A self-detection optimization method and system of a high-speed high-precision structured light measurement system

Abstract

The application belongs to the technical field of three-dimensional measurement, and particularly discloses a self-detection optimization method and system of a high-speed and high-precision structured light measurement system. The structured light measurement system is constructed and calibrated; a periodic fringe pattern is pre-projected to the surface of a to-be-measured object, a measurement light intensity function under environmental interference is acquired, and a light intensity noise model under environmental interference is constructed; according to the light intensity noise model, the environmental light frequency is solved through the light intensity function under environmental interference, and the best exposure time is obtained, and the threshold range of a saturated mask image is calculated through the light intensity function; the best exposure of the measurement system and the best fringe frequency are obtained according to environmental parameters, so as to compensate for the original measurement parameters; and the point cloud of the target object is calculated according to the compensated original measurement parameters and the light intensity function of the to-be-measured object acquired. The environmental light influence frequency is calculated through the pre-projection mode, the collection parameters of the binocular structured light system are compensated through the interference frequency, and the reconstruction effect is optimized.

Description

Technical Field

[0001] This invention belongs to the field of three-dimensional measurement technology, and more specifically, relates to a self-detection optimization method and system for a high-speed, high-precision structured light measurement system. Background Technology

[0002] Three-dimensional measurement technology plays a vital role in industrial automation, including process guidance, process status monitoring, process dimension detection and defect identification, and process quality assessment. It is an indispensable part of realizing automated intelligent manufacturing.

[0003] However, for structured light 3D measurement systems in complex ambient lighting scenarios, linear shifts in measurement results and insufficient robustness of measurement accuracy still occur. Therefore, for structured light measurement systems operating under different light field environments in the measurement of large moving components, it is still necessary to optimize the accuracy of the measurement system.

[0004] For mobile structured light measurement systems in this situation, there are several ways to optimize accuracy: reconstructing surface information with dense point clouds by using high-frequency fringe phase profilometry; and improving the surface reflection efficiency of the object under test by using special material surface spraying, thereby improving the acquisition efficiency. However, the computational and operational complexity of the above two methods are relatively high, and the improvement effect on measurement accuracy under indoor light fields is not significant.

[0005] Chinese patent CN106840036A provides a binary structured light optimization method suitable for rapid 3D topography measurement, including the following steps: camera and projector λ calibration; camera calibration and distortion correction; projection of phase-shifting grating fringes; projection of binary structured light encoding map; determination of grating wrapping phase using the phase-shifting method, and determination of grating order using the encoding fringes; determination of absolute phase, and completion of phase unwrapping; phase mapping to true height, and reconstruction of 3D contour. However, under the mixed interference of projection nonlinearity error and environmental influence, the height information mapped by phase may change, thus lacking universality. Chinese patent CN112611320A discloses a multi-frequency heterodyne surface structured light 3D reconstruction method based on gradient light intensity. The steps include: 1) building a structured light measurement system; 2) calibrating the entire structured light measurement system; 3) determining the different frequencies of m sets of structured light fringe patterns to be generated; 4) generating m sets of projection images based on gradient light intensity; 5) projecting the projection images onto the surface of the object under test, and then modulating the images by the object under test to obtain a total acquired image; 6) calculating the gradient light intensity of the total acquired image; 7) calculating the wrapping phase of each set of acquired images and obtaining the highest frequency wrapping phase; 8) calculating the integer fringe level of the highest frequency acquired image; 9) calculating the absolute phase corresponding to the highest frequency acquired image and performing 3D reconstruction. The process of solving for the absolute phase using the multi-frequency phase-shifting method results in an excessively dense point cloud on the structured light measurement surface, leading to excessive computational time costs.

[0006] Based on the above-mentioned defects and shortcomings, there is an urgent need in this field to propose a self-testing optimization method for high-speed and high-precision structured light measurement systems, which adds a compensation factor to the original four-step phase shift method, thereby avoiding the influence of the measurement environment under different ambient light conditions on the measurement results of mobile robots. Summary of the Invention

[0007] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a self-detection optimization method and system for a high-speed, high-precision structured light measurement system. Based on the principle of binocular structured light, a compensation algorithm for environmental interference is proposed, and the structured light 3D reconstruction system is optimized accordingly. By designing a pre-projection mode to solve for the ambient light frequency, the compensation is incorporated into the structured light image acquisition program, thereby improving the accuracy of the reconstruction effect of the structured light 3D measurement system under various ambient light conditions. Furthermore, by designing an algorithm to plan the reconstruction fringe frequency within an effective threshold range, the accuracy requirements are maintained while improving the structured light scanning and reconstruction operation speed with the lowest possible binocular structured light phase shift frequency.

[0008] To achieve the above objectives, according to one aspect of the present invention, a self-testing optimization method for a high-speed, high-precision structured light measurement system is proposed, characterized by comprising the following steps:

[0009] S1 constructs a structured light measurement system and calibrates the system.

[0010] S2 pre-projects a periodic stripe pattern onto the surface of the object under test, obtains the measured light intensity function under environmental interference, and constructs a light intensity noise model under environmental interference based on the measured light intensity function.

[0011] S3 calculates the ambient light frequency and the optimal exposure time based on the light intensity noise model and the light intensity function under environmental interference, and calculates the threshold range of the saturated mask image using the light intensity function.

[0012] S4 obtains the optimal exposure and optimal stripe frequency of the measurement system based on environmental parameters to compensate for the original measurement parameters;

[0013] S5 calculates the point cloud of the target object based on the compensated original measurement parameters and the collected light intensity function of the object under test.

[0014] As a further preferred embodiment, in step S1, the optical measurement system includes: a binocular camera arranged non-collinearly and a measurement object, and a fringe generating device is also arranged between the binocular cameras for periodically projecting fringe onto the measurement object.

[0015] As a further preferred option, step S2 includes the following steps:

[0016] S21 obtains the grayscale information of the pixels of the object under test by encoding the grayscale points of the image captured by the camera, and represents the grayscale information as a measurement light intensity function;

[0017] S22 constructs a light intensity noise model under the combined effects of ambient light and surface reflection from the object being measured;

[0018] S23 solves the relationship between the package phase error and the light intensity error based on the measured light intensity function and the light intensity noise model;

[0019] Based on the relationship between the packaged phase error and the light intensity error, S24 divides the phase error under the influence of multiple noises into phase error under the overall influence and phase error under the influence of periodic ambient light.

[0020] As a further preferred embodiment, in step S21, the measured light intensity function is:

[0021] I n (x,y)=A(x,y)+B(x,y)cos[φ(x,y)-2πn / N]

[0022] In the formula, A(x,y) is the background light intensity, B(x,y) is the modulation intensity, and φ(x,y) is the phase information;

[0023] Preferably, in step S22, the light intensity noise model includes:

[0024] A n =A(x) p y p )

[0025] =A 0 (x p y p )+ΔA n (x p y p )

[0026] In the formula, A n For the actual acquisition of background light intensity model, A 0 (x p y p ) represents the ideal background light intensity, ΔA n (x p y p Let be the background light intensity error function. The ambient light noise is posited as one-dimensional superimposed Gaussian noise. The probability density influence function of the light intensity noise under the influence of the Gaussian mixture model is expressed as:

[0027]

[0028] In the formula, α kLet x be the probability that the observed data belongs to the Kth sub-model, and θ be the target observed data in the sample data. k Let be the Gaussian distribution density function of the k-th sub-model.

[0029] As a further preferred embodiment, in step S23, the light intensity noise model is substituted into the measured light intensity function, and the partial derivative of the light intensity error function is solved to obtain the relationship between the wrapping phase error and the light intensity error:

[0030]

[0031] In the formula, Δφ is the phase error, and ΔA is the phase error. n Let A be the background light intensity error function, φ be the phase information, and A be the background light intensity error function. n This is a model of the actual background light intensity.

[0032] Preferably, in step S24, the phase error Δφ under the influence of multiple noises is modeled into two types of errors:

[0033] Δφ=Δφ L +Δφ T

[0034] In the formula, Δφ represents the phase error under the overall influence. L The phase error Δφ is caused by the influence of periodic ambient light. T The phase error is caused by reflection from the surface of the object under test.

[0035] As a further preferred embodiment, the phase error Δφ under the influence of periodic ambient light... L The calculation model is as follows:

[0036]

[0037] The phase error Δφ under the influence of surface reflection of the object under test T The calculation model is as follows:

[0038]

[0039] In the formula, N is the phase shift step, B is the intensity modulation, and ΔA n Let ΔA be the background light intensity error function under the influence of periodic ambient light. i Let φ be the background light intensity error function under the influence of surface reflection, and φ be the phase information.

[0040] As a further preferred option, step S3 includes:

[0041] S31 solves for the ambient light frequency and obtains the least desirable exposure time by using the light intensity function under environmental interference. Through analysis of the pre-projected light intensity function, it is found that in the four-step phase-shift method of the light measurement system under periodic ambient light influence, the light intensity of adjacent fringe projection steps differs from that of adjacent steps due to the periodic ambient light influence. Let the frequency of the periodic ambient light influence be F. L F is obtained by solving for adjacent phases. L :

[0042]

[0043] In the formula, N is the number of phase shift steps, and A n For the actual acquisition of background light intensity model, A 0 For the ideal background light intensity model,

[0044] According to the F L Identify and compensate for the interval between the projected exposure and the camera acquisition trigger, thereby obtaining the optimal exposure time;

[0045] S32 determines the effective range of stripe projection acquisition by setting a mask image based on the light intensity noise model:

[0046]

[0047] In the formula, ΔM k (x p y p M represents the mask range corresponding to pixel P. k-1 (x p y p M is the effective upper bound corresponding to pixel P. k (x p y p ) represents the effective lower bound corresponding to pixel P. The upper limit represents the maximum projector brightness. This represents the maximum projector brightness corresponding to the lower bound.

[0048] As a further preferred option, the following also includes:

[0049] By calculating the effective range in real time, an adaptive compensation function is set for the acquired background light intensity A and modulation intensity B, where the modulation intensity B is set as follows:

[0050]

[0051] The average strength is:

[0052]

[0053] In the formula, Bp (x p y p () represents the modulation intensity. L represents the maximum capture brightness corresponding to the saturated pixels of the matched camera. min (x p y p A represents the minimum brightness of the projected stripe pattern. p (x p y p ) is the background light intensity function of pixel P collected by the device.

[0054] As a further preferred embodiment, step S4 includes the following steps:

[0055] S41 processes the acquired fringes using an exposure function and a mask image to obtain a set of multi-parameter linear programming equations for the optimized projected fringe frequencies:

[0056]

[0057] In the formula, N is the number of phase shift steps, σ n Let F be a Gaussian distributed variable. L Let F be the frequency of the periodic ambient light effect, I0 be the light intensity function at point P, and A be the frequency of the ambient light effect. p Let B be the background light intensity function for pixel P. p Let Exp be the modulation intensity function for the acquisition of pixel P. pro For the projector exposure time, Exp L This refers to the periodic exposure of Huang Jing's case;

[0058] Based on the absolute phase required in structured light reconstruction, S42 expands the absolute phase range from 2π to 2πF, where F is the total number of periods of the set fringe pattern. The optimal fringe frequency within this range is as follows:

[0059]

[0060] In the formula, (B P ) min σ is the minimum modulation intensity function within the threshold range. n Let σ be a Gaussian distributed variable. φ This represents the standard deviation of the phase error.

[0061] According to another aspect of the present invention, a self-testing optimization system for a high-speed, high-precision structured light measurement system is also provided, comprising:

[0062] The first main control module is used to build the optical measurement system and calibrate the measurement system;

[0063] The second main control module is used to pre-project a periodic stripe pattern onto the surface of the object to be measured, obtain the measured light intensity function under environmental interference, and construct a light intensity noise model under environmental interference based on the measured light intensity function.

[0064] The third main control module is used to solve the ambient light frequency and obtain the optimal exposure time based on the light intensity noise model and the light intensity function under environmental interference, and to calculate the threshold range of the saturated mask image through the light intensity function.

[0065] The fourth main control module is used to obtain the optimal exposure and optimal stripe frequency of the measurement system based on environmental parameters, in order to compensate for the original measurement parameters;

[0066] The fifth main control module is used to calculate the point cloud of the target object based on the compensated original measurement parameters and the acquired light intensity function of the object under test.

[0067] In summary, compared with the prior art, the above-described technical solutions conceived by this invention mainly possess the following technical advantages:

[0068] 1. This invention addresses the measurement system under varying ambient light noise conditions within a binocular structured light measurement system. It calculates the ambient light interference frequency through pre-projection and uses this interference frequency to compensate for the acquisition parameters of the binocular structured light system, thereby optimizing the reconstruction effect. Specifically, this invention designs a compensation algorithm for environmental interference based on the principle of binocular structured light and optimizes the structured light 3D reconstruction system based on this algorithm. By designing a pre-projection mode to solve for the ambient light frequency, the compensation is incorporated into the structured light image acquisition program, improving the accuracy of the reconstruction effect of the structured light 3D measurement system under various ambient light conditions. Furthermore, by designing an algorithm to plan the reconstruction fringe frequency within an effective threshold range, the invention improves the structured light scanning and reconstruction speed while maintaining accuracy requirements with the lowest possible binocular structured light phase shift frequency.

[0069] 2. This invention obtains the measured light intensity function under environmental interference through pre-projection, that is, obtains the interference parameters under the current measurement environment to compensate for the light intensity function under the influence of the environment, thereby improving the measurement accuracy. At the same time, the relationship between the phase light intensity error and the pre-projection process is obtained, which facilitates the subsequent decomposition and optimization of the phase error under the influence of light intensity error.

[0070] 3. This invention uses a masking process in the acquired projection pattern to divide the acquired stripe image into effective regions, thereby extracting effective information and avoiding the problem of uneven grayscale distribution in the acquired stripe image.

[0071] 4. In this invention, a three-dimensional measurement device is used to compensate for periodic environmental noise, thereby avoiding phase errors caused by uneven distribution of light intensity functions of each pixel in the acquired image due to uneven measurement projection and periodic environmental noise.

[0072] 5. This invention optimizes the computational speed of the reconstruction algorithm by planning the projection frequency of the structured light measurement system within the effective reconstruction area, using the smallest fringe projection frequency within the effective range, and performing template calculations.

[0073] 6. This invention optimizes the phase error caused by reflection by setting a filtering range during the pre-projection process and optimizing the projection exposure time and projection fringe frequency for the light intensity function within the effective range. This controls the measurement results of the three-dimensional measurement system for the target environment to be within the effective range, thereby achieving optimization of the phase error caused by reflection. Attached Figure Description

[0074] Figure 1 This is a flowchart of a self-detection optimization method for a high-speed, high-precision structured light measurement system according to a preferred embodiment of the present invention;

[0075] Figure 2 This is a schematic diagram of the measurement system involved in the self-detection optimization system of the high-speed and high-precision structured light measurement system of the present invention.

[0076] In all the accompanying drawings, the same reference numerals denote the same technical features, specifically: 1-camera, 2-target object, 3-projection stripe generating device. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0078] like Figure 1 and Figure 2 As shown in the figure, the self-detection optimization method of a high-speed, high-precision structured light measurement system provided by this invention includes: a projection fringe generator and two cameras, typically monochrome cameras. This method optimizes the structured light 3D reconstruction system by designing a compensation algorithm for environmental interference based on the principle of binocular structured light. It solves for the ambient light frequency by designing a pre-projection mode, thereby compensating for it in the structured light image acquisition program. This improves the accuracy of the reconstruction effect of the structured light 3D measurement system under various ambient light conditions. Furthermore, by designing an algorithm to plan the reconstruction fringe frequency within an effective threshold range, it improves the structured light scanning and reconstruction speed while maintaining accuracy requirements with the lowest possible binocular structured light phase shift frequency. The following is a specific implementation scheme of this patent.

[0079] Step 1: Construct and calibrate the measurement system. The measurement system involved in this invention includes a binocular camera, a projection fringe generator, and a target object. First, the camera is calibrated and its calibration model and parameters are acquired. In this invention, conventional calibration methods are used to acquire the calibration model and parameters. The purpose is to calculate the camera position and acquire the measurement equipment parameters before acquiring the measurement point cloud, and to calibrate the camera based on the distortion parameters within the binocular camera's internal parameters. In one embodiment of this invention, Zhang's calibration method is used for calibration, specifically:

[0080]

[0081] The original expression can be expressed as:

[0082]

[0083]

[0084] in: The homogeneous coordinates represent the pixel coordinates (u, v, 1) of the image plane; The homogeneous coordinates represent the coordinates (X, Y, Z, 1) of the world coordinate system; R represents the rotation matrix; t represents the translation matrix; s represents the scale factor. A represents the camera intrinsic parameters.

[0085] Step two involves pre-projecting the target object and acquiring the measured light intensity function under environmental interference. The purpose of this step is to obtain the interference parameters under the current measurement environment to compensate for the light intensity function under environmental influence, thereby improving measurement accuracy. In a practical binocular structured light measurement system, the acquisition of the fringe pattern at each point of the object is affected by various parameters. The main principle is to obtain the grayscale information of the target pixels by encoding and sampling grayscale points through the camera. This grayscale information can be represented by a light intensity function. Therefore, the height information represented by each pair of corresponding pixels in the binocular camera is represented by the light intensity function.

[0086] I n (x, y) = A(x, y)

[0087] +B(x,y)cos[φ(x,y)-2πn / N] (0-4)

[0088] In the formula, A(x,y) is the background light intensity, B(x,y) is the modulation intensity, φ(x,y) is the phase information, and N is the phase shifting step.

[0089] Binocular structured light systems used in working environments are inevitably affected by the scanning environment. Specifically, measurement systems on AGV platforms are affected by ambient light under different measurement conditions, the influence of the stripe generator, and the combined effect of reflection from the surface of the projected object. Therefore, all error influences can be uniformly processed to obtain the following light intensity noise model acting on the background light intensity:

[0090] A n =A(x) p y p )

[0091] =A 0 (x p y p )+ΔA n (x p y p (0-5)

[0092] In the formula, A n For the actual acquisition of background light intensity model, A 0 (x p y p ) represents the ideal background light intensity, ΔA n (x p y p Let be the background light intensity error function. The ambient light noise is posited as one-dimensional superimposed Gaussian noise. The probability density influence function of the background light intensity noise under the influence of the Gaussian mixture model can be expressed as:

[0093]

[0094] In the formula, α k Let x be the probability that the observed data belongs to the Kth sub-model, and θ be the target observed data in the sample data. k Let be the Gaussian distribution density function of the k-th sub-model.

[0095] Substitute the acquired background light intensity error model (0-5) into the acquired light intensity function and apply it to the light intensity error function ΔA. n (x p y p By performing partial derivatives, the relationship between the wrapping phase error and the light intensity error can be obtained as follows:

[0096]

[0097] In the formula, Δφ is the phase error, and ΔA is the phase error. n Let A be the background light intensity error function, φ be the phase information, and A be the background light intensity error function. n This is a model of the actual background light intensity.

[0098] In this invention, the phase is considered under the influence of periodic ambient light and the reflection from the target object's surface. Therefore, the phase error Δφ under the influence of multiple noises can be modeled as two types of errors:

[0099] Δφ=Δφ L +Δφ T (0-8)

[0100] Where Δφ is the phase error under the overall influence, Δφ L The phase error Δφ is caused by the influence of periodic ambient light. T The phase error is caused by reflection from the surface of the object under test.

[0101] Phase error Δφ under the influence of periodic ambient light L By taking the partial derivative of the light intensity function, we can obtain the following result:

[0102]

[0103] in:

[0104]

[0105]

[0106]

[0107]

[0108] In the four-step phase shift, the partial derivative of the phase information φ yields the phase error model under periodic influence: the phase difference between adjacent projected fringes is π / 2. Therefore, the difference between the intensity functions acquired from adjacent π / 2 acquisitions can be used to obtain the phase error Δφ. L :

[0109]

[0110] Phase error Δφ due to reflection from the surface of the object under test L The effective interval can be divided into the fringe acquisition image in the pre-projection four-step phase-shifting method, and the light intensity error can be processed by planning the target projection frequency within the effective interval. The following model is set for fitting, and the light intensity error can be expressed by the following model:

[0111]

[0112] The partial derivative function of the phase under this error model can be obtained as follows:

[0113]

[0114] This leads to the expression for the phase error caused by reflection from the surface of the object under test:

[0115]

[0116] This process is carried out during the pre-projection work to obtain the relationship between light intensity error and phase, which facilitates the subsequent decomposition and optimization of phase error under the influence of light intensity error.

[0117] Step 3: Solve the ambient light frequency and obtain the optimal exposure time using the light intensity function under interference. Analyze the pre-projected light intensity function to find that in the four-step phase-shift method of the measurement system under the influence of periodic light noise, the camera acquisition between adjacent fringe projection steps will be affected by periodic light noise. The difference between the fringe acquisition light intensity and the fringe acquisition light intensity in adjacent steps can be calculated. The frequency of the periodic noise influence can be set as F. L F is obtained by solving for adjacent phases. L Expression:

[0118]

[0119] Identify and compensate for the interval between the projected exposure and the camera acquisition trigger, i.e., satisfy the following conditions:

[0120]

[0121] By using a 3D measurement device that compensates for periodic environmental noise, phase errors caused by uneven distribution of light intensity functions in individual pixels of the acquired image due to the unevenness between the measurement projection and periodic environmental noise are avoided.

[0122] Step four involves determining the threshold range by calculating the saturated mask image using the light intensity function. Specifically, this can be implemented during the pre-projection process by setting a filtering range for the error interval affected by surface reflection. For light intensity functions within the effective range, the projection exposure time and projection fringe frequency are optimized. This ensures that the measurement results of the 3D measurement system for the target environment are all within the effective range, thereby optimizing the phase error caused by reflection.

[0123] In this embodiment, the error model under the influence of surface reflection light error is set by equation (0-13), and the mask image M is set. k (x, y) determines the effective range in fringe projection acquisition:

[0124]

[0125] In the formula, ΔM k (x p y p M represents the mask range corresponding to pixel P. k-1 (xp y p M is the effective upper bound corresponding to pixel P. k (x p y p ) represents the effective lower bound corresponding to pixel P. The upper limit represents the maximum projector brightness. This represents the maximum projector brightness corresponding to the lower bound.

[0126] Its maximum reflected brightness is the maximum acquisition brightness corresponding to the saturated pixels of the matched camera. The effective brightness range is calculated in real time, and an adaptive compensation function is set for the collected background light intensity A and modulation intensity B, where the modulation intensity is set as follows:

[0127]

[0128] The average strength is:

[0129]

[0130] In the formula, B p (x p y p () represents the modulation intensity. L represents the maximum capture brightness corresponding to the saturated pixels of the matched camera. min (x p y p A represents the minimum brightness of the projected stripe pattern. p (x p y p ) is the background light intensity function of pixel P collected by the device.

[0131] Step 5: Based on environmental parameters, obtain the optimal exposure and optimal fringe frequency compensation parameters for the measurement system. Specifically, process the acquired fringes using an exposure function and a mask image to obtain a set of multi-parameter linear programming equations for the optimized projected fringe frequency:

[0132]

[0133] Where N is the phase shift step, B is the intensity modulation, and σ is a Gaussian distributed variable. n I0 is the light intensity function at point P, and A p Let B be the background light intensity function for pixel P. p Let Exp be the modulation intensity function for the acquisition of pixel P. pro For the projector exposure time, Exp L This refers to the periodic exposure of Huang Jing's case.

[0134] To address the absolute phase required in structured light reconstruction, its defined phase range needs to be expanded from 2π to 2πF, where F is the total number of periods of the desired fringe pattern. The optimal fringe frequencies within this range are as follows:

[0135]

[0136] Among them (B) P ) min The minimum modulation intensity function within the threshold range is used to maximize the frequency of the projected fringe, reduce the standard deviation of the error, and thus improve the measurement accuracy of the measurement system.

[0137] Step six: Calculate the point cloud based on the real-time acquired light intensity function. For a binocular structured light system, the imaging planes of the left and right cameras need to be rotated to achieve epipolar alignment, so as to satisfy the triangulation between the same image points of the left and right cameras in triangulation to calculate the three-dimensional height of the target image point. In the multi-frequency four-step phase-shifting method, the relationship between the gradient levels k between adjacent layers is used for encoding operations to complete the unwrapping phase in multi-frequency phase shifting:

[0138]

[0139] Where, φ h (x, y), φ l (x, y) represents the absolute phase. For the wrapping phase.

[0140] The height information of the corresponding image points is calculated by searching along the corresponding epipolar lines between the corrected image planes. The coordinate parameters of the target can be calculated using the principle of similarity as follows:

[0141]

[0142]

[0143]

[0144] Where f is the camera focal length, and b is the camera baseline distance. The focal length of the left camera is f. l =f / d x Right camera focal length f r =f / d y The matching point pairs of the left and right images are (u l ,v l ) and (u r ,v r Z c This refers to the height information of a three-dimensional point.

[0145] The three-dimensional point set obtained by matching the light intensity information of each pixel in the absolute phase map is the measurement-optimized point cloud space:

[0146]

[0147] In summary, the method of this invention obtains environmental variables by pre-projection and optimizes the accuracy of the surface structured light measurement system by real-time compensation to the parameters of the surface structured light measurement system. It adds a compensation factor to the original four-step phase shift method, thereby avoiding the influence of the measurement environment on the measurement effect of the mobile robot under different ambient light conditions and ensuring the efficiency of measurement. Furthermore, by setting an effective measurement threshold, it further avoids the error caused by surface structured light projection of highly reflective objects.

[0148] According to another aspect of the present invention, a self-detection optimization system for a high-speed, high-precision structured light measurement system is also provided, for implementing any combination of the methods in the above embodiments, comprising: a first main control module for constructing the light measurement system and calibrating the measurement system; a second main control module for pre-projecting a periodic stripe pattern onto the surface of the object to be measured, obtaining the measured light intensity function under environmental interference, and constructing a light intensity noise model under environmental interference based on the measured light intensity function; a third main control module for solving the ambient light frequency and obtaining the optimal exposure time based on the light intensity noise model and the light intensity function under environmental interference, and calculating the threshold range of the saturated mask image based on the light intensity function; a fourth main control module for obtaining the optimal exposure and optimal stripe frequency of the measurement system based on environmental parameters to compensate for the original measurement parameters; and a fifth main control module for calculating the point cloud of the target object based on the compensated original measurement parameters and the acquired light intensity function of the object to be measured.

[0149] The optical measurement system comprises a projection fringe generator and two cameras. The projection fringe generator projects periodic fringe patterns onto the surface of the object under test, while the two cameras acquire images of the object's surface in real time. The system also includes at least one control system to coordinate the actions of the various devices and execute calculations or computational programs as described above.

[0150] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A self-detection optimization method of a high-speed high-precision structured light measurement system, characterized in that, The method comprises the following steps: S1, constructing a structured light measurement system and calibrating the measurement system; S2, projecting a periodic fringe pattern to the surface of a to-be-measured object in advance, obtaining a measurement light intensity function under environmental interference, and constructing a light intensity noise model under environmental interference according to the measurement light intensity function; Step S2 comprises the following steps: S21, obtaining gray scale information of a pixel point of the to-be-measured object by sampling the encoding of a gray scale point of a picture taken by a camera, and representing the gray scale information as a measurement light intensity function; S22, constructing a light intensity noise model under the joint action of environmental light influence and reflection influence of the surface of the measured object; S23, solving the relationship between wrapped phase error and light intensity error based on the measurement light intensity function and the light intensity noise model; S24, according to the relationship between the wrapped phase error and the light intensity error, dividing the phase error under multiple noise influences into phase error under overall influence and phase error under periodic environmental light influence; S3, according to the light intensity noise model, solving the environmental light frequency through the light intensity function under environmental interference, obtaining the best exposure time, and calculating the threshold range of the saturated mask image through the light intensity function; step S3 comprises: S31, the light intensity function under the environmental interference is used to solve the environmental light frequency and obtain the best exposure time. Through the analysis of the pre-projected light intensity function, it is found that in the four-step phase shift method of the light measurement system under the influence of periodic environmental light, the light intensity collected by the adjacent stripe projection steps will be different due to the influence of periodic environmental light. The frequency of the periodic environmental light is set as , and the adjacent phase is solved to obtain : wherein N is the phase shift step, is the actual collected background light intensity model, A 0 is the ideal background light intensity model, according to the identifies and compensates for the projection exposure to camera acquisition trigger interval, resulting in the optimal exposure time, is the phase information; S32, according to the light intensity noise model, determining the effective range in the fringe projection collection by setting a mask image: ， wherein, is a mask range corresponding to the pixel P point, is an effective upper bound corresponding to the pixel P point, is an effective lower bound corresponding to the pixel P point, is a maximum projector brightness corresponding to the upper bound, is a maximum projector brightness corresponding to the lower bound; S4, obtaining the best exposure time and the best fringe frequency of the measurement system according to environmental parameters to compensate for the original measurement parameters; S5, calculating the point cloud of the target object according to the compensated original measurement parameters and the collected light intensity function of the to-be-measured object.

2. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 1, characterized in that, In step S1, the light measurement system comprises a binocular camera and a measurement object arranged in a non-collinear manner, and a fringe generating device is further arranged between the binocular camera, which is used for periodic fringe projection to the measurement object.

3. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 2, characterized in that, In step S21, the measurement light intensity function is: ， wherein is the background light intensity, is the modulation intensity, is the phase information, and N is the number of phase shifting steps.

4. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 3, characterized in that, In step S22, the light intensity noise model comprises: ， In the formula, is the actual background light intensity model, is the ideal background light intensity model, is the background light intensity error function, is the coordinate of the pixel P, the ambient light noise is designed as one-dimensional superimposable Gaussian noise, and the light intensity noise under the influence of the probability density of the Gaussian mixture model is represented as a probability density influence function: ， In the formula, is the probability that the observation data belongs to the Kth sub-model, is the target observation data in the sample data, is the Gaussian distribution density function of the kth sub-model, is the phase information.

5. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 4, characterized in that, In step S23, the light intensity noise model is brought into the measurement light intensity function, and the partial derivative of the light intensity error function is solved to obtain the relationship between the wrapped phase error and the light intensity error: ， wherein is a phase error, is a background light intensity error function, is phase information, is an actual collected background light intensity model.

6. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 5, characterized in that, In step S24, the phase error under multiple noise influences is divided into two types of errors for modeling: ， wherein is the phase error under the overall influence, is the phase error under the periodic ambient light influence, is the phase error under the surface reflection of the object under test influence.

7. The self-detection optimization method of the high-speed and high-precision structured light measurement system according to claim 6, characterized in that, phase error under the influence of the periodic ambient light The computational model is as follows: ， Phase error under influence of surface reflection of the object to be measured The computational model is as follows: ， wherein N is the phase shift step size, B is the intensity modulation, is the background light intensity error function under periodic ambient light influence, is the background light intensity error function under surface reflection influence, is the phase information.

8. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 7, characterized in that, Further comprising: By real-time calculation of the effective range collected, an adaptive compensation function is set for the background light intensity A and the modulation intensity B collected, wherein the modulation intensity B is set as: ， The average intensity is: ， In the formula, is the modulation intensity, is the maximum acquisition brightness corresponding to the matched camera saturated pixel, wherein k = 1, 2, …, K, K is the sample number, is the minimum brightness of the projected fringe pattern, is the background light intensity function acquired by the device at the pixel P point.

9. The self-detection optimization method of a high-speed high-precision structured light measurement system according to claim 8, wherein, Step S4 comprises the following steps: S41, processing the collected fringe by the exposure function and the mask image to obtain a multi-parameter linear programming equation set of the optimized projection fringe frequency: ， where N is the number of phase shift steps, is a Gaussian distributed variable, is the frequency of the periodic ambient light influence, is the light intensity function for the P-point, is the acquisition background light intensity function for the P-point pixel, is the acquisition modulated light intensity function for the P-point pixel, is the projector exposure time, is the periodic ambient exposure time; S42 Based on the absolute phase required in the structured light reconstruction, the absolute phase range is unfolded from to where, F is the total number of periods of the set of fringe patterns, the best fringe frequency in the range is obtained as follows: ， wherein is the minimum modulation light intensity function within the threshold range, is a Gaussian distributed variable, is the standard deviation of the phase error.

10. A self-detection optimization system for a high-speed high-precision structured light measurement system, characterized in that, Comprise: A first main control module is configured to construct a light measurement system and calibrate the measurement system; A second main control module is configured to project a periodic fringe pattern to the surface of a to-be-measured object in advance, obtain a measurement light intensity function under environmental interference, and construct a light intensity noise model under environmental interference according to the measurement light intensity function; specifically, The gray scale information of the pixel point of the object to be measured is obtained by sampling the encoding of the gray scale point of the picture taken by the camera, and the gray scale information is expressed as a measured light intensity function; An ambient light influence and a measured object surface reflection influence jointly acting light intensity noise model is constructed; The relationship between the wrapped phase error and the light intensity error is solved based on the measured light intensity function and the light intensity noise model; According to the relationship between the wrapped phase error and the light intensity error, the phase error under the influence of multiple noises is divided into the phase error under the influence of the overall influence and the phase error under the influence of the periodic ambient light; The environmental light frequency is solved by the light intensity function under the environmental interference, and the best exposure time is obtained. Through the analysis of the pre-projected light intensity function, in the four-step phase shift method of the light measurement system under the influence of periodic environmental light, the number of adjacent stripe projection steps will be affected by the periodic environmental light, and the light intensity of the collected stripes will be different from the light intensity of the collected stripes in the adjacent steps. The frequency of the periodic environmental light is , and the adjacent phase is solved to obtain : ， wherein N is the phase shift step size, is the actual background light intensity model, A 0 is the ideal background light intensity model, according to the identifying and compensating for the projection exposure and camera acquisition trigger interval, resulting in an optimal exposure time, is the phase information; A third main control module is configured to solve the ambient light frequency and obtain the best exposure time through the light intensity function under the influence of the ambient interference according to the light intensity noise model, and calculate the threshold range of the saturated mask image through the light intensity function; ， wherein, is a mask range corresponding to the pixel P point, is an effective upper bound corresponding to the pixel P point, is an effective lower bound corresponding to the pixel P point, is a maximum projector brightness corresponding to the upper bound, is a maximum projector brightness corresponding to the lower bound; According to the light intensity noise model, the effective range in the stripe projection collection is determined by setting the mask image: A fourth main control module is configured to obtain the best exposure and the best stripe frequency of the measurement system according to the environmental parameters to compensate for the original measurement parameters; A fifth main control module is configured to calculate the point cloud of the target object according to the compensated original measurement parameters and the collected light intensity function of the object to be measured.

Images