduce the labor cost, and can be widely applied to the technical field of out-of-plane displacement measurement.
Description
Technical Field
The invention relates to the technical field of out-of-plane displacement measurement, in particular to a neural network-based out-of-plane displacement measurement method and device.
Background
Scientific research and rapid development of engineering technology put higher requirements on high resolution, wide range and real-time performance of the out-of-plane displacement measurement. Among the non-contact measurement methods, laser interferometry with high sensitivity is widely used. Fringe segmentation and counting methods, phase shifting methods, time phase estimation, fourier transform methods are often used for fringe pattern analysis for off-plane measurements. The stripe subdivision technology in the laboratory environment achieves the measurement accuracy of one percent of wavelength magnitude, but the measuring range is influenced by speckle decorrelation. The related art provides a vibration measurement method based on an interference fringe subdivision technology, and measurement accuracy of 1/400 wavelength is achieved. The time series fringe pattern phase extraction method achieves a measurement range of hundreds of micrometers, but the real-time performance is not high. The resolution of the measuring system based on laser feedback interference is 0.51nm, the measuring range is 850 mu m, and the positioning precision of 2 minutes is 5 nm. It follows that achieving high resolution, large range, real-time out-of-plane displacement measurements simultaneously remains a challenge.
With the wide application and rapid development of the out-of-plane displacement measurement, an effective, high-resolution, large-range and real-time out-of-plane displacement measurement method is increasingly needed, and the existing out-of-plane displacement measurement method has various defects, which are specifically shown as follows:
(1) it is difficult to meet the requirements of high resolution, real-time performance and large range.
(2) The level of high machine processing is not achieved, the efficiency is low, and a large amount of manual participation is required.
Disclosure of Invention
In view of this, the embodiments of the present invention provide an off-plane displacement measurement method and apparatus based on a neural network, which is efficient and low in labor cost.
One aspect of the present invention provides an out-of-plane displacement measurement method based on a neural network, including:
acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes;
constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflection surface interference fringes and the diffuse reflection interference fringes;
performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and translating a displacement end to obtain a fringe image;
according to the stripe image, identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network;
and determining the out-of-plane displacement direction by adopting a convolution neural network pair according to the fringe image.
Optionally, the single-point out-of-plane displacement measurement system includes a michelson interferometer, a piezoelectric ceramic nano translation stage, a CCD camera, and a computing device;
in the single-point off-plane displacement measurement system, emitted laser is divided into two paths through a beam splitter prism, a displacement end light path returns through a first reflector, and a reference end light path returns through a second reflector;
wherein the first mirror and the second mirror are coupled on the piezoceramic nano translation stage.
Optionally, the performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, translating the displacement end, and acquiring a fringe image includes:
taking the normalized convolution value of the interferogram subarea at any moment and the interferogram subarea at the initial moment as a reference value of the similarity degree;
and tracking the displacement according to the nonlinear mapping relation between the normalized convolution value and the displacement to obtain the displacement of the measured object at any moment.
Optionally, the performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, translating the displacement end, and acquiring a fringe image further includes:
in one period, dividing the fringe pattern into two states according to the gray matrix of the central sub-area of the fringe pattern, moving the fringe pattern leftwards for a half period and moving rightwards for a half period, and converting the problem of judging the displacement direction into the problem of image identification.
Optionally, the identifying, according to the fringe image, the magnitude of the out-of-plane displacement in one period by using a back propagation neural network includes:
adopting a five-layer BP neural network to approximate the nonlinear mapping relation between the normalized convolution value and the displacement;
calculating the normalized convolution value of the interferogram at each moment and the interferogram at the initial moment, taking the normalized convolution value as an input value of a BP neural network, and taking the corresponding known displacement as an output value of the network for network training; wherein the learning rate is set to be 0.001 in the training process;
wherein, the input layer of the five-layer BP neural network comprises a neuron corresponding to the input normalized convolution value; the neuron numbers of the three hidden layers of the five-layer BP neural network are respectively 100, 200 and 200; the output layer of the five-layer BP neural network comprises a neuron corresponding to the output displacement; any two layers are completely connected, the initial value of the weight of the connection parameter is a random number distributed in a Gaussian way, and the initial value of the bias term is a constant.
Optionally, the convolutional neural network includes three convolutional layers, three pooling layers, a full-link layer, and a softmax function layer, and determining an out-of-plane displacement direction using the convolutional neural network pair according to the fringe image includes:
taking the interference pattern subarea as an input value of the network, when the interference pattern subarea passes through a first convolution layer, the convolution kernel size is (3, 3, 1, 32), and the image is changed into a three-dimensional matrix with the size of 600 multiplied by 32;
the three-dimensional matrix passes through a maximum pooling layer, the step length in the length and width directions is 2, and the size of the three-dimensional matrix is changed into 300 multiplied by 32;
after passing through the second convolution layer, the convolution kernel size is (3, 3, 32, 64), and the three-dimensional matrix size is changed into 300 × 300 × 64;
through a maximum pooling layer with the length and width direction step length of 2, the size of the three-dimensional matrix is changed into 150 multiplied by 64;
after passing through the third convolution layer, the convolution kernel size is (3, 3, 64, 64), and the three-dimensional matrix size becomes 300 × 300 × 64;
after passing through the maximum pooling layer, the size of the three-dimensional matrix is changed to 75 multiplied by 64;
passing through a full connection layer with the number of neurons being 1000 and a softmax function layer;
the output is probability distribution through an output layer with the neuron number of 2, and the probability distribution is used for representing the probability of the left or right direction of the out-of-plane displacement direction;
and determining the out-of-plane displacement direction according to the direction with the larger probability value in the probability distribution.
Another aspect of the embodiments of the present invention further provides an out-of-plane displacement measurement apparatus based on a neural network, including:
the first module is used for acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes;
the second module is used for constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflecting surface interference fringes and the diffuse reflection interference fringes;
the third module is used for carrying out a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and carrying out translation on a displacement end to obtain a fringe image;
the fourth module is used for identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network according to the stripe image;
and the fifth module is used for determining the out-of-plane displacement direction by adopting a convolutional neural network pair according to the fringe image.
Another aspect of the embodiments of the present invention further provides an electronic device, including a processor and a memory;
the memory is used for storing programs;
the processor executes the program to implement the method as described above.
Yet another aspect of the embodiments of the present invention provides a computer-readable storage medium, which stores a program, which is executed by a processor to implement the method as described above.
A further aspect of embodiments of the present invention also provides a computer program product comprising a computer program which, when executed by a processor, implements a method as described above.
The embodiment of the invention firstly obtains the characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes; constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflection surface interference fringes and the diffuse reflection interference fringes; performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and translating a displacement end to obtain a fringe image; according to the stripe image, identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network; and determining the out-of-plane displacement direction by adopting a convolution neural network pair according to the fringe image. The invention can improve the processing efficiency and reduce the labor cost.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a flowchart illustrating the overall steps provided by an embodiment of the present invention;
FIG. 2 is a schematic diagram of a diffuse reflecting surface interferometry system provided in accordance with an embodiment of the present invention;
FIG. 3 is a schematic diagram of a single-point out-of-plane displacement measurement system according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating a normalized convolution value-displacement curve provided by an embodiment of the present invention;
fig. 5 is a schematic diagram of a BP network according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a CNN network according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In view of the problems in the prior art, an aspect of the present invention provides an out-of-plane displacement measurement method based on a neural network, as shown in fig. 1, the method includes the following steps:
acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes;
constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflection surface interference fringes and the diffuse reflection interference fringes;
performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and translating a displacement end to obtain a fringe image;
according to the stripe image, identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network;
and determining the out-of-plane displacement direction by adopting a convolution neural network pair according to the fringe image.
Optionally, the single-point out-of-plane displacement measurement system includes a michelson interferometer, a piezoelectric ceramic nano translation stage, a CCD camera, and a computing device;
in the single-point off-plane displacement measurement system, emitted laser is divided into two paths through a beam splitter prism, a displacement end light path returns through a first reflector, and a reference end light path returns through a second reflector;
wherein the first mirror and the second mirror are coupled on the piezoceramic nano translation stage.
Optionally, the performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, translating the displacement end, and acquiring a fringe image includes:
taking the normalized convolution value of the interferogram subarea at any moment and the interferogram subarea at the initial moment as a reference value of the similarity degree;
and tracking the displacement according to the nonlinear mapping relation between the normalized convolution value and the displacement to obtain the displacement of the measured object at any moment.
Optionally, the performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, translating the displacement end, and acquiring a fringe image further includes:
in one period, dividing the fringe pattern into two states according to the gray matrix of the central sub-area of the fringe pattern, moving the fringe pattern leftwards for a half period and moving rightwards for a half period, and converting the problem of judging the displacement direction into the problem of image identification.
Optionally, the identifying, according to the fringe image, the magnitude of the out-of-plane displacement in one period by using a back propagation neural network includes:
adopting a five-layer BP neural network to approximate the nonlinear mapping relation between the normalized convolution value and the displacement;
calculating the normalized convolution value of the interferogram at each moment and the interferogram at the initial moment, taking the normalized convolution value as an input value of a BP neural network, and taking the corresponding known displacement as an output value of the network for network training; wherein the learning rate is set to be 0.001 in the training process;
wherein, the input layer of the five-layer BP neural network comprises a neuron corresponding to the input normalized convolution value; the neuron numbers of the three hidden layers of the five-layer BP neural network are respectively 100, 200 and 200; the output layer of the five-layer BP neural network comprises a neuron corresponding to the output displacement; any two layers are completely connected, the initial value of the weight of the connection parameter is a random number distributed in a Gaussian way, and the initial value of the bias term is a constant.
Optionally, the convolutional neural network includes three convolutional layers, three pooling layers, a full-link layer, and a softmax function layer, and determining an out-of-plane displacement direction using the convolutional neural network pair according to the fringe image includes:
taking the interference pattern subarea as an input value of the network, when the interference pattern subarea passes through a first convolution layer, the convolution kernel size is (3, 3, 1, 32), and the image is changed into a three-dimensional matrix with the size of 600 multiplied by 32;
the three-dimensional matrix passes through a maximum pooling layer, the step length in the length and width directions is 2, and the size of the three-dimensional matrix is changed into 300 multiplied by 32;
after passing through the second convolution layer, the convolution kernel size is (3, 3, 32, 64), and the three-dimensional matrix size is changed into 300 × 300 × 64;
through a maximum pooling layer with the length and width direction step length of 2, the size of the three-dimensional matrix is changed into 150 multiplied by 64;
after passing through the third convolution layer, the convolution kernel size is (3, 3, 64, 64), and the three-dimensional matrix size becomes 300 × 300 × 64;
after passing through the maximum pooling layer, the size of the three-dimensional matrix is changed to 75 multiplied by 64;
passing through a full connection layer with the number of neurons being 1000 and a softmax function layer;
the output is probability distribution through an output layer with the neuron number of 2, and the probability distribution is used for representing the probability of the left or right direction of the out-of-plane displacement direction;
and determining the out-of-plane displacement direction according to the direction with the larger probability value in the probability distribution.
Another aspect of the embodiments of the present invention further provides an out-of-plane displacement measurement apparatus based on a neural network, including:
the first module is used for acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes;
the second module is used for constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflecting surface interference fringes and the diffuse reflection interference fringes;
the third module is used for carrying out a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and carrying out translation on a displacement end to obtain a fringe image;
the fourth module is used for identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network according to the stripe image;
and the fifth module is used for determining the out-of-plane displacement direction by adopting a convolutional neural network pair according to the fringe image.
Another aspect of the embodiments of the present invention further provides an electronic device, including a processor and a memory;
the memory is used for storing programs;
the processor executes the program to implement the method as described above.
Yet another aspect of the embodiments of the present invention provides a computer-readable storage medium, which stores a program, which is executed by a processor to implement the method as described above.
Yet another aspect of embodiments of the present invention provides a computer program product comprising a computer program which, when executed by a processor, implements a method as described above.
The following detailed description of the specific implementation principles of the present invention is made with reference to the accompanying drawings:
the invention provides a fringe movement detection and displacement measurement method based on a neural network, which is applied to various off-plane displacement measurement works, and the specific implementation principle can be divided into the following steps:
a. according to the optical principle, the characteristics of light intensity and the like of the interference fringes of the reflecting surface and the diffuse reflection interference fringes are combed so as to facilitate subsequent experimental design and implementation.
b. An off-plane displacement measurement system is designed, and experimental procedures are respectively designed for reflecting surface interference and diffuse reflecting surface interference.
c. An experiment is carried out, after the displacement end M2 is translated, the moved image of the interference fringe and the central sub-area gray matrix are acquired, and the normalized convolution value Conv of the acquired interference pattern sub-area at any moment and the initial moment (600 × 600 pixels at the center of the image) is used as a reference value of the similarity degree.
d. The magnitude of the out-of-plane displacement within a period is identified using a back-propagation neural network (BP).
e. The judgment of the displacement direction is regarded as a binary problem in image recognition, and a Convolutional Neural Network (CNN) is adopted to learn and judge the displacement direction.
In the step (a), the optical principles and characteristics of the reflective surface interference fringes and the diffuse reflection interference fringes are combed. The method comprises the following specific steps:
1) principle of interference fringe formation of reflecting surface
When the two light waves meeting the conditions of same frequency, same vibration direction and stable phase difference are superposed in space, the light intensity is re-stabilized and unevenly distributed, and the phenomenon of alternate bright and dark or colored fringes is called interference. Light waves, which are electromagnetic waves, can be described as:
here, ,
is the vector of the electric field and,
is the complex amplitude of the electric field;
is the vector of the magnetic field and,
is the complex amplitude of the magnetic field, w,
angular velocity and initial phase are represented, respectively, and p represents an arbitrary position in space. The optical disturbance referred to herein is affected by the electric field in the light wave, so that the vibration of the electric field vector is discussed, i.e. the optical vector at a certain time is only represented as:
when light propagates in a certain direction, the scalar wave function of its planar monochromatic lightwave can be expressed as:
wherein r, k represent propagation distance and wave number respectively,
is the initial phase. When two light waves with the same vibration direction propagate in space:
expressed as follows by a complex amplitude method:
when the encounter superposition occurs, the intensity of the superposed light is as follows:
order to
Represents the phase difference of two beams at point p, according to
The light intensity can be converted into:
I(P)＝E10 2(p)+E20 2(p)+2E10E20cosθ(p)
wherein E is10 2(p)，E20 2(p) represents the light intensity of the two light waves at the point p, respectively. The superposition of two coherent beams thus causes an increase of the term 2 instead of a simple addition of the light intensities10E20cos θ (p), where θ (p) is the propagation distance from the point source and the initial positionAre related.
At this time when
θ(p)＝2nπ(n＝0，±1，±2…)
Maximum is got to the light intensity, and two vector vibrations paces are unanimous, and the complete stack strengthens each other, and intensity is:
I(P)＝E10 2(p)+E20 2(p)+2E10E20＝(E10+E20)2
when in
θ(p)＝(2n+1)π(n＝0，±1，±2…)
The light intensity takes minimum value, and the vibration cadence is opposite totally, weakens each other, and intensity this moment is:
I(P)＝E10 2(p)+E20 2(p)-2E10E20＝(E10-E20)2
therefore, the light intensity of the space point of the overlapping region can be obtained through calculation. Two beams of light meeting the coherence condition at the receiving end interfere to form interference fringes with alternate light and shade, and the interference fringes are recorded by the CCD. The shape of the interference fringes, i.e. the intensity of light at each point in the overlapping area, is determined by the respective optical path difference.
2) Principle of formation of diffuse reflection interference fringes
Considering that the measured object is a diffuse reflection surface, the M2 and M3 are replaced by ground glass, and other optical path elements are not changed, as shown in fig. 2, the CCD position is adjusted, and speckle fields on the surfaces of M2 and M3 are imaged at the same time, and at this time, the diffuse reflection speckle fields are mutually superposed. LASER in fig. 2 represents a LASER; dis represents the out-of-plane displacement to be solved; m2(Object) represents ground glass; BS represents a semi-transparent semi-reflecting mirror; SF stands for spatial filter; LENS stands for prism; compensation stands for compensating end displacement; m3 represents ground glass; CCD stands for camera;
the laser light irradiates on the surface of the diffuse reflection object, and the randomly distributed bright and dark spots formed on the surface and in the front space due to interference are called laser speckles, which can be described as follows:
wherein the sum of xi, A,
respectively representing complex amplitude, amplitude and phase. Xi can also be expressed as
In an electronic speckle interference (ESPI) experiment, light and dark interference fringes can be obtained by subtracting speckle fields at different time points, and the mathematical expression of the interference fringes at the time point is as follows:
wherein, Ia-IbRepresenting the subtraction of the speckle field intensities at different times a, b, and z1，z2Representing the michelson interferometer ends M2, M3 (imaging at the same position on the CCD).
Due to the fact that
The following results were obtained:
when the speckle patterns are correlated at two moments in the viewing position, i.e.
The above formula can be rewritten as
Wherein
Respectively represent the phase change of the speckle interferogram at different moments (before and after displacement),
representing the difference of phase change of the interference speckle pattern, which is a slow-varying term, and
is a fast varying term and therefore the above equation indicates that the speckle (fast varying term) is modulated by the fringes (slow varying term).
When the pixel point phase satisfies
(m ═ 0, ± 1, ± 2 …); i.e. the phase change caused by the displacement is exactly an integer multiple of 2 pi
Therefore, the light intensity of these pixels is not changed at different time (before and after displacement), which is called dark point.
From the above analysis, when no displacement occurs, the subtracted speckle pattern will be all dark spots, and when displacement occurs, the light intensity will change, and spots with alternate light and dark appear. However, it is difficult to extract the accurate displacement size and direction from the interference field with changing brightness, so carrier fringes are introduced here, the size and direction of displacement are judged by the movement of the carrier fringes when displacement occurs, and the light intensity of the interference field after the carrier fringes are introduced is as follows:
wherein,
representing the carrier stripe phase. Therefore, carrier fringes are introduced into the diffuse reflection surface interferometry system, an image obtained by subtracting the speckle field at the moment from the original speckle field is displayed on a picture display interface of the compiling software in real time, and the light intensity of each point on the carrier fringes of the image is determined by the displacement of the corresponding pixel point.
In the step (b), a single-point out-of-plane displacement measurement system is designed, as shown in fig. 3. In fig. 3, LASER stands for LASER; SF stands for spatial filter; dis represents the out-of-plane displacement to be measured; m2(Object) represents a mirror; LENS stands for prism; BS represents a semi-transparent semi-reflecting mirror; compensation stands for compensating end displacement; m3 denotes a mirror; the CCD stands for camera.
The measuring system consists of a Michelson interferometer, a piezoelectric ceramic (PZT) nano translation table, a CCD camera and computing equipment. As shown in fig. 3, a laser beam is split into two paths by a beam splitter prism, the light path at the displacement end returns by a reflector M2, the light path at the reference end returns by a reflector M3, and a reflector M3 is coupled on a PZT nano translation stage (model: tomorrow, measurement range: 210 microns). When the measured object is considered to be a mirror surface, the reflector M2 is taken as the measured object and is coupled on a nano translation table (model number is PI: P-622.1CL, the measuring range is 800 microns, and the positioning precision is 10 nanometers). The two paths of light are combined through a beam splitter prism BS and recorded by a CCD at a receiving end.
In the step (c), an experiment is performed, after the displacement end M2 is translated, the image of the interference fringe after being moved and the central sub-area gray-scale matrix are acquired, and the normalized convolution value Conv of the sub-area of the interference pattern at any time and the sub-area of the interference pattern at the initial time (600 × 600 pixels at the center of the image) is used as a reference value of the similarity degree:
since M2 only undergoes translation in the direction of the optical axis, the fringes move, and the fringe spacing and angle do not change, i.e., only the phase changes. When no translation occurs, the fringe state is not changed, the interferogram at any moment and the interferogram at the initial moment are kept consistent, the similarity degree is maximum, and the similarity degree gradually decreases along with the gradual increase of the displacement in a half period, so that the normalized convolution value Conv of the subregion of the interferogram at any moment and the subregion of the interferogram at the initial moment (600 × 600 pixels at the center of the image) is used as a reference value of the similarity degree, and is defined as follows:
where M represents the total number of pixels in the sub-area,
the gray values of each pixel point of the interference fringe pattern in the current state and the initial state are respectively represented.
Fig. 4 shows the relationship between the normalized convolution value and the displacement, and it can be seen that the normalized convolution value has symmetry in one period, when the stripe moves for a half period, the convolution value is minimum, and when the stripe moves for a period, the convolution value returns to the maximum. Therefore, accurate tracking displacement can be given according to the nonlinear mapping relation between the normalized convolution value and the displacement, so that the displacement of the measured object M2 at any moment is obtained, and complex phase calculation is avoided. Similarly, in one period, the fringe pattern is divided into two states according to the gray matrix of the central sub-area of the fringe pattern, and the displacement direction judgment problem is converted into an image identification problem in a half period of leftward movement and a half period of rightward movement.
In the step (d), the magnitude of the out-of-plane displacement within one period is identified by using a back propagation neural network (BP).
General approximation theory shows that a 3-layer perceptron neural network can approximate a multidimensional continuous nonlinear function with arbitrary precision. A five-layer BP neural network is employed herein to approximate the nonlinear mapping relationship between normalized convolution values and displacements. As shown in fig. 5, the input layer contains a neuron corresponding to the input normalized convolution value, the three hidden layers contain the neuron numbers of 100, 200 and 200, respectively, and the output layer contains a neuron corresponding to the output displacement value. The adjacent two layers are completely connected. The initial value of the connection parameter weight is a random number distributed in a Gaussian way, and the initial value of the bias term is a constant. Considering that the nature of the problem is a regression problem, the mean square error (MSE value) is taken as a loss function of the network.
For a michelson interferometer, the object to be measured is shifted by half a wavelength and the fringes are shifted by one period. The experiment adopts a helium-neon laser, the laser wavelength is 632.8nm, and therefore, the stripe moves for one period corresponding to the displacement of 316.4nm of the measured object. Within a half period, the larger the displacement, the smaller the convolution value. Moving the measured object by 10nm each time until the measured object moves by 158.2nm (1/4 wavelength), calculating the normalized convolution value of the interferogram at each moment and the interferogram at the initial moment, taking the normalized convolution value as the input value of the BP neural network, taking the corresponding known displacement as the output value of the network for network training, and setting the learning rate to be 0.001 in the training process.
In the step (e), the judgment of the displacement direction is regarded as a binary classification problem in image recognition, and a Convolutional Neural Network (CNN) is adopted to learn and judge the problem.
As shown in fig. 6, the convolutional neural network employed herein contains three convolutional layers, three pooling layers, one fully-connected layer and one softmax function layer. The interference pattern subareas (with the size of 600 × 600 pixels) are used as input values of the network, when the interference pattern subareas pass through a first convolution layer, the convolution kernel size is (3, 3, 1, 32), the image is a three-dimensional matrix with the size of 600 × 600 × 32, the length and width directions of the three-dimensional matrix pass through a maximum pooling layer are all 2, and the matrix size is 300 × 300 × 32. Then, after passing through the second convolution layer, the convolution kernel size is (3, 3, 32, 64), the image becomes a three-dimensional matrix with the size of 300 × 300 × 64, and after passing through a maximum pooling layer with the length and width direction steps of 2, the matrix size becomes 150 × 150 × 64. The convolution kernel size of the third convolution layer is (3, 3, 64, 64), the image becomes a three-dimensional matrix of 300 × 300 × 64 size, and likewise, after passing through the maximum pooling layer, the size is reduced to 75 × 75 × 64. Then, the data passes through a full connection layer with the number of the neurons being 1000 and a softmax function layer, finally passes through an output layer with the number of the neurons being 2, the output is probability distribution which respectively represents the probability size of the left and the right, and the final displacement direction is given according to the side with the large probability value. After layer-by-layer convolution, the features of the interference pattern are gradually extracted and combined, and finally the extracted features are classified according to a certain classifier rule. And identifying the displacement direction of the image by taking the cross entropy as a loss function of the network.
In summary, the invention provides a fringe motion detection method based on a neural network to realize displacement compensation measurement, and the out-of-plane displacement measurement with high resolution, wide range and real-time performance is realized. Compared with the traditional out-of-plane displacement measurement method, the method can give consideration to multiple indexes such as high resolution, large range and real-time performance, and meanwhile, machine processing is carried out through a neural network method, so that the overall measurement efficiency is improved.
In order to verify the effectiveness of the tracking measurement algorithm, the embodiment of the invention respectively performs a reflecting surface interference experiment and a diffuse reflecting surface interference experiment.
(1) Interference test of reflecting surface
In this embodiment, a reflective surface interference experiment system constructed based on michelson interference is adopted, a helium-neon laser (HNL210L) is adopted as a light source, a beam of laser is emitted from the helium-neon laser, expanded into a group of parallel light after passing through a spatial filter, divided into two beams through a half-transmitting and half-reflecting mirror, and converged on a CCD target surface after passing through a reflector respectively. The off-plane displacement of the measured end is provided by the nano translation stage and applied to the measured end reflecting mirror, and the compensation displacement of the compensation end is provided by the PZT nano translation stage and applied to the compensation end reflecting mirror. The CCD end digital signal is transmitted to the computing equipment in real time to monitor the moving size and direction of the interference fringes in real time, and once the tracking displacement is detected to be larger than a set threshold (10 nm in the text), a tracking instruction is triggered, so that the interference image is always kept unchanged. The CCD industrial camera used in this experiment was from Basler corporation, model number Basler-ace 1600-20gm (resolution 1600X 1200 pixels, maximum frame rate 20 frame/s). The CPU of the computer is Intel i5-4460, and the display card model is GeForce GTX 1080. The tracking algorithm comprises the judgment of the size and the direction of displacement and the sending of a command to actuate the PZT nano translation stage, which are written by adopting Python language.
Considering that the measuring range of the PZT nano translation stage at the tracking end is 210 microns, the translation displacement of the measured end is set to be 40nm each time, the process is repeated 5000 times, and the interval time is 400 ms.
(2) Interference experiment of diffuse reflection surface
This embodiment is based on diffuse reflection surface interference experimental system that michelson interfered and built, measured end and tracking end change the speculum for ground glass, adjust the focus of receiving terminal CCD camera, to ground glass surface imaging for diffuse reflection interference speckle field superposes each other, shows the image after current stack speckle field subtracts with initial stack speckle field in real time at computer software end, applys the carrier wave after, appears light and shade alternate stripe. After the displacement end translates, the carrier stripes move. And the software monitors the moving size and direction of the carrier stripes in real time, and once the tracking displacement is detected to be larger than a set threshold (10 nm in the text), a tracking instruction is triggered, so that the image is kept unchanged all the time. And similarly, the translation displacement of the measured end is set to be 40nm every time, the repetition is 5000 times, and the interval time is 400 ms.
In the experiment, it is difficult to ensure that the optical axes of the compensation end and the measured end are completely vertical, so a correction coefficient K is introduced to correct the error between the displacement of the compensation end and the displacement of the measured end. The correction coefficient K can be obtained through theoretical derivation of the geometric relationship of the arrangement positions of the experimental device or a calibration experiment. The correction coefficient K is obtained by adopting a calibration experiment, the two systems respectively carry out five groups of calibration experiments, the ratio of the known measured end displacement to the tracking displacement obtained by the calculation of the compensation end is calculated, and the average value is calculated to obtain the correction coefficient K. The K value calculated by the diffuse reflection surface interference experiment system is 0.954, and the K value of the reflection surface interference experiment system is 0.912.
The time interval of every two times of translation set by the experiment is 400ms, and the translation is repeated 5000 times. According to the experimental result, the compensation measurement displacement curve is well matched with the real displacement curve. The measurement resolution of the device depends on a pre-set displacement threshold. When the detected displacement is larger than 10nm, the PZT nano translation stage is actuated to restore the stripe pattern to the initial state. When the displacement is less than 10nm, the state is regarded as unchanged. When the fluctuation is about 10nm, the measurement resolution is corresponded, and the fluctuation is kept stable and does not increase along with the increase of the displacement. Because the compensation displacement calculation of each step is compared with the initial fringe state, the residual error of the previous step can be compensated in the next compensation tracking, and therefore, no accumulated error exists. When the error of the error displacement is maintained to be about 10nm, the effectiveness of the tracking process is verified. Errors of the reflecting surface interference experiment and the diffuse reflecting surface interference experiment are influenced by background noise, including environmental vibration, air disturbance, temperature fluctuation and random CCD noise.
Table 1 shows the data comparison of the tracking measurement displacement and the real displacement in the reflective surface interference experiment and the diffuse reflective surface interference experiment. Relative errors of less than 0.5% are seen. It can be seen that the system can achieve accurate tracking measurements for both reflective surface objects and diffusely reflective surface objects within the maximum travel of the PZT unit (210 microns for the system). If the displacement of the object continues to increase, the reflecting surface interference system is still effective, and decorrelation occurs in the diffuse reflecting surface interference system, so that a wider range of measurement can be realized through segmented measurement, but error accumulation is introduced.
TABLE 1
Table 2 shows the tracking time consumption of each step, wherein the set program is equal to 200ms, and includes the time required for the nano translation stage to move, the PZT driving time, the mode stabilization time, and the time required for image acquisition. And the sum of the average calculation time of the displacement and the average judgment time of the displacement direction is less than 200ms, so that the tracking measurement algorithm can realize real-time tracking measurement.
TABLE 2
| Time consuming (ms) | Determination of displacement direction | Calculation of displacement magnitude | Program latency |
| Diffuse reflection surface | 105 | 63 | 200 |
| Reflecting surface | 123 | 55 | 200 |
The resolution of the measuring system is determined by a set displacement threshold, when the displacement of the measured object is larger than the set threshold, the compensation condition is triggered, the stripe returns to the initial position at the moment, and the compensation displacement is recorded. However, the displacement threshold is limited by background noise, including environmental vibration, air disturbance, temperature fluctuation and random CCD noise, and excessive noise may affect the accuracy of displacement calculation and even cause displacement calculation errors. The resolution can thus be further improved by effective noise suppression.
The maximum measuring range of the system is only 210 mu m and is limited by the maximum stroke of the PZT nano translation table, so that the measuring range can be increased by replacing the nano translation table with larger stroke. The time consumed by each tracking of the system is about 400ms, wherein the time consumed by the movement of the nano translation platform, the time consumed by the calculation of the displacement size and direction, the PZT driving time, the mode stabilizing time and the image acquisition time are included. The image acquisition time can be effectively reduced by using a higher frame rate (20 frames per second is used here) CCD camera. The fringe settling time can be reduced by using PZT devices with faster drive response. By improving the hardware configuration of the computer, including a CPU and a GPU, the calculation time can be effectively reduced.
The compensation algorithm realizes the out-of-plane displacement measurement by accumulating each compensation displacement, processes the interference fringe pattern by utilizing the neural network, and outputs the magnitude and the direction of the compensation displacement, thereby avoiding accurate phase calculation and overcoming the influence of noise to a certain extent.
In alternative embodiments, the functions/acts noted in the block diagrams may occur out of the order noted in the operational illustrations. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality/acts involved. Furthermore, the embodiments presented and described in the flow charts of the present invention are provided by way of example in order to provide a more thorough understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed, and in which sub-operations described as part of larger operations are performed independently.
Furthermore, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the described functions and/or features may be integrated in a single physical device and/or software module, or one or more functions and/or features may be implemented in a separate physical device or software module. It will also be appreciated that a detailed discussion of the actual implementation of each module is not necessary for an understanding of the present invention. Rather, the actual implementation of the various functional modules in the apparatus disclosed herein will be understood within the ordinary skill of an engineer, given the nature, function, and internal relationship of the modules. Accordingly, those skilled in the art can, using ordinary skill, practice the invention as set forth in the claims without undue experimentation. It is also to be understood that the specific concepts disclosed are merely illustrative of and not intended to limit the scope of the invention, which is defined by the appended claims and their full scope of equivalents.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. An out-of-plane displacement measurement method based on a neural network is characterized by comprising the following steps:
acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes;
constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflection surface interference fringes and the diffuse reflection interference fringes;
performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and translating a displacement end to obtain a fringe image;
according to the stripe image, identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network;
and determining the out-of-plane displacement direction by adopting a convolution neural network pair according to the fringe image.
2. The method for measuring the out-of-plane displacement based on the neural network is characterized in that the single-point out-of-plane displacement measuring system comprises a Michelson interferometer, a piezoelectric ceramic nano translation stage, a CCD camera and a computing device;
in the single-point off-plane displacement measurement system, emitted laser is divided into two paths through a beam splitter prism, a displacement end light path returns through a first reflector, and a reference end light path returns through a second reflector;
wherein the first mirror and the second mirror are coupled on the piezoceramic nano-translation stage.
3. The method according to claim 1, wherein the step of performing a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point out-of-plane displacement measurement system, performing translation on a displacement end, and acquiring a fringe image comprises:
taking the normalized convolution value of the interferogram subarea at any moment and the interferogram subarea at the initial moment as a reference value of the similarity degree;
and tracking the displacement according to the nonlinear mapping relation between the normalized convolution value and the displacement to obtain the displacement of the measured object at any moment.
4. The method according to claim 3, wherein the off-plane displacement measurement method based on the neural network is characterized in that a reflecting surface interference experiment and a diffuse reflecting surface interference experiment are performed according to the single-point off-plane displacement measurement system, a displacement end is translated, and a fringe image is obtained, and the method further comprises:
in one period, dividing the fringe pattern into two states according to the gray matrix of the central sub-area of the fringe pattern, moving the fringe pattern leftwards for a half period and moving rightwards for a half period, and converting the problem of judging the displacement direction into the problem of image identification.
5. The method according to claim 1, wherein the identifying the magnitude of the out-of-plane displacement in one period by using a back propagation neural network according to the fringe image comprises:
adopting a five-layer BP neural network to approximate the nonlinear mapping relation between the normalized convolution value and the displacement;
calculating the normalized convolution value of the interferogram at each moment and the interferogram at the initial moment, taking the normalized convolution value as an input value of a BP neural network, and taking the corresponding known displacement as an output value of the network for network training; wherein the learning rate is set to be 0.001 in the training process;
wherein, the input layer of the five-layer BP neural network comprises a neuron corresponding to the input normalized convolution value; the neuron numbers of the three hidden layers of the five-layer BP neural network are respectively 100, 200 and 200; the output layer of the five-layer BP neural network comprises a neuron corresponding to the output displacement; any two layers are completely connected, the initial value of the weight of the connection parameter is a random number distributed in a Gaussian way, and the initial value of the bias term is a constant.
6. The method according to claim 1, wherein the convolutional neural network comprises three convolutional layers, three pooling layers, a full connection layer and a softmax function layer, and the determining the direction of the out-of-plane displacement by using the convolutional neural network pair according to the fringe image comprises:
taking the interference image subarea as the input value of the network, when the interference image passes through a first convolution layer, the convolution kernel size is (3, 3, 1, 32), and the image becomes a three-dimensional matrix with the size of 600 multiplied by 32;
the three-dimensional matrix passes through a maximum pooling layer, the step length in the length and width directions is 2, and the size of the three-dimensional matrix is changed into 300 multiplied by 32;
after passing through the second convolution layer, the convolution kernel size is (3, 3, 32, 64), and the three-dimensional matrix size is changed into 300 × 300 × 64;
through a maximum pooling layer with the length and width direction step length of 2, the size of the three-dimensional matrix is changed into 150 multiplied by 64;
after passing through the third convolution layer, the convolution kernel size is (3, 3, 64, 64), and the three-dimensional matrix size becomes 300 × 300 × 64;
after passing through the maximum pooling layer, the size of the three-dimensional matrix is changed to 75 multiplied by 64;
passing through a full connection layer with the number of neurons being 1000 and a softmax function layer;
the output is probability distribution through an output layer with the neuron number of 2, and the probability distribution is used for representing the probability of the left or right direction of the out-of-plane displacement direction;
and determining the out-of-plane displacement direction according to the direction with the larger probability value in the probability distribution.
7. An out-of-plane displacement measuring device based on a neural network, comprising:
the first module is used for acquiring characteristic information of the interference fringes of the reflecting surface and the diffuse reflection interference fringes;
the second module is used for constructing a single-point off-plane displacement measurement system according to the characteristic information of the reflecting surface interference fringes and the diffuse reflection interference fringes;
the third module is used for carrying out a reflecting surface interference experiment and a diffuse reflecting surface interference experiment according to the single-point off-plane displacement measurement system, and carrying out translation on a displacement end to obtain a fringe image;
the fourth module is used for identifying the size of the out-of-plane displacement in one period by adopting a back propagation neural network according to the stripe image;
and the fifth module is used for determining the out-of-plane displacement direction by adopting a convolutional neural network pair according to the fringe image.
8. An electronic device comprising a processor and a memory;
the memory is used for storing programs;
the processor executing the program realizes the method of any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that the storage medium stores a program, which is executed by a processor to implement the method according to any one of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program realizes the method of any of claims 1 to 6 when executed by a processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210277755.0A CN114608451A (en) | 2022-03-21 | 2022-03-21 | Neural network-based off-plane displacement measurement method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210277755.0A CN114608451A (en) | 2022-03-21 | 2022-03-21 | Neural network-based off-plane displacement measurement method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114608451A true CN114608451A (en) | 2022-06-10 |
Family
ID=81864474
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210277755.0A Pending CN114608451A (en) | 2022-03-21 | 2022-03-21 | Neural network-based off-plane displacement measurement method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114608451A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112505148A (en) * | 2020-12-13 | 2021-03-16 | 河南省科学院应用物理研究所有限公司 | Active safety detection system for service state of glass curtain wall based on intelligent vision and big data |
| CN113375582A (en) * | 2021-05-13 | 2021-09-10 | 中国电子科技集团公司第三十八研究所 | Fuzzy neural network-based real-time monitoring method and system for flexible antenna array surface |
-
2022
- 2022-03-21 CN CN202210277755.0A patent/CN114608451A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112505148A (en) * | 2020-12-13 | 2021-03-16 | 河南省科学院应用物理研究所有限公司 | Active safety detection system for service state of glass curtain wall based on intelligent vision and big data |
| CN113375582A (en) * | 2021-05-13 | 2021-09-10 | 中国电子科技集团公司第三十八研究所 | Fuzzy neural network-based real-time monitoring method and system for flexible antenna array surface |
Non-Patent Citations (1)
| Title |
|---|
| 胡文欣: ""神经网络法在位移测量中的若干应用研究"", 《中国博士学位论文全文数据库 基础科学辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | A fast reconstruction method for three-dimensional shape measurement using dual-frequency grating projection and phase-to-height lookup table | |
| WO2021184686A1 (en) | Single-frame stripe analysis method based on multi-scale generative adversarial neural network | |
| CN108955575B (en) | A method of corrugated is accurately restored based on single width interference fringe | |
| Gurrola-Ramos et al. | U-Net based neural network for fringe pattern denoising | |
| Chen et al. | Method for eliminating zero spectrum in Fourier transform profilometry | |
| CN107228632A (en) | A kind of displacement field laminar analysis measurement device and method based on adding window Fourier transformation | |
| Yu et al. | Untrained deep learning-based phase retrieval for fringe projection profilometry | |
| Zhao et al. | Accurate fringe projection profilometry using instable projection light source | |
| Zhang et al. | Accurate measurements of droplet volume with coherence scanning interferometry | |
| CN114608451A (en) | Neural network-based off-plane displacement measurement method and device | |
| Ballester et al. | Single-shot synthetic wavelength imaging: Sub-mm precision ToF sensing with conventional CMOS sensors | |
| CN117629104A (en) | High-precision dynamic interferometry method and device based on deep learning | |
| Wu et al. | Precise 3-D microscopic profilometry using diffractive image microscopy and artificial neural network in single-exposure manner | |
| Boukamcha et al. | Robust technique for 3D shape reconstruction | |
| CN116625269A (en) | Absolute detection method for plane surface shape of large-caliber optical element | |
| Ganotra et al. | Object reconstruction in multilayer neural network based profilometry using grating structure comprising two regions with different spatial periods | |
| Wang et al. | High-efficiency 3D shape measurement based on redesigned Gray code and aligned phase unwrapping method | |
| Picart et al. | Basic fundamentals of digital holography | |
| Chen et al. | Error model and two-phase correction method for indirect illumination | |
| Qiao et al. | Snapshot interferometric 3D imaging by compressive sensing and deep learning | |
| Meijer et al. | Determination of the phase in the center of a circular two-beam interference pattern to determine the displacement of a rough surface | |
| JP4344849B2 (en) | Optical phase distribution measurement method | |
| Gao et al. | Image signal denoising method of grating linear displacement sensor based on NLM | |
| Pineda et al. | Toward the generation of reproducible synthetic surface data in optical metrology | |
| Qian et al. | Color deep learning profilometry for single-shot 3D shape measurement |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220610 |
|
| RJ01 | Rejection of invention patent application after publication |