Machine vision-based industrial robot intelligent control system and method

Abstract

The application relates to the technical field of industrial robots, in particular to an intelligent control system and method for an industrial robot based on machine vision, which comprises a polarization interference imaging module, a dielectric characteristic analysis module, a dynamic clamping force field generation module, a multi-view field fusion module, a sub-pixel compensation module and an electromagnetic shielding cooperation module; wherein: the polarization interference imaging module is used for acquiring a polarization characteristic spectrum of a workpiece surface; the dielectric characteristic analysis module is used for calculating a surface dielectric constant distribution; the dynamic clamping force field generation module is used for generating a vacuum adsorption force field distribution graph with a pressure gradient; and the multi-view field fusion module generates an anti-slip three-dimensional grabbing track through a parallax compensation algorithm. Through the combination of the polarization interference imaging, dielectric characteristic analysis, sub-pixel compensation and electromagnetic shielding technologies, the grabbing precision and stability of the industrial robot in a complex environment are significantly improved, and external interference is effectively inhibited through electromagnetic shielding.

Description

Technical Field

[0001] This invention relates to the field of industrial robot technology, and in particular to an intelligent control system and method for industrial robots based on machine vision. Background Technology

[0002] With the continuous improvement of industrial automation, industrial robots are being used more and more widely in production lines, especially in scenarios requiring high precision and high efficiency. Robot gripping tasks play an important role in fields such as electronics manufacturing, automobile assembly, and precision machining. Traditional industrial robot gripping technology usually relies on the mechanical control of the robotic arm and a simple visual feedback system to obtain the position and posture information of the workpiece through external sensors, thereby completing the gripping. However, due to the complexity of the environment and the variety of workpiece shapes, the gripping accuracy and stability of these traditional systems are often limited. Especially when dealing with workpieces with smooth, transparent, or irregular shapes, existing gripping methods often cannot guarantee accuracy and stability.

[0003] Existing robotic grasping methods typically rely on a single vision or mechanical sensor for workpiece identification and path planning, which often struggles to simultaneously meet the requirements of high-precision grasping and adaptability to complex working conditions. For example, traditional vision systems are easily affected by changes in lighting, workpiece surface reflection, or transparent materials, leading to misjudgments. Mechanical sensors, on the other hand, cannot accurately capture minute deformation information, affecting grasping accuracy. Furthermore, existing systems often neglect electromagnetic interference in the environment during grasping path planning, and the microscopic vision unit may be affected by external electromagnetic interference, resulting in a decrease in image acquisition quality. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides an intelligent control system and method for industrial robots based on machine vision.

[0005] An intelligent control system for industrial robots based on machine vision includes a polarization interferometric imaging module, a dielectric feature analysis module, a dynamic clamping force field generation module, a multi-field-of-view fusion module, a sub-pixel compensation module, and an electromagnetic shielding coordination module; wherein:

[0006] Polarization interferometric imaging module: used to acquire the polarization feature spectrum of the workpiece surface by rotating the polarizer array and extract the Stokes parameters of each pixel;

[0007] Dielectric characteristic analysis module: used to receive Stokes parameters, calculate the surface dielectric constant distribution based on Fresnel reflection law, and identify the workpiece material type;

[0008] Dynamic clamping force field generation module: used to match the preset stress-deformation curve according to the identified workpiece material type and generate a vacuum adsorption force field distribution map with pressure gradient;

[0009] Multi-field fusion module: used to receive vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generate anti-slip 3D grasping trajectory through parallax compensation algorithm;

[0010] Subpixel compensation module: Based on the microscopic vision unit to capture the microscopic deformation features of the adsorption contact surface, the module performs subpixel-level path correction on the 3D grasping trajectory generated by the multi-field fusion module.

[0011] Electromagnetic shielding coordination module: used to activate the annular electromagnetic barrier when the vacuum adsorption force field is activated, suppress the influence of electromagnetic interference from surrounding equipment on the microscopic vision unit, and synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator.

[0012] Optionally, the polarization interferometry imaging module includes a tunable wavelength laser source unit, a four-way polarizer array unit, an image acquisition unit, and a Stokes parameter calculation unit; wherein:

[0013] Tunable wavelength laser source unit: used to emit tunable laser with wavelengths covering 400-1000nm to irradiate the surface of the workpiece;

[0014] Four-way polarizer array unit: By rotating the polarizer array, the polarization direction is switched in 30° increments to acquire four sets of interference images respectively;

[0015] Image acquisition unit: Works synchronously with the four-way polarizer array unit to trigger and receive interference images in each polarization direction, and acquire surface information at different angles;

[0016] Stokes parameter calculation unit: used to process four sets of interferometric images and calculate the Stokes parameters for each pixel, including polarization angle and phase difference.

[0017] Optionally, the Stokes parameter calculation unit includes:

[0018] Image preprocessing: Each set of interferometric images is preprocessed to obtain the polarization information of each pixel. Specifically, image difference processing is performed on the four interferometric images to obtain the light intensity value in each polarization direction, denoted as... ;

[0019] Calculate Stokes parameters: Calculate Stokes parameters based on the light intensity values ​​in each polarization direction. The specific formulas include:

[0020] ,in, It represents the total light intensity, which is the sum of the light intensities under all polarization directions;

[0021] ,in, The first component representing the difference in light intensity is used to represent the difference between horizontal and vertical polarization;

[0022] ,in, The second component representing the difference in light intensity is used to represent and Differences in polarization;

[0023] ,in, The third component, representing the difference in light intensity, is used to represent the difference between right-handed and left-handed circular polarization.

[0024] Optionally, the dielectric characteristic analysis module includes a Stokes parameter receiving unit, a dielectric constant calculation unit, and a material identification unit; wherein:

[0025] Stokes parameter receiving unit: used to receive Stokes parameters provided by the polarization interferometric imaging module. ;

[0026] Dielectric constant calculation unit: Used to calculate the surface dielectric constant distribution based on Fresnel reflection law, specifically calculating the polarization angle of each pixel based on the previously received Stokes parameters. and phase difference Then, the surface dielectric constant is calculated using Fresnel's law of reflection, with the following formula: ,in, The dielectric constant of the workpiece surface. It is the polarization angle. Phase difference;

[0027] Material identification unit: used to process the calculated dielectric constant The dielectric constant range of various materials in the preset material database is compared; based on the comparison results, the material type of the workpiece is identified and the corresponding material label is output.

[0028] Optionally, the dynamic clamping force field generation module includes a material type receiving unit, a stress-deformation curve matching unit, and a pressure gradient generation unit; wherein:

[0029] Material type receiving unit: used to receive workpiece material type information provided by the dielectric characteristic analysis module;

[0030] Stress-deformation curve matching unit: Used to select the stress-deformation curve corresponding to the received workpiece material type from a preset material database. The expression for the stress-deformation curve is: ,in, For the stress of the material, The Young's modulus of the material. For the strain of the material;

[0031] Pressure gradient generation unit: Used to generate a vacuum adsorption force field distribution map with pressure gradient based on the matched stress-deformation curve and material type. During this process, the adsorption pressure... Calculate using the following formula: ,in, Indicates adsorption pressure. For safety reasons, The maximum withstand stress of the material. The dielectric constant is calculated for the surface of the workpiece.

[0032] Optionally, the multi-field fusion module includes a vacuum adsorption force field distribution map receiving unit, a parallax compensation unit, and a grasping trajectory generation unit; wherein:

[0033] Vacuum adsorption force field distribution map receiving unit: used to receive vacuum adsorption force field distribution maps captured by industrial cameras from at least three directions;

[0034] Parallax compensation unit: used to perform parallax compensation based on the received multi-view vacuum adsorption force field distribution map. By calculating the parallax between each pair of adjacent view images, it eliminates image distortion caused by different camera angles.

[0035] Grasping trajectory generation unit: Based on the compensated multi-view vacuum adsorption force field distribution map, the iterative nearest point algorithm is used to fuse multi-view data to generate an anti-slip 3D grasping trajectory.

[0036] Optionally, the grasping trajectory generation unit includes:

[0037] Data preparation: Receive compensated vacuum adsorption force field distribution images from multiple industrial cameras, each corresponding to image data from different viewpoints. Each image contains a set of three-dimensional coordinate points of the workpiece surface and force field distribution, denoted as... ;

[0038] Preliminary registration: By performing preliminary registration on the data of each viewpoint, the three-dimensional coordinate point sets of multiple viewpoints are connected. Project onto a unified coordinate system; at this point, define the initial transformation matrix. It is used to roughly align data from different viewpoints to obtain a preliminary aligned point set. ;

[0039] Iterative matching: The iterative nearest point algorithm is used to perform matching by minimizing the distance between each pair of corresponding point sets. The goal of the iterative nearest point algorithm is to solve for the optimal rigid transformation matrix. The minimization function is defined as: ,in, It is the target point of the current viewpoint. These are the points to be aligned. It is the transformation matrix of the current iteration step. Represents Euclidean distance. The total number of points in the set;

[0040] Trajectory generation: After completing iterative matching, a three-dimensional grasping trajectory with anti-slip function is generated by combining the mechanical constraints of the adsorption force field.

[0041] Optionally, the sub-pixel compensation module includes a microscopic vision unit, a deformation feature extraction unit, and a path correction unit; wherein:

[0042] Microscopic vision unit: used to capture the microscopic deformation features of the adsorption contact surface. This microscopic vision unit is equipped with a high-speed microscope camera and a high-precision optical zoom lens, which can perform high-resolution imaging of the contact surface, with an imaging resolution of 2μm / pixel.

[0043] Deformation Feature Extraction Unit: Utilizes the Hough transform algorithm to extract concentric circle distortion features on the adsorption contact surface, thereby determining the microscopic deformation on the contact surface. ;

[0044] Path correction unit: Based on deformation, performs sub-pixel-level path correction on the 3D grasping trajectory generated by the multi-view fusion module to obtain the corrected 3D grasping trajectory; the path correction calculation formula is: ,in, To correct the displacement, The deformation is obtained through the deformation feature extraction unit.

[0045] Optionally, the electromagnetic shielding coordination module includes an electromagnetic interference detection unit, an electromagnetic shielding activation unit, and a trajectory synchronization unit; wherein:

[0046] Electromagnetic Interference Detection Unit: Used to monitor the electromagnetic environment of the working area, especially the electromagnetic interference level in the area where the microscopic vision unit is located. It detects the magnetic field strength of the working area through a Hall sensor. When the magnetic field strength is detected When the electromagnetic interference level exceeds the preset threshold, it is determined to be high, and electromagnetic shielding measures need to be activated.

[0047] Electromagnetic shielding activation unit: When the electromagnetic interference level in the working area reported by the electromagnetic interference detection unit exceeds the threshold, it is used to activate the annular electromagnetic barrier to generate a reverse magnetic field. The reverse magnetic field is generated according to the following formula: ,in, It is a reverse magnetic field. The original magnetic field strength detected in the working area;

[0048] Trajectory synchronization unit: used to synchronize the 3D grasping trajectory corrected by the sub-pixel compensation module to the actuator to perform the grasping operation.

[0049] A machine vision-based intelligent control method for industrial robots, implemented by the aforementioned machine vision-based intelligent control system for industrial robots, includes the following steps:

[0050] S1: Obtain the polarization feature map of the workpiece surface. Specifically, the polarization information of each pixel on the workpiece surface is collected by rotating the polarizer array, and the Stokes parameter of each pixel is calculated.

[0051] S2: Based on the Stokes parameters obtained in S1, the Stokes parameters of each pixel are calculated using the Jones matrix method, thereby obtaining the dielectric constant distribution of the workpiece surface;

[0052] S3: Based on the dielectric constant distribution obtained in S2, identify the material type of the workpiece and select the matching material type according to the preset material library;

[0053] S4: Based on the workpiece material type identified in S3, match the preset stress-deformation curve and generate a vacuum adsorption force field distribution map with pressure gradient.

[0054] S5: Receives vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generates an anti-slip three-dimensional grasping trajectory based on three-dimensional parallax fusion using a parallax compensation algorithm.

[0055] S6: Capture the microscopic deformation features of the adsorption contact surface through a microscopic vision unit, extract the concentric circle distortion features on the contact surface using the Hough transform algorithm, and calculate the microscopic deformation.

[0056] S7: Based on the micro-deformation obtained in S6, perform sub-pixel-level path correction on the 3D grasping trajectory generated in S5;

[0057] S8: Synchronize the corrected 3D grasping trajectory from S7 to the actuator to ensure that the industrial robot performs the grasping task according to the corrected trajectory.

[0058] The beneficial effects of this invention are:

[0059] This invention introduces polarization interferometry imaging technology, combined with precise dielectric feature analysis and multi-field fusion methods. The system can acquire workpiece surface information in real time, accurately identify workpieces of different materials, and then generate an appropriate gripping trajectory based on the identification results. The innovation of this technology lies in its ability to adapt to different workpiece surface shapes, especially when dealing with smooth, transparent or irregularly shaped workpieces, ensuring gripping accuracy and stability.

[0060] This invention captures microscopic deformation features through a microscopic vision unit, achieving sub-pixel-level path correction for 3D grasping trajectories and effectively avoiding grasping failures caused by minute errors. Simultaneously, when the vacuum adsorption force field is activated, the electromagnetic shielding coordination module effectively suppresses electromagnetic interference from surrounding devices through a ring-shaped electromagnetic barrier, ensuring high-quality imaging and path calculation by the microscopic vision unit. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of an intelligent control system for an industrial robot according to an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of an intelligent control method for industrial robots according to an embodiment of the present invention. Detailed Implementation

[0064] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0065] It should be noted that the use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0066] Generally, terms can be understood at least partly from their use in context. For example, depending at least partly on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in a singular sense, or a combination of features, structures, or characteristics in a plural sense. Additionally, the term "based on" can be understood not necessarily to convey an exclusive set of factors, but rather, alternatively, depending at least partly on the context, to allow for the presence of other factors that are not necessarily explicitly described.

[0067] like Figure 1 As shown, an intelligent control system for industrial robots based on machine vision includes a polarization interferometric imaging module, a dielectric feature analysis module, a dynamic clamping force field generation module, a multi-field-of-view fusion module, a sub-pixel compensation module, and an electromagnetic shielding coordination module; wherein:

[0068] Polarization interferometric imaging module: used to acquire the polarization feature spectrum of the workpiece surface by rotating the polarizer array and extract the Stokes parameters of each pixel;

[0069] Dielectric characteristic analysis module: used to receive Stokes parameters, calculate the surface dielectric constant distribution based on Fresnel reflection law, and identify the workpiece material type;

[0070] Dynamic clamping force field generation module: used to match the preset stress-deformation curve according to the identified workpiece material type and generate a vacuum adsorption force field distribution map with pressure gradient;

[0071] Multi-field fusion module: used to receive vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generate anti-slip 3D grasping trajectory through parallax compensation algorithm;

[0072] Subpixel compensation module: Based on the microscopic vision unit to capture the microscopic deformation features of the adsorption contact surface, the module performs subpixel-level path correction on the 3D grasping trajectory generated by the multi-field fusion module.

[0073] Electromagnetic shielding coordination module: used to activate the annular electromagnetic barrier when the vacuum adsorption force field is activated, suppress the influence of electromagnetic interference from surrounding equipment on the microscopic vision unit, and synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator.

[0074] The polarization interferometry imaging module includes a tunable wavelength laser source unit, a four-way polarizer array unit, an image acquisition unit, and a Stokes parameter calculation unit; wherein:

[0075] Tunable wavelength laser source unit: used to emit tunable laser with wavelengths covering 400-1000nm to irradiate the surface of the workpiece;

[0076] Four-way polarizer array unit: By rotating the polarizer array, the polarization direction is switched in 30° increments to acquire four sets of interference images respectively;

[0077] Image acquisition unit: Works synchronously with the four-way polarizer array unit to trigger and receive interference images in each polarization direction, and acquire surface information at different angles;

[0078] Stokes parameter calculation unit: This unit processes four sets of interferometric images and calculates the Stokes parameters for each pixel, including the polarization angle and phase difference. Specifically, the four-axis polarizer array unit rotates at different angles to image different polarization directions. After each rotation of the polarizer array, the image acquisition unit triggers the CCD camera to capture the interferometric image at each angle. These four sets of interferometric images are then input into the Stokes parameter calculation unit, which calculates the Stokes parameters for each pixel using the Jones matrix method. These parameters can be further used for inversion of the surface dielectric constant and identification of the material type. Through the design of the above polarization interferometric imaging module, the polarization feature spectrum of the workpiece surface can be extracted with high precision, and the Stokes parameters for each pixel can be accurately obtained, providing reliable data support for subsequent material analysis and clamping force field generation.

[0079] The Stokes parameter calculation unit includes:

[0080] Image preprocessing: Each set of interferometric images is preprocessed to obtain the polarization information of each pixel. Specifically, image difference processing is performed on the four interferometric images to obtain the light intensity value in each polarization direction, denoted as... ,in, This represents the light intensity when the polarizer array is in the first direction. This represents the light intensity when the polarizer array is in the second direction. This represents the light intensity when the polarizer array is in the third direction. This represents the light intensity when the polarizer array is in the fourth direction.

[0081] Calculate Stokes parameters: Calculate Stokes parameters based on the light intensity values ​​in each polarization direction. The specific formulas include:

[0082] ,in, It represents the total light intensity, which is the sum of the light intensities under all polarization directions;

[0083] ,in, The first component representing the difference in light intensity is used to represent the difference between horizontal and vertical polarization;

[0084] ,in, The second component representing the difference in light intensity is used to represent and Differences in polarization;

[0085] ,in, The third component, representing the difference in light intensity, is used to represent the difference between right-handed and left-handed circular polarization. Through the above steps, the Stokes parameter calculation unit can accurately calculate the Stokes parameters of each pixel, providing high-precision optical information for subsequent workpiece surface material identification.

[0086] The dielectric characteristic analysis module includes a Stokes parameter receiving unit, a dielectric constant calculation unit, and a material identification unit; wherein:

[0087] Stokes parameter receiving unit: used to receive Stokes parameters provided by the polarization interferometric imaging module. ;

[0088] Dielectric constant calculation unit: Used to calculate the surface dielectric constant distribution based on Fresnel reflection law, specifically calculating the polarization angle of each pixel based on the previously received Stokes parameters. and phase difference The formula is:

[0089] ,in, The polarization angle represents the angle indicating the direction of polarization of the light wave.

[0090] ,in, Phase difference represents the difference in phase of a light wave in different polarization directions; The arctangent function is defined as follows: That is, given input value When the time comes, output the corresponding angle value. , making ,in The range of values ​​is Then, the surface dielectric constant is calculated using Fresnel's law of reflection, with the following formula: ,in, The dielectric constant of the workpiece surface. It is the polarization angle. The phase difference is used; this formula is based on the polarization angle and phase difference to invert the dielectric constant of the workpiece surface;

[0091] Material identification unit: used to process the calculated dielectric constant The dielectric constant range of various materials in the preset material database is compared; based on the comparison result, the material type of the workpiece is identified, and the corresponding material label is output; for example, if If the value is less than 3.0, it is identified as a transparent material; if If the value is greater than 3.0, it is identified as an opaque material. Through the above technical solution, the dielectric feature analysis module can accurately calculate the dielectric constant of the workpiece surface and achieve high-precision identification of the workpiece material type based on this parameter. This process provides key material data support for the subsequent dynamic clamping force field generation module.

[0092] The dynamic clamping force field generation module includes a material type receiving unit, a stress-deformation curve matching unit, and a pressure gradient generation unit; wherein:

[0093] Material type receiving unit: used to receive workpiece material type information provided by the dielectric characteristic analysis module;

[0094] Stress-deformation curve matching unit: Used to select the stress-deformation curve corresponding to the received workpiece material type from a preset material database. The stress-deformation curve describes the relationship between stress and deformation of a material under external force. The expression for the stress-deformation curve is: ,in, For the stress of the material, The Young's modulus of the material. The strain of the material is taken as the stress-strain curve; the adsorption pressure distribution in different regions is determined based on the selected stress-strain curve.

[0095] Pressure gradient generation unit: Used to generate a vacuum adsorption force field distribution map with pressure gradient based on the matched stress-deformation curve and material type. During this process, the adsorption pressure... Calculate using the following formula: ,in, Indicates adsorption pressure. For safety reasons, The maximum withstand stress of the material. Based on the dielectric constant calculated for the workpiece surface, the dynamic clamping force field generation module can accurately match the stress-deformation curve according to different workpiece material types through the above technical solution, thereby generating a vacuum adsorption force field distribution map with pressure gradient. This force field distribution map provides precise adsorption force support for industrial robots, ensuring that the workpiece surface can be uniformly stressed during the gripping process, and avoiding stability problems caused by uneven adsorption.

[0096] The multi-field fusion module includes a vacuum adsorption force field distribution map receiving unit, a parallax compensation unit, and a grasping trajectory generation unit; wherein:

[0097] Vacuum adsorption force field distribution map receiving unit: used to receive vacuum adsorption force field distribution maps taken by industrial cameras from at least three directions. The industrial cameras are located at different angles and can simultaneously acquire force field information from different perspectives on the workpiece surface.

[0098] Parallax compensation unit: Used to perform parallax compensation based on the received multi-view vacuum adsorption force field distribution map. By calculating the parallax between each pair of adjacent viewpoints, it eliminates image distortion caused by different camera angles; the parallax compensation formula is: ,in, The target coordinates after compensation. Original coordinates As the refraction compensation factor, Distance from the current viewpoint As the reference distance, The dielectric constant is calculated for the workpiece surface; this formula is used to calculate camera parallax at different angles, and after compensation, more accurate image data is obtained.

[0099] The grasping trajectory generation unit: Based on the compensated multi-view vacuum adsorption force field distribution map, an iterative nearest point algorithm is used to fuse multi-view data to generate an anti-slip 3D grasping trajectory. This iterative nearest point algorithm continuously optimizes the point-to-point matching relationship to finally obtain a 3D grasping trajectory that conforms to the surface morphology of the workpiece, ensuring that no slippage or deviation occurs during the grasping process. Through the above technical solution, the vacuum adsorption force field distribution map obtained by the multi-view camera, combined with the parallax compensation algorithm, can effectively eliminate image errors caused by different viewpoints, thereby generating an accurate anti-slip 3D grasping trajectory. This solution ensures that the robot can maintain stability when grasping in complex environments and avoids grasping failures caused by parallax or uneven adsorption force fields.

[0100] The grasping trajectory generation unit includes:

[0101] Data Preparation: Receive compensated vacuum adsorption force field distribution images from multiple industrial cameras (420, 430, 440), corresponding to image data from different viewpoints. Each image contains a set of three-dimensional coordinate points of the workpiece surface and force field distribution, denoted as... ;

[0102] Preliminary registration: By performing preliminary registration on the data of each viewpoint, the three-dimensional coordinate point sets of multiple viewpoints are connected. Project onto a unified coordinate system; at this point, define the initial transformation matrix. It is used to roughly align data from different viewpoints to obtain a preliminary aligned point set. ;

[0103] Iterative matching: The iterative nearest point algorithm is used to perform matching by minimizing the distance between each pair of corresponding point sets. The goal of the iterative nearest point algorithm is to solve for the optimal rigid transformation matrix. The minimization function is defined as: ,in, It is the target point of the current viewpoint. These are the points to be aligned. It is the transformation matrix of the current iteration step. Represents Euclidean distance. Given the total number of point sets, the transformation matrix is ​​updated in each iteration. This causes the distance error between each pair of points to continuously decrease;

[0104] Trajectory Generation: After iterative matching, a 3D grasping trajectory with anti-slip function is generated by combining the mechanical constraints of the adsorption force field. Specifically, the mechanical constraints of the adsorption force field include the mechanical balance between the adsorption force and the contact surface, and the force distribution; for each point set... The adsorption force field provides a force vector at the contact point. The magnitude and direction of this force vector vary depending on the characteristics of the workpiece surface and material. To prevent slippage, the direction of the contact force must be consistent with the trajectory path during the generation of the gripping trajectory. Combining mechanical constraints, the formula for generating the anti-slip three-dimensional gripping trajectory is as follows: ,in, Indicates at point Adsorption force at the location, It is the adjustment coefficient of adsorption force. It refers to the material properties at that point (such as hardness, elasticity, etc.). It is the displacement gradient at that point, representing the degree of deformation at that point; when generating the trajectory, the force vector at the contact point between the adsorption force field and the workpiece surface is adjusted. This ensures that the force field provides sufficient friction during the grasping process to prevent slippage. Finally, the trajectory generation unit optimizes the trajectory path based on this mechanical constraint to ensure that the robot can firmly grasp the workpiece without slippage during the grasping process.

[0105] The subpixel compensation module includes a microscopic vision unit, a deformation feature extraction unit, and a path correction unit; among which:

[0106] Microscopic vision unit: used to capture the microscopic deformation features of the adsorption contact surface. This microscopic vision unit is equipped with a high-speed microscope camera and a high-precision optical zoom lens, which can perform high-resolution imaging of the contact surface, with an imaging resolution of 2μm / pixel.

[0107] Deformation Feature Extraction Unit: Utilizes the Hough transform algorithm to extract concentric circle distortion features on the adsorption contact surface, thereby determining the microscopic deformation on the contact surface. The formula for calculating the Hough transform is: ,in, Let these be the coordinates of a point on the contact surface. For the angle parameters of the circle, The distance parameter to the circle; by detecting the distortion features on the contact surface, the deformation of the concentric circles on the contact surface is obtained, and the deformation of the contact surface is... Calculated using the following formula: ,in, Let be the radius of the circle after deformation. Let be the radius of the circle before deformation. This represents the microscopic deformation on the contact surface, and the magnitude of the deformation.

[0108] Path correction unit: Based on deformation, performs sub-pixel-level path correction on the 3D grasping trajectory generated by the multi-view fusion module to obtain the corrected 3D grasping trajectory; the path correction calculation formula is: ,in, To correct the displacement, The deformation is obtained through the deformation feature extraction unit; the two-dimensional grasping trajectory is corrected based on the corrected displacement; the above steps capture microscopic deformation features through the microscopic vision unit, extract concentric circle distortion features by combining Hough transform and calculate microscopic deformation. The sub-pixel compensation module can accurately identify the microscopic deformation of the adsorption contact surface and correct the grasping trajectory. This technology ensures that the robot can accurately avoid the trajectory deviation caused by the deformation of the contact surface during the grasping process, which significantly improves the accuracy and stability of grasping.

[0109] The electromagnetic shielding coordination module includes an electromagnetic interference detection unit, an electromagnetic shielding activation unit, and a trajectory synchronization unit; among which:

[0110] Electromagnetic Interference Detection Unit: Used to monitor the electromagnetic environment of the working area, especially the electromagnetic interference level in the area where the microscopic vision unit is located. It detects the magnetic field strength of the working area through a Hall sensor. When the magnetic field strength is detected Exceeding a preset threshold (e.g., When the electromagnetic interference level is determined to be high, electromagnetic shielding measures need to be activated.

[0111] Electromagnetic shielding activation unit: When the electromagnetic interference level in the working area reported by the electromagnetic interference detection unit exceeds the threshold, it is used to activate the annular electromagnetic barrier to generate a reverse magnetic field. The reverse magnetic field is generated according to the following formula: ,in, It is a reverse magnetic field. The ring-shaped electromagnetic barrier generates a reverse magnetic field to suppress external electromagnetic interference, thereby ensuring that the microscopic vision unit can still work stably in a high electromagnetic interference environment and guaranteeing the quality of image acquisition.

[0112] The trajectory synchronization unit synchronizes the 3D grasping trajectory corrected by the sub-pixel compensation module to the actuator for grasping operations. This unit monitors the electromagnetic environment in real time through an electromagnetic interference detection unit, and the electromagnetic shielding activation unit effectively suppresses electromagnetic interference from surrounding equipment by generating a reverse magnetic field, thus ensuring the efficient and stable operation of the microscopic vision unit in complex electromagnetic environments. Furthermore, the trajectory synchronization unit accurately synchronizes the 3D grasping trajectory corrected by the sub-pixel compensation module to the actuator, ensuring the robot can perform precise grasping tasks and avoiding operational errors caused by electromagnetic interference or trajectory deviations.

[0113] like Figure 2 As shown, an intelligent control method for industrial robots based on machine vision, implemented by the aforementioned intelligent control system for industrial robots based on machine vision, includes the following steps:

[0114] S1: Obtain the polarization feature map of the workpiece surface. Specifically, the polarization information of each pixel on the workpiece surface is collected by rotating the polarizer array, and the Stokes parameter of each pixel is calculated.

[0115] S2: Based on the Stokes parameters obtained in S1, the Stokes parameters of each pixel are calculated using the Jones matrix method, thereby obtaining the dielectric constant distribution of the workpiece surface;

[0116] S3: Based on the dielectric constant distribution obtained in S2, identify the material type of the workpiece and select the matching material type according to the preset material library;

[0117] S4: Based on the workpiece material type identified in S3, match the preset stress-deformation curve and generate a vacuum adsorption force field distribution map with pressure gradient.

[0118] S5: Receives vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generates an anti-slip three-dimensional grasping trajectory based on three-dimensional parallax fusion using a parallax compensation algorithm.

[0119] S6: Capture the microscopic deformation features of the adsorption contact surface through a microscopic vision unit, extract the concentric circle distortion features on the contact surface using the Hough transform algorithm, and calculate the microscopic deformation.

[0120] S7: Based on the micro-deformation obtained in S6, perform sub-pixel-level path correction on the 3D grasping trajectory generated in S5;

[0121] S8: Synchronize the corrected 3D grasping trajectory from S7 to the actuator to ensure that the industrial robot performs the grasping task according to the corrected trajectory.

[0122] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0123] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An intelligent control system for industrial robots based on machine vision, characterized in that, It includes a polarization interferometric imaging module, a dielectric feature analysis module, a dynamic clamping force field generation module, a multi-field fusion module, a sub-pixel compensation module, and an electromagnetic shielding coordination module; among which: Polarization interferometric imaging module: used to acquire the polarization feature spectrum of the workpiece surface by rotating the polarizer array and extract the Stokes parameters of each pixel; Dielectric characteristic analysis module: used to receive Stokes parameters, calculate the surface dielectric constant distribution based on Fresnel reflection law, and identify the workpiece material type; Dynamic clamping force field generation module: used to match the preset stress-deformation curve according to the identified workpiece material type and generate a vacuum adsorption force field distribution map with pressure gradient; Multi-field fusion module: used to receive vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generate anti-slip 3D grasping trajectory through parallax compensation algorithm; Subpixel compensation module: Based on the microscopic vision unit to capture the microscopic deformation features of the adsorption contact surface, the module performs subpixel-level path correction on the 3D grasping trajectory generated by the multi-field fusion module. Electromagnetic shielding coordination module: used to activate the annular electromagnetic barrier when the vacuum adsorption force field is activated, suppress the influence of electromagnetic interference from surrounding equipment on the microscopic vision unit, and synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator.

2. The intelligent control system for industrial robots based on machine vision according to claim 1, characterized in that, The polarization interferometry imaging module includes a tunable wavelength laser source unit, a four-way polarizer array unit, an image acquisition unit, and a Stokes parameter calculation unit; wherein: Tunable wavelength laser source unit: used to emit tunable laser with wavelengths covering 400-1000nm to irradiate the surface of the workpiece; Four-way polarizer array unit: By rotating the polarizer array, the polarization direction is switched in 30° increments to acquire four sets of interference images respectively; Image acquisition unit: Works synchronously with the four-way polarizer array unit to trigger and receive interference images in each polarization direction, and acquire surface information at different angles; Stokes parameter calculation unit: used to process four sets of interferometric images and calculate the Stokes parameters for each pixel, including polarization angle and phase difference.

3. The intelligent control system for industrial robots based on machine vision according to claim 2, characterized in that, The Stokes parameter calculation unit includes: Image preprocessing: Each set of interferometric images is preprocessed to obtain the polarization information of each pixel. Specifically, image difference processing is performed on the four interferometric images to obtain the light intensity value in each polarization direction, denoted as... ; Calculate Stokes parameters: Calculate Stokes parameters based on the light intensity values ​​in each polarization direction. The specific formulas include: ,in, It represents the total light intensity, which is the sum of the light intensities under all polarization directions; ,in, The first component representing the difference in light intensity is used to represent the difference between horizontal and vertical polarization; ,in, The second component representing the difference in light intensity is used to represent and Differences in polarization; ,in, The third component, representing the difference in light intensity, is used to represent the difference between right-handed and left-handed circular polarization.

4. The intelligent control system for industrial robots based on machine vision according to claim 1, characterized in that, The dynamic clamping force field generation module includes a material type receiving unit, a stress-deformation curve matching unit, and a pressure gradient generation unit; wherein: Material type receiving unit: used to receive workpiece material type information provided by the dielectric characteristic analysis module; Stress-deformation curve matching unit: Used to select the stress-deformation curve corresponding to the received workpiece material type from a preset material database. The expression for the stress-deformation curve is: ,in, For the stress of the material, The Young's modulus of the material. For the strain of the material; Pressure gradient generation unit: Used to generate a vacuum adsorption force field distribution map with pressure gradient based on the matched stress-deformation curve and material type. During this process, adsorption pressure Calculate using the following formula: ,in, Indicates adsorption pressure. For safety reasons, The maximum withstand stress of the material. The dielectric constant is calculated for the surface of the workpiece.

5. The intelligent control system for industrial robots based on machine vision according to claim 1, characterized in that, The multi-field fusion module includes a vacuum adsorption force field distribution map receiving unit, a parallax compensation unit, and a grasping trajectory generation unit; wherein: Vacuum adsorption force field distribution map receiving unit: used to receive vacuum adsorption force field distribution maps captured by industrial cameras from at least three directions; Parallax compensation unit: used to perform parallax compensation based on the received multi-view vacuum adsorption force field distribution map. By calculating the parallax between each pair of adjacent view images, it eliminates image distortion caused by different camera angles. Grasping trajectory generation unit: Based on the compensated multi-view vacuum adsorption force field distribution map, the iterative nearest point algorithm is used to fuse multi-view data to generate an anti-slip 3D grasping trajectory.

6. The intelligent control system for industrial robots based on machine vision according to claim 5, characterized in that, The grasping trajectory generation unit includes: Data preparation: Receive compensated vacuum adsorption force field distribution images from multiple industrial cameras, each corresponding to image data from different perspectives. Each image contains a set of three-dimensional coordinate points of the workpiece surface and force field distribution, denoted as... ; Preliminary registration: By performing preliminary registration on the data of each viewpoint, the three-dimensional coordinate point sets of multiple viewpoints are connected. Project onto a unified coordinate system; at this point, define the initial transformation matrix. It is used to roughly align data from different viewpoints to obtain a preliminary aligned point set. ; Iterative matching: The iterative nearest point algorithm is used to perform matching by minimizing the distance between each pair of corresponding point sets. The goal of the iterative nearest point algorithm is to solve for the optimal rigid transformation matrix. The minimization function is defined as: ,in, It is the target point of the current viewpoint. These are the points to be aligned. It is the transformation matrix of the current iteration step. Represents Euclidean distance. The total number of points in the set; Trajectory generation: After completing iterative matching, a three-dimensional grasping trajectory with anti-slip function is generated by combining the mechanical constraints of the adsorption force field.

7. The intelligent control system for industrial robots based on machine vision according to claim 1, characterized in that, The sub-pixel compensation module includes a microscopic vision unit, a deformation feature extraction unit, and a path correction unit; wherein: Microscopic vision unit: Used to capture the microscopic deformation features of the adsorption contact surface. This microscopic vision unit is equipped with a high-speed microscope camera and a high-precision optical zoom lens, enabling high-resolution imaging of the contact surface, with an imaging resolution reaching [missing information]. ; Deformation Feature Extraction Unit: Utilizes the Hough transform algorithm to extract concentric circle distortion features on the adsorption contact surface, thereby determining the microscopic deformation on the contact surface. ; Path correction unit: Based on deformation, performs sub-pixel-level path correction on the 3D grasping trajectory generated by the multi-view fusion module to obtain the corrected 3D grasping trajectory; the path correction calculation formula is: ,in, To correct the displacement, The deformation is obtained through the deformation feature extraction unit.

8. The intelligent control system for industrial robots based on machine vision according to claim 1, characterized in that, The electromagnetic shielding coordination module includes an electromagnetic interference detection unit, an electromagnetic shielding activation unit, and a trajectory synchronization unit; wherein: Electromagnetic Interference Detection Unit: Used to monitor the electromagnetic environment of the working area, especially the electromagnetic interference level in the area where the microscopic vision unit is located. It detects the magnetic field strength of the working area through a Hall sensor. When the magnetic field strength is detected When the electromagnetic interference level exceeds the preset threshold, it is determined to be high, and electromagnetic shielding measures need to be activated. Electromagnetic shielding activation unit: When the electromagnetic interference level in the working area reported by the electromagnetic interference detection unit exceeds the threshold, it is used to activate the annular electromagnetic barrier to generate a reverse magnetic field. The reverse magnetic field is generated according to the following formula: ,in, It is a reverse magnetic field. The original magnetic field strength detected in the working area; Trajectory synchronization unit: used to synchronize the 3D grasping trajectory corrected by the sub-pixel compensation module to the actuator to perform the grasping operation.

9. A machine vision-based intelligent control method for industrial robots, implemented by a machine vision-based intelligent control system for industrial robots as described in any one of claims 1-8, characterized in that, Includes the following steps: S1: Obtain the polarization feature map of the workpiece surface. Specifically, the polarization information of each pixel on the workpiece surface is collected by rotating the polarizer array, and the Stokes parameter of each pixel is calculated. S2: Based on the Stokes parameters obtained in S1, the Stokes parameters of each pixel are calculated using the Jones matrix method, thereby obtaining the dielectric constant distribution of the workpiece surface; S3: Based on the dielectric constant distribution obtained in S2, identify the material type of the workpiece and select the matching material type according to the preset material library; S4: Based on the workpiece material type identified in S3, match the preset stress-deformation curve and generate a vacuum adsorption force field distribution map with pressure gradient. S5: Receives vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generates an anti-slip three-dimensional grasping trajectory based on three-dimensional parallax fusion using a parallax compensation algorithm. S6: Capture the microscopic deformation features of the adsorption contact surface through a microscopic vision unit, extract the concentric circle distortion features on the contact surface using the Hough transform algorithm, and calculate the microscopic deformation. S7: Based on the micro-deformation obtained in S6, perform sub-pixel-level path correction on the 3D grasping trajectory generated in S5; S8: Synchronize the corrected 3D grasping trajectory from S7 to the actuator to ensure that the industrial robot performs the grasping task according to the corrected trajectory.

Images