CN110598834A - Binocular vision detection system structure optimization method - Google Patents
Binocular vision detection system structure optimization method Download PDFInfo
- Publication number
- CN110598834A CN110598834A CN201910886407.1A CN201910886407A CN110598834A CN 110598834 A CN110598834 A CN 110598834A CN 201910886407 A CN201910886407 A CN 201910886407A CN 110598834 A CN110598834 A CN 110598834A
- Authority
- CN
- China
- Prior art keywords
- binocular
- parameters
- optimization
- influence
- particle swarm
- 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
- 238000005457 optimization Methods 0.000 title claims abstract description 92
- 238000001514 detection method Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000002245 particle Substances 0.000 claims abstract description 97
- 238000005259 measurement Methods 0.000 claims abstract description 64
- 238000004458 analytical method Methods 0.000 claims abstract description 41
- 238000004088 simulation Methods 0.000 claims abstract description 13
- 230000008878 coupling Effects 0.000 claims description 12
- 238000010168 coupling process Methods 0.000 claims description 12
- 238000005859 coupling reaction Methods 0.000 claims description 12
- 230000002068 genetic effect Effects 0.000 claims description 12
- 230000003287 optical effect Effects 0.000 claims description 11
- 238000013461 design Methods 0.000 abstract description 15
- 230000000007 visual effect Effects 0.000 abstract description 6
- 230000001419 dependent effect Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 20
- 230000000694 effects Effects 0.000 description 15
- 150000001875 compounds Chemical class 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000012546 transfer Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000002028 premature Effects 0.000 description 2
- 238000010845 search algorithm Methods 0.000 description 2
- 241000251468 Actinopterygii Species 0.000 description 1
- 241000270298 Boidae Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000024703 flight behavior Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012916 structural analysis Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 238000002198 surface plasmon resonance spectroscopy Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a binocular vision detection system structure optimization method, which comprises the following steps: establishing a binocular detection system performance analysis model; analyzing the influence of the system structure on the measurement precision; analyzing the influence of index parameters; optimizing structural parameters; simulation analysis is carried out, a binocular system precision analysis structure model is established, the relation between the structural parameters of a binocular structure and system index parameters is obtained, single variable numerical analysis is carried out on the binocular system index parameters respectively, the influence of each structural parameter on the index parameters is determined, finally, the optimal solution of the binocular structure parameters is obtained by applying an improved particle swarm optimization algorithm under the conditions of meeting the actual industrial requirements and guaranteeing the effective visual field and resolution ratio, and through the method, the structural uncertainty caused by the original binocular system structure designed through manual experience is solved, the design data of the binocular system structure is made to be dependent, and the measurement precision of the binocular detection system is improved.
Description
Technical Field
The invention belongs to the technical field of binocular vision detection systems, and particularly relates to a structure optimization method of a binocular vision detection system.
Background
The visual detection system is a complex mechanical system, and the structural parameters of the detection system have a large influence on the performance of the binocular system. The indexes for describing the performance of the binocular system mainly comprise an effective view field, resolution and detection precision, the indexes for describing the performance of the binocular system mainly comprise the effective view field, the resolution and the detection precision, the influence parameters of the effective view field, the resolution and the detection precision comprise a binocular system base line distance B, a base line and optical axis included angle alpha and a camera focal length f, the coupling performance among all the parameters is high, the design difficulty of the binocular detection system with high precision, large view field and high resolution is increased, along with the development of scientific technology, the binocular detection becomes a main technical means to be applied to hotspot tasks such as non-cooperative target on-orbit capture, space garbage removal, trajectory planning and the like, and the space target is far in detection distance and greatly influenced by external factors, so that the improvement of the measurement precision of the binocular detection.
In the aspect of structure optimization intelligent algorithm, Peng D M and Fairfield C A propose a structure optimization algorithm combining a finite element analysis method and a genetic algorithm to optimize the arch bridge structure. The design variables to be optimized are discretized and the discretized variables are analyzed using a hybrid genetic algorithm. An Interactive Search Algorithm (ISA) is proposed by Xiao R B and Tao Z W, two different designed search schemes are designed, and the competitiveness of the algorithm is improved. The algorithm combining the Xiao R B, Tao Z W Whale Optimization Algorithm (WOA) and the differential evolution (de) improves the convergence speed and avoids local optimization. Javaid S, Ali, Mushtaq N provide free flight constraints for the crow search algorithm, and a personal ceiling strategy, thereby eliminating unnecessary structural analysis. Yashar D a, Wojtusiak J proposes a new integer permutation-based genetic algorithm (IPGA) that employs a constraint control ordering technique to handle manufacturability constraints during the fitness assignment phase. Khoshnoudian F, Talaii S, Fallahian M proposes a heuristic algorithm based on surface plasmon resonance particle swarm optimization (SPR-PSO) to design different fitness functions aiming at different modulation modes. SeifiH, Rezaee java a seeks to improve structural performance and design efficiency by using a transition section method and a Bidirectional Evolutionary Structure Optimization (BESO) method, so that a customized structural node can be manufactured quickly and accurately. And Ho-Huu V and Duong-Gia D solve the multi-target design optimization problem by using a multi-target evolutionary optimization algorithm, adopt a reliability analysis method and evaluate the reliability of a solution set. WangS, Wang M, Xin G U performs multi-objective structural optimization on the spoiler using a response surface optimization method. And the Jing Z, Chen J and Li solve the constraint optimization problem by adopting a Genetic Algorithm (GA), and have higher accuracy and robustness. Wang C, Yu T and Shao G combine geometric analysis (XIGA) such as expansion with chaotic particle swarm optimization algorithm, and provide a new method capable of effectively getting rid of local optimization.
However, two problems in the above analysis methods are not solved, that is, the problem of establishing a multi-parameter structure error model and how to optimize the strong coupling parameters, and these methods are prone to fall into premature and local optimal conditions when solving the multi-parameter, strong coupling and nonlinear optimization problems, resulting in low precision of final structure optimization.
Therefore, in view of the above, the present invention provides a binocular vision detection system structure optimization method by improving the existing structure and defects, so as to achieve the purpose of higher practical value.
Disclosure of Invention
In order to solve the technical problems, the invention provides a binocular vision detection system structure optimization method to solve the problems that the existing binocular vision detection system is easy to fall into precocity and local optimization when solving the problems of multi-parameter, strong coupling and nonlinear optimization.
The invention relates to a binocular vision detection system structure optimization method, which is achieved by the following specific technical means:
a binocular vision detection system structure optimization method comprises the following steps:
s1, establishing a binocular detection system performance analysis model:
the method comprises the steps of evaluating main performance indexes of the binocular detection system, namely effective view field R, horizontal resolution delta x, vertical resolution delta y and system detection precision delta, and respectively establishing performance analysis models of the binocular detection system aiming at the performance indexes;
s2, analyzing the influence of the system structure on the measurement precision:
(1) analyzing the precision of a binocular detection system;
(2) analyzing the influence of the effective field of view of the binocular system on the measurement precision;
(3) analyzing the influence of the resolution of the binocular system on the measurement precision;
s3, index parameter influence analysis:
the binocular system structural parameters comprise a base line included angle alpha, a base line distance B and a camera focal length f;
(1) analyzing the influence of the baseline included angle alpha in the structural parameters on the index parameters of the binocular system;
(2) analyzing the influence of the base line B in the structural parameters on the index parameters of the binocular system;
(3) analyzing the influence of the camera focal length f on the index parameters of the binocular system in the structural parameters;
s4, structural parameter optimization:
the method is improved on the basis of a basic particle swarm algorithm, and structural parameters are optimized by applying an improved particle swarm algorithm (IEPSO);
s5, simulation analysis:
globally searching for an optimal solution in multiple groups of feasible solutions by adopting an improved particle swarm optimization (IEPSO), and rapidly and efficiently completing the optimal solution of the binocular system structure configuration;
(1) optimizing an objective function;
the system structure is determined by an optical axis baseline included angle alpha, a focal length f and a baseline distance B, but the influence of the focal length f and the baseline distance B on the measurement precision is a pair of mutual constraint parameters which mainly influence the measurement precision and the detection depth, the optimization of the mutual constraint variables can be solved through a particle swarm algorithm, but the fitness function of the particle swarm algorithm must be related to the structure parameters and the constraint variables, at least two parameters are mutually coupled to be used as an optimization condition, and a representative effective view field, a system resolution and a binocular system measurement precision are set on the basis of the two points to serve as coupling parameters, so that the purpose of optimization is achieved;
(2) optimizing parameter setting;
obtaining an optimal solution of binocular structure parameters by applying an improved particle swarm optimization (IEPSO);
(3) simulation analysis;
respectively applying a Genetic Algorithm (GA), a particle swarm algorithm (PSO) and an improved particle swarm algorithm (IEPSO) to obtain a simulation optimization result, and analyzing; based on the analysis, an improved particle swarm optimization (IEPSO) is adopted to globally search for an optimal solution in multiple groups of feasible solutions, and the structural parameter optimization of the binocular detection system is rapidly and efficiently completed; and comparing parameters before and after optimization, and determining that the structural precision of the system obtained by applying the improved particle swarm optimization (IEPSO) is higher.
Compared with the prior art, the invention has the following beneficial effects:
establishing a binocular system precision analysis structure model, obtaining the relationship between the structural parameters of a binocular structure and system index parameters, respectively analyzing the univariate numerical values of the binocular system index parameters to determine the influence of each structural parameter on the index parameters, finally obtaining the optimal solution of the binocular structure parameters by setting an optimization objective function and applying an improved particle swarm optimization (IEPSO) under the conditions of meeting the actual industrial requirements and ensuring the effective field of view and resolution, wherein the measurement error of the binocular detection system reaches 0.01mm under the conditions of 597.05mm of effective field of view radius, 2.0738mm of horizontal resolution and 0.0960mm of vertical resolution, and by the method, the structural uncertainty caused by the original binocular system structure designed by manual experience is solved, the design of the binocular system structure is based on the design data, and the measurement precision of the binocular detection system is improved, the problems that the existing binocular detection system is easy to fall into precocity and local optimization when solving the problems of multi-parameter, strong coupling and nonlinear optimization are solved.
Drawings
FIG. 1 is a diagram of a performance analysis model of a binocular detection system according to the present invention;
fig. 2 is a schematic view of the effective field of view of the binocular system of the present invention;
FIG. 3 is a schematic diagram of coordinate transformation of a binocular system according to the present invention;
FIG. 4 is a graph of the effect of the baseline angle α on the effective field radius R of the present invention;
FIG. 5 is a graph illustrating the effect of the baseline angle α on the horizontal resolution Δ x in accordance with the present invention;
FIG. 6 is a graph of the effect of the baseline angle α on the vertical resolution Δ y of the present invention;
FIG. 7 is a graph of the effect of the included baseline angle α on measurement error in accordance with the present invention;
FIG. 8 is a graph of the effect of the baseline distance B on the effective field of view radius R in accordance with the present invention;
FIG. 9 is a graph of the effect of baseline distance B on horizontal resolution Δ x in accordance with the present invention;
FIG. 10 is a graph of the effect of baseline distance B on vertical resolution Δ y in accordance with the present invention;
FIG. 11 is a diagram showing the effect of B/Z structural parameters on measurement errors according to the present invention;
FIG. 12 is a graph of the effect of the focal length f of the camera of the present invention on the effective field radius R;
FIG. 13 is a graph of the effect of the focal length f of the camera of the present invention on the horizontal resolution Δ x;
FIG. 14 is a graph of the effect of the focal length f of the camera of the present invention on the vertical resolution Δ y;
FIG. 15 is a graph showing the effect of the included baseline angle α on the measurement error Δ in accordance with the present invention;
FIG. 16 is a flow chart of an improved particle swarm algorithm of the present invention;
FIG. 17(a) is a graph of convergence of objective functions of a genetic algorithm, a particle swarm algorithm, and an improved particle swarm algorithm;
FIG. 17(b) is a graph of optimization of measurement errors for genetic algorithm, particle swarm algorithm and modified particle swarm algorithm;
FIG. 18(a) is a graph of convergence of the focal length f of the camera of the present invention;
FIG. 18(B) is a graph of the convergence of the baseline distance B of the present invention;
FIG. 18(c) is a graph of convergence of the baseline angle α of the present invention;
FIG. 18(d) is a graph of the effective field radius R optimization of the present invention;
FIG. 18(e) is a horizontal resolution optimization plot of the present invention;
FIG. 18(f) is a graph of the vertical resolution optimization of the present invention;
FIG. 18(g) is an iterative graph of the objective function tz of the present invention;
FIG. 18(h) is a graph of the measurement error optimization of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings:
example (b):
as shown in the accompanying fig. 1 to 18 (h): the invention provides a binocular vision detection system structure optimization method, which specifically comprises the following working principles:
1. establishing a binocular detection system performance analysis model
The main performance indexes for evaluating the binocular detection system are effective view field R, horizontal resolution delta x, vertical resolution delta y and system detection precision delta. Aiming at the performance indexes, respectively establishing each performance analysis model of the binocular detection system as follows:
1.1 binocular detection System principle
The binocular system structure is a strong coupling model, and a system performance analysis model is established in the text for analyzing the influence of the system structure on the measurement accuracy, as shown in fig. 1. The origin of coordinates of the measurement system is established in the projection center oc1, xc1 of one of the cameras along the optical centers oc1, oc2, yc1 vertically downwards, zc1 and xc1, yc1 form a right-hand spiral coordinate system. The finally derived binocular system structure accuracy relationship is shown in formulas (1) to (5).
The position of the spatial point p (xc1, yc1, zc1) may be represented by α 1, ω 1, Φ 1;
since f1 ═ f2 ═ f, we can:
in the formula (I), the compound is shown in the specification,
xc1, yc1, zc1, which is the position coordinates of any point p (xc1, yc1, zc1) in space;
b represents a base line and is the length of the line from the optical centers oc1 to oc 2;
ω 1, ω 2 represents the horizontal projection angle, which is the angle between the projection of a point p (xc1, yc1, zc1) in space on the xc1, zc1 plane and the optical axis respectively;
phi 1 and phi 2 represent vertical projection angles;
alpha 1 and alpha 2 are included angles between the base line B and the optical axis of the two cameras;
f1, f2 denotes the two-camera focal length;
(x1, y1), (x2, y2) represent the position coordinates of a point p (xc1, yc1, zc1) in space under the image coordinate system.
1.2 binocular detection System accuracy analysis
As can be seen from the above binocular system performance analysis model, the position information of a spatial point P is determined by factors of B, α 1, α 2, f1, f2, x1, x2, y1 and y2, i.e., P (xc1, yc1, zc1) ═ f (B, α 1, α 2, f1, f2, x1, x2, y1 and y2), the system structure accuracy for the P point is expressed as:
in the formula (I), the compound is shown in the specification,
i＝B，α1，α2，f1，f2；
x1, x2, y1 and y2 are influence factors of the measurement system;
δ i is the extraction accuracy of each influencing factor, i.e., the system-inherent error.
To simplify the research process, the influence of the inherent error of the system is not considered herein for the moment.
The measurement error transfer coefficient of the space point p coordinate obtained by equation (6) is:
in the formula (I), the compound is shown in the specification,
θ1＝ω1+α1；
θ2＝ω2+α2。
the errors in the three coordinate axis directions are:
the measurement error of the whole system is as follows:
1.3 effective field of view of binocular System
The effective field of view of the binocular system is described by the effective field of view radius, the inscribed circle of the effective field of view of fig. 2 is a circle HEFG, and the inscribed circle radius is calculated from the area of any quadrangle, which is the effective field of view radius.
As can be seen from the arbitrary quadrilateral area formula, when four sides are known,
in the formula (I), the compound is shown in the specification,
s: the area of an arbitrary quadrilateral;
r: an effective field radius;
k＝1/2(AB+BC+CD+AD)。
in the formula, θ is the camera CCD field angle, and is determined by the camera focal length f and the CCD size.
1.4 binocular System resolution
The resolution of a binocular system refers to the ability to resolve the smallest change in the x, y coordinates of a point p within the effective field of view. Assuming that we use CCD parameters of 2048 × 2048 and 5.5 μm × 5.5 μm pixels, this is defined by the resolution: the resolution for an object is the size represented by a unit pixel, i.e.:
in the formula (I), the compound is shown in the specification,
x 1: measuring the length of an object on the CCD in the x direction;
x: measuring the length of an object in the x direction under a structure precision coordinate system;
Δ x: the length of an image element in the x direction is the resolution in the x direction;
Δ ω: the size of the CCD pixel.
In the same way, the method for preparing the composite material,
as can be seen from the system structure analysis, when the binocular system is a symmetric structure (α 1 ═ α 2,), the measurement accuracy is high, and the conversion between the binocular structure model coordinate system and the image coordinate system can be seen as follows:
then the resolution in the x, y directions is expressed as:
in the formula (I), the compound is shown in the specification,
Δ x, Δ y are respectively the system resolution in the x direction and the y direction under the structure precision coordinate system;
and omega is the horizontal projection angle of the binocular vision system.
From the geometric relationship:
the solution of the omega angle in the formula is a nonlinear solution process, the equation has a plurality of solutions, and an unsatisfactory solution set can be filtered out from the actual constraint range of the omega angle in an optimization program.
2. Index parameter impact analysis
The binocular system structural parameters comprise a base line included angle alpha, a base line distance B and a camera focal length f, and in order to know the influence of the structural parameters on index parameters of the binocular system, namely effective field radius, resolution and detection errors, the three parameters are analyzed respectively, a numerical analysis method of a single variable is applied for research convenience, and therefore an experiment rule is found.
2.1 Baseline Angle α
The size of the system structure parameter alpha angle directly influences the structure of the binocular system, the larger the alpha angle, the larger the detection field range, and the farther the detection depth, so that the important significance is realized in the research of the value of the alpha angle. From the formulas (10) - (18), the effective field radius R of the binocular system is related to the system resolution and three structural parameters, for the convenience of research, a single variable numerical analysis is set, and it can be known from FIG. 4 that the alpha angle has a plurality of extreme points for R, and the maximum value in the range is too large and does not accord with the actual situation, additional constraint conditions are needed, and the extreme point which accords with the constraint conditions is selected to obtain the maximum value of R. As can be seen from fig. 5 and 6, the larger the angle α, the larger the horizontal resolution, and the vertical resolution has no linear relationship.
As shown in equations (6) to (9), the detection errors caused by α 1 and α 2 are:
the measurement error versus angle α is shown in fig. 7. As can be seen from the figure, the measurement error is the smallest when α 1 ═ α 2, that is, when the binocular system is structurally symmetric (α 1 ═ α 2), the binocular system has high measurement accuracy.
2.2 base line distance B
The change of the binocular baseline distance can cause the change of the detection depth z of the binocular system and the included angles alpha 1 and alpha 2 between the optical axis and the baseline, so the influence of the baseline distance B on index parameters is very complicated, fig. 8 is an influence graph of the baseline B on R, the longer the available baseline is, the larger the effective field radius is, fig. 9 and 10 are influence graphs of the baseline B on the resolution of the system, the longer the available baseline is, the lower the resolution of the system is, therefore, how to set the system to meet the requirements of the larger field of view and the higher resolution is very important.
To find out the influence rule of the baseline B on the detection accuracy more clearly, we assume that α 1 ═ α 2, ω 1 ═ ω 2 ═ 0,i.e. assuming that the binocular camera is a parallel binocular system. The error transfer coefficient can be expressed as:
in the formula (I), the compound is shown in the specification,
the total error of this system with respect to the measurements of B and z can be expressed as:
in the formula (I), the compound is shown in the specification,
the measurement error of the system changes with the change of k, the measurement error curve is shown in fig. 11, it can be known from the figure that when the value of k is near 1.3, the error can reach the minimum value, which explains the importance of the system structure design, but the measurement accuracy is reduced directly due to the overlarge baseline distance, the effective visual field of the binocular measurement is also greatly influenced, and the size of the baseline distance is directly limited by the weight, the volume and the system space size of the camera, therefore, the design of the system structure needs to be considered under the condition of multiple factors
2.3 focal Length f of Camera
The focal length of the camera is an important parameter of the optical lens, the size of the focal length has a large influence on the measurement accuracy, the field range and the resolution, fig. 12 is an influence curve of the focal length f on R, a plurality of extreme points exist in the value range, the design is required according to the actual situation, fig. 13 and 14 are influence curves of the focal length f on the resolution of the system, the larger the focal length f is, the lower the horizontal resolution is, and the higher the vertical resolution is.
The error transfer function of the focal length obtained from equations (6) to (9) is:
the measurement error distribution about the focal length is shown in fig. 15, it can be known from fig. 15 that the system measurement error decreases with the increase of the focal length, that is, the larger the effective focal length of the camera in the binocular vision system, the higher the measurement accuracy, but by the optical camera principle, the larger the focal length of the camera, the larger the load of the camera inevitably increases, which not only is the maneuverability poor, but also is not practical.
In summary, when the binocular system is a symmetrical structure, that is, α 1 is α 2, the measurement accuracy is high, and it can be known from the analysis of the baseline distance B that the measurement accuracy is high when the baseline distance is smaller, but the detection range is smaller when the baseline distance is smaller and the effective field radius R is smaller. From the analysis of the focal length, the longer the focal length, the higher the measurement accuracy, the higher the resolution, but the larger the focal length of the camera, the larger the load, which is not in accordance with the actual requirement, therefore, if a larger detection depth is desired and a larger detection range is desired, a large baseline is inevitably required, which will cause the reduction of the measurement accuracy, and meanwhile, the focal length of the camera is inevitably in a certain range, which cannot realize infinity, so how to design the binocular detection system structure, so that it can satisfy the requirements of the effective field of view, the resolution and the measurement accuracy under the condition of satisfying the structural design, which is important, at this time, the univariate numerical analysis can not satisfy the requirements, in order to solve the problem, the present document applies the improved particle swarm optimization method to optimize the binocular detection system structure, wherein the coupling condition of the binocular system is used as the fitness function of the algorithm, and the focal length and the baseline range are used, thereby achieving the purpose of finding the optimal solution.
3. Structural parameter optimization method
The particle swarm optimization algorithm is provided under the inspiration of behavior rules of bird swarms, fish swarms and human society, a Boids model is mainly used for simulating bird swarms flight behavior, in the model, the behavior of each individual is only related to the behavior of adjacent individuals around the individual, three principles need to be followed, namely collision avoidance, speed consistency and principle of gathering to the center, the basic idea of the most original particle swarm algorithm, namely the PSO algorithm, is to randomly initialize a group of particles without volume and mass, regard each particle as a feasible solution of the optimization problem, determine the quality of the particle by a preset fitness function, move in a feasible solution space, determine the direction and distance of the particle by a speed variable, usually follow the current optimal particle and obtain the optimal solution after generation-by-generation searching. The particle swarm algorithm adopts a random initial population, and updates the position and the speed of the particle swarm through adaptive learning, as shown in formulas (26) and (27).
In the formula (I), the compound is shown in the specification,
omega is the inertial weight;
c1, C2 are acceleration terms;
r1 and R2 are random numbers between [0,1 ];
pgt is the global optimum position;
the Pit is the historical optimal position found by the particle;
xid is the current iteration particle position;
and vidt +1 is the next iteration speed.
The particle swarm optimization reduces the iteration times of optimization by means of the characteristic that particles have memory, can find an optimal solution in a short time, but also generates premature convergence, has poor local optimization capability and is easy to fall into the problem of local optimization. The method is improved on the basis of a basic particle swarm algorithm, and the improved particle swarm algorithm is applied to optimize the structural parameters. The improved particle swarm optimization (IEPSO) algorithm provided by the method increases local-global information sharing items while maintaining the local development capability of the traditional particle swarm optimization algorithmTo improve the global exploration capability of the algorithm. And on the basis of the thought of genetic algorithm population variation, the diversity of the population is kept by adopting the principle of final elimination. The global optimization performance of the particle swarm optimization algorithm is improved by the improved method, and fig. 16 is a specific process for realizing the IEPSO algorithm.
And randomly initializing the positions and the speeds of the population particles, calculating the fitness values of the individual particles, and keeping the positions and the fitness values of the individual particles and the global optimal particles of the current iteration. And then performing particle swarm operation. Increased local-global information sharing itemsIs the signal between the locally optimal particle and the globally optimal particle obtained from the current iterationAnd the information exchange is used for balancing the exploration and development capability of the particles in the global optimization process of the algorithm. The IEPSO algorithm updates the speed and position using equations (27), (28).
The formula (29) is composed of three parts, the first part is 'inheritance' of previous speed, the second part is 'self-cognition' of the particle, and the third part is 'local information sharing'. The fourth part is "local-global information sharing".
The IEPSO algorithm is not limited to one-way communication between the global particles and the individual particles, and the added local-global information sharing itemAnd (4) information exchange between the local optimal particle found for the particle and the global optimal particle obtained by the current iteration. Updating the population speed by the formula (3), searching the whole search space by the early-stage particles of the algorithm at a higher speed, and determining the approximate range of the optimal solution, which is beneficial to global search; the search space of most particles in the later period is gradually reduced and the particles are concentrated in the neighborhood of the optimal value for deep search, so that local search is facilitated.
And continuously keeping the original speed of the particle which does not exceed the preset range after the speed is updated, and assigning the maximum value of the speed to the particle when the speed of the particle exceeds the speed boundary. And continuously keeping the original positions of the particles which do not exceed the preset range after the positions are updated, and when the particles exceed the preset range, eliminating inferior particles and supplementing new particles in the preset range to form a new population. And recalculating the fitness value of the new population, and storing the individual particles and the global optimal particle position and fitness value information obtained by current iteration.
As population diversity decreases, the algorithm easily converges around a local optimum. In order to keep the diversity of the population, the fitness value is used as an evaluation standard, particles with poor objective function values are eliminated, the particles are selected in a preset range to be supplemented into the population, the particle swarm operation is executed again, iteration is stopped when the convergence condition reaches the convergence precision, and the global optimum value is obtained.
4. Simulation analysis
And globally searching for an optimal solution in a plurality of groups of feasible solutions by adopting an IEPSO algorithm, and quickly and efficiently finishing the optimal solution of the binocular system structure configuration. The method is characterized in that a certain type of binocular system is designed as example simulation optimization, wherein B is 1000mm, alpha is 450, horizontal resolution delta x is 1.1564mm, vertical resolution delta y is 4.5505mm, f is 10mm, and detection error delta of the binocular system is 10 mm.
4.1 optimizing the objective function
The system structure is determined by an optical axis baseline included angle alpha, a focal length f and a baseline distance B, but the influence of the focal length f and the baseline distance B on the measurement precision is a pair of mutual constraint parameters which mainly influence the measurement precision and the detection depth, the optimization of the mutual constraint variables can be solved by a particle swarm algorithm, but the fitness function of the particle swarm algorithm must be related to the structure parameters and the constraint variables, at least two parameters are mutually coupled to be used as an optimization condition, a representative effective visual field is set based on the two points, the system resolution and the binocular system measurement precision are used as coupling parameters, so that the purpose of optimization is achieved, wherein the larger the effective visual field is, the larger the detection range is, the higher the resolution and the measurement precision are, the larger the baseline distance is, the larger the focal length is required by the resolution and the measurement precision is, this complies with the coupling parameter criterion. The objective of the method is to improve the measurement accuracy of a binocular system structure as much as possible, and reduce the measurement error of the system while improving the effective visual field and the system resolution as much as possible.
The optimization objective function herein based on the above analysis is:
tz＝μ1Δ+μ2Δx+μ3Δy-μ4R
in the formula (I), the compound is shown in the specification,
Δ: measuring errors of a binocular system, wherein the smaller the measuring error is, the higher the system precision is;
Δ x: binocular system horizontal direction resolution, Δ y: the vertical direction resolution of the binocular system is smaller, namely the smaller the delta x and the delta y is, the stronger the resolution capability of the system is;
r: the effective field radius of the binocular system is larger, and the detection range of the binocular system is wider when R is larger;
the weight coefficients mu 1+ mu 2+ mu 3+ mu 4 of the parameters mu 1, mu 2, mu 3 and mu 4 are 1, and the weight coefficients are set after normalization processing, namely, the optimal structure of the binocular system with the particles within the parameter range and the minimum optimized objective function is found.
4.2 optimizing parameter settings
The influence analysis of the index parameters can obtain that the detection precision of the symmetrical binocular system structure is highest, so that the binocular system structure parameters are set to be alpha 1-alpha 2-alpha in the optimization process; f1 ═ f2 ═ f; and the three parameters are structural parameters for particle swarm optimization, the specific process is to set a three-dimensional particle parameter range of alpha, f and B, the size of the optimized population in each time in the parameter range is 500, an optimization function is applied to find a global minimum value, the weight occupied by a system measurement error delta is the maximum, namely a three-variable optimal solution which enables the measurement error to be the minimum is found, and the solution is the optimal solution of the binocular system structure. Specific parameter settings are shown in table 1.
Table 1: parameter setting
4.3 simulation analysis
The simulation optimization results obtained by applying the Genetic Algorithm (GA), the particle swarm algorithm (PSO), and the improved particle swarm algorithm (IEPSO) are shown in fig. 17(a) and 17 (b).
Fig. 17(a) is a graph of convergence of objective functions of three intelligent optimization algorithms, which shows that a Genetic Algorithm (GA) and a particle swarm algorithm (PSO) converge in 25 generations, and an improved particle swarm algorithm (IEPSO) converges in 14 generations, which shows that the convergence speed of the improved particle swarm algorithm is faster, and that an objective function tz of the improved particle swarm algorithm is smallest, which shows that the optimization effect of the improved particle swarm algorithm is the best, and fig. 17(b) is a graph of measurement error optimization of the three algorithms, which shows that, with respect to a measurement error, the measurement error of the Genetic Algorithm (GA) and the measurement error of the particle swarm algorithm (PSO) can reach 0.03mm, and the measurement error of the improved particle swarm algorithm (IEPSO) can reach 0.01mm, i.e., the detection accuracy is the highest by applying the improved particle swarm algorithm.
Based on the analysis, an IEPSO algorithm is adopted to globally search for an optimal solution among a plurality of groups of feasible solutions, so that the structural parameter optimization of the binocular detection system is rapidly and efficiently completed, and the simulation result is shown in fig. 18.
Fig. 18 is a simulation analysis diagram of a system applying the improved particle swarm optimization, and fig. 18(a), fig. 18(b), fig. 18(c) are structural parameter convergence curves, the result is converged, and the optimization result has practical significance, fig. 18(d), fig. 18(e), fig. 18(f) are index parameter optimization curves, fig. 18(d) is an effective field radius, as compared with the baseline distance optimization curve diagram of fig. 18(b), the maximum effective field radius is about half of the baseline distance, fig. 18(e) and fig. 18(f) are a horizontal resolution and a vertical resolution, respectively, and the system resolution set according to the parameters can resolve at least 3.2mm moving objects. Fig. 18(g) is an iteration curve of the objective function tz, which converges around 14 generations, and has a fast operation speed and a good convergence effect, indicating that the fitness function has strong coupling, meets the convergence condition, and conforms to the particle swarm optimization standard, fig. 18(h) is a measurement error optimization diagram, from which it can be known that the measurement error can converge to 0.01mm, and compared with a pure empirical manufacturing system, the system has higher design accuracy.
Table 2: optimized pre-and post-parameter comparison
Table 2 shows that comparing the empirical setting with a set of parameters before and after the optimization by applying the particle swarm algorithm, the effective field R is increased by 98.1371mm, the horizontal resolution is decreased by 0.9174mm, the vertical resolution is increased by 4.4545mm, and the measurement error is 0.01mm from 10mm when the structural parameters (B, α, f) meet the actual industrial requirements, so that the structural accuracy of the system obtained by applying the improved particle swarm algorithm is higher when the difference range between the effective field and the resolution is smaller.
5. Summary of the invention
Establishing a binocular system precision analysis structure model to obtain the relationship between the structural parameters of the binocular structure and the system index parameters, then analyzing the single variable numerical values of the index parameters of the binocular system respectively, determining the influence of each structural parameter on the index parameters, finally obtaining the optimal solution of the binocular structural parameters by setting an optimization objective function and applying an improved particle swarm optimization algorithm (IEPSO) under the condition of meeting the actual industrial requirements and ensuring the effective field of view and resolution, under the conditions that the effective field radius is 597.05mm, the horizontal resolution is 2.0738mm, and the vertical resolution is 0.0960mm, the measurement error of the binocular detection system is 0.01mm, by the method, the structural uncertainty caused by the original binocular system structure designed by artificial experience is solved, the design of the binocular system structure can be based on, and the measurement precision of the binocular detection system is improved.
The technical solutions of the present invention or similar technical solutions designed by those skilled in the art based on the teachings of the technical solutions of the present invention are all within the scope of the present invention to achieve the above technical effects.
Claims (1)
1. A binocular vision detection system structure optimization method is characterized by comprising the following steps:
s1, establishing a binocular detection system performance analysis model:
the method comprises the steps of evaluating main performance indexes of the binocular detection system, namely effective view field, horizontal resolution, vertical resolution and system detection precision, and respectively establishing performance analysis models of the binocular detection system aiming at the performance indexes;
s2, analyzing the influence of the system structure on the measurement precision:
(1) analyzing the precision of a binocular detection system;
(2) analyzing the influence of the effective field of view of the binocular system on the measurement precision;
(3) analyzing the influence of the resolution of the binocular system on the measurement precision;
s3, index parameter influence analysis:
the binocular system structural parameters comprise a base line included angle, a base line distance and a camera focal length;
(1) analyzing the influence of the base line clamp alpha in the structural parameters on the index parameters of the binocular system;
(2) analyzing the influence of the base line in the structural parameters on the index parameters of the binocular system;
(3) analyzing the influence of the camera focal length in the structural parameters on the index parameters of the binocular system;
s4, structural parameter optimization:
the method comprises the following steps of improving on the basis of a basic particle swarm algorithm, and optimizing structural parameters by applying the improved particle swarm algorithm;
s5, simulation analysis:
globally searching for an optimal solution in a plurality of groups of feasible solutions by adopting an improved particle swarm algorithm, and quickly and efficiently finishing the optimal solution of the binocular system structure configuration;
(1) optimizing an objective function;
the system structure is determined by an optical axis baseline included angle, a focal length and a baseline distance, the influence of the focal length and the baseline distance on the measurement precision is a pair of mutual constraint parameters which mainly influence the measurement precision and the detection depth, the optimization of the mutual constraint variables is solved through a particle swarm algorithm, a fitness function of the particle swarm algorithm is related to the structure parameters and the constraint variables, at least two parameters are mutually coupled to be used as an optimization condition, and a representative effective view field, a system resolution and a binocular system measurement precision are set on the basis of the two points to serve as coupling parameters, so that the purpose of optimization is achieved;
(2) optimizing parameter setting;
obtaining an optimal solution of binocular structure parameters by applying an improved particle swarm optimization algorithm;
(3) simulation analysis;
respectively applying a genetic algorithm, a particle swarm algorithm and an improved particle swarm algorithm to obtain a simulation optimization result, and analyzing; based on the analysis, an improved particle swarm algorithm is adopted to globally search for an optimal solution in multiple groups of feasible solutions, and the structural parameter optimization of the binocular detection system is completed quickly and efficiently; and comparing parameters before and after optimization, and determining that the structural precision of the system obtained by applying the improved particle swarm optimization is higher.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910886407.1A CN110598834A (en) | 2019-09-19 | 2019-09-19 | Binocular vision detection system structure optimization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910886407.1A CN110598834A (en) | 2019-09-19 | 2019-09-19 | Binocular vision detection system structure optimization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110598834A true CN110598834A (en) | 2019-12-20 |
Family
ID=68861156
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910886407.1A Pending CN110598834A (en) | 2019-09-19 | 2019-09-19 | Binocular vision detection system structure optimization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110598834A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110929427A (en) * | 2019-12-23 | 2020-03-27 | 吉林大学 | Remote sensing satellite video imaging rapid simulation method |
| CN111274673A (en) * | 2020-01-07 | 2020-06-12 | 上海索辰信息科技有限公司 | Optical product model optimization method and system based on particle swarm optimization |
| CN114027974A (en) * | 2021-09-15 | 2022-02-11 | 苏州中科华影健康科技有限公司 | Multi-focus endoscope path planning method, device and terminal |
| CN115100365A (en) * | 2022-08-25 | 2022-09-23 | 国网天津市电力公司高压分公司 | Camera optimal baseline acquisition method based on particle swarm optimization |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740947A (en) * | 2016-02-01 | 2016-07-06 | 金陵科技学院 | Robot performance optimization method based on Pareto entropy particle swarm |
| JP2018128739A (en) * | 2017-02-06 | 2018-08-16 | キヤノン株式会社 | Image processing apparatus, image processing method, computer program and storage medium |
| CN109524693A (en) * | 2018-11-13 | 2019-03-26 | 吉林大学 | Fuel battery air feed system model predictive control method |
| CN109978957A (en) * | 2019-03-22 | 2019-07-05 | 青岛鑫慧铭视觉科技有限公司 | Biocular systems scaling method based on quantum behavior population |
-
2019
- 2019-09-19 CN CN201910886407.1A patent/CN110598834A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740947A (en) * | 2016-02-01 | 2016-07-06 | 金陵科技学院 | Robot performance optimization method based on Pareto entropy particle swarm |
| JP2018128739A (en) * | 2017-02-06 | 2018-08-16 | キヤノン株式会社 | Image processing apparatus, image processing method, computer program and storage medium |
| CN109524693A (en) * | 2018-11-13 | 2019-03-26 | 吉林大学 | Fuel battery air feed system model predictive control method |
| CN109978957A (en) * | 2019-03-22 | 2019-07-05 | 青岛鑫慧铭视觉科技有限公司 | Biocular systems scaling method based on quantum behavior population |
Non-Patent Citations (6)
| Title |
|---|
| SHUANG NING等: "Analysis and optimization of the performance parameters of noncooperative target location detection system", 《INTERNATIONAL JOURNAL FOR LIGHT AND ELECTRON OPTICS》 * |
| ZHANG ,GM等: "Camera Calibration of the Stereo-Vision System with Large Field of View Based on Parallel Particle Swarm Optimization", 《2ND INTERNATIONAL CONFERENCE AND INFORMATION SCIENCE》 * |
| 刘世晶等: "基于全视域GA-SVR模型的鱼类行为双目视觉观测系统标定", 《农业工程学报》 * |
| 宁爽: "基于双目视觉对非合作目标的定轨", 《中国优秀博硕士学位论文全文数据库（硕士）工程科技II辑》 * |
| 支帅等: "非合作目标超近距离双目测量系统误差分析", 《中国空间科学学会2013年空间光学与机电技术研讨会会议论文集》 * |
| 肖志涛等: "双目视觉系统测量精度分析", 《光电工程》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110929427A (en) * | 2019-12-23 | 2020-03-27 | 吉林大学 | Remote sensing satellite video imaging rapid simulation method |
| CN111274673A (en) * | 2020-01-07 | 2020-06-12 | 上海索辰信息科技有限公司 | Optical product model optimization method and system based on particle swarm optimization |
| CN114027974A (en) * | 2021-09-15 | 2022-02-11 | 苏州中科华影健康科技有限公司 | Multi-focus endoscope path planning method, device and terminal |
| CN114027974B (en) * | 2021-09-15 | 2023-10-13 | 苏州中科华影健康科技有限公司 | Endoscope path planning method, device and terminal for multiple lesion sites |
| CN115100365A (en) * | 2022-08-25 | 2022-09-23 | 国网天津市电力公司高压分公司 | Camera optimal baseline acquisition method based on particle swarm optimization |
| CN115100365B (en) * | 2022-08-25 | 2023-01-20 | 国网天津市电力公司高压分公司 | Camera optimal baseline acquisition method based on particle swarm optimization |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110598834A (en) | Binocular vision detection system structure optimization method | |
| CN109410321B (en) | Three-dimensional reconstruction method based on convolutional neural network | |
| CN107199569B (en) | Joint robot trajectory planning method based on joint energy balanced distribution | |
| CN107590827A (en) | A kind of indoor mobile robot vision SLAM methods based on Kinect | |
| CN104020665B (en) | Mechanical arm minimum jerk track optimizing method based on multi-objective particle swarm algorithm | |
| CN108994844A (en) | A kind of scaling method and device of sanding operation arm trick relationship | |
| CN111192364A (en) | Low-cost mobile multi-robot vision simultaneous positioning and map creating method | |
| CN112497216A (en) | Industrial robot pose precision compensation method based on deep learning | |
| CN111076724A (en) | Three-dimensional laser positioning method and system | |
| Yu et al. | A hybrid point cloud alignment method combining particle swarm optimization and iterative closest point method | |
| CN116945177B (en) | Feeding robot feeding path planning method, feeding robot feeding path planning device, electronic equipment and medium | |
| CN113910001A (en) | Numerical control machine tool space error identification method | |
| TWI625700B (en) | Redundant point detection method for point cloud data bonding | |
| CN114310872B (en) | Automatic vegetable-beating method for mechanical arm based on DGG point cloud segmentation network | |
| CN109084751B (en) | High-energy-efficiency satellite attitude determination algorithm based on box particle filter | |
| CN110849351B (en) | Method for constructing grid map by using depth camera and binocular camera | |
| Hlavac | Kinematics control of a redundant planar manipulator with a MLP neural network | |
| CN113312821A (en) | B-spline density method-based three-dimensional self-supporting structure topology optimization design method | |
| CN117075617B (en) | Robot track planning method and device, storage medium and electronic equipment | |
| CN117649495B (en) | Indoor three-dimensional point cloud map generation method and system based on point cloud descriptor matching | |
| CN113920180B (en) | Point cloud registration optimization method based on normal distribution transformation hypothesis verification | |
| CN115375902B (en) | Multi-spectral laser radar point cloud data-based over-point segmentation method | |
| CN112329167B (en) | Magnetorheological polishing removal function deduction method based on curvature and immersion depth | |
| CN118821356A (en) | Spring structure design method and system of gravity sensing unit | |
| Ghazanfari et al. | Optimal rastering orientation in freeform extrusion fabrication processes |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191220 |