Industrial robot intelligent control system and method based on machine vision

Abstract

The invention relates to the technical field of industrial robots, in particular to an industrial robot intelligent control system and method based on machine vision, and the system comprises a polarization interference imaging module, a dielectric feature 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 the surface of a workpiece; the dielectric characteristic analysis module is used for calculating surface dielectric constant distribution; the dynamic clamping force field generation module is used for generating a vacuum adsorption force field distribution diagram with a pressure gradient; and the multi-view-field fusion module is used for generating an anti-slip three-dimensional grabbing track through a parallax compensation algorithm. According to the invention, by combining the technologies of polarization interference imaging, dielectric characteristic analysis, sub-pixel compensation, electromagnetic shielding and the like, the grabbing precision and stability of the industrial robot in a complex environment are remarkably improved, and meanwhile, external interference is effectively inhibited through electromagnetic shielding.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial robots, and in particular to an industrial robot intelligent control system and method based on machine vision. Background Art

[0002] With the continuous improvement of the level of industrial automation, industrial robots are increasingly used in production lines, especially in scenarios that require high-precision and high-efficiency operations; robot grasping tasks occupy an important position in the fields of electronic manufacturing, automobile assembly and precision machining; traditional industrial robot grasping technology usually relies on the mechanical control of the robotic arm and a simple visual feedback system, and obtains the position and posture information of the workpiece through external sensors to complete the grasping; however, due to the complex environment and the changeable shape of the workpiece, the grasping accuracy and stability of these traditional systems are often limited; especially when dealing with workpieces with smooth surfaces, transparent or irregular shapes, the grasping methods of the existing technology are often difficult to guarantee accuracy and stability.

[0003] The robot grasping methods in the existing technology usually rely on a single visual sensor or mechanical sensor for workpiece identification and path planning, which often cannot meet the requirements of high-precision grasping and adapting to complex working conditions at the same time. For example, traditional visual systems are easily disturbed by changes in lighting, reflections on the workpiece surface or transparent materials, resulting in misjudgment. Mechanical sensors cannot accurately capture tiny deformation information, affecting grasping accuracy. In addition, existing systems often ignore the electromagnetic interference problem in the environment during the grasping path planning process. The microscopic vision unit may be affected by external electromagnetic interference, resulting in a decrease in image acquisition quality. Summary of the invention

[0004] Based on the above objectives, the present invention provides an industrial robot intelligent control system and method based on machine vision.

[0005] An industrial robot intelligent control system based on machine vision includes a polarization interference imaging module, a dielectric characteristic 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; wherein: Polarization interference imaging module: used to obtain the polarization characteristic spectrum of the workpiece surface by rotating the polarizer array and extract the Stokes parameters of each pixel point; Dielectric characteristic analysis module: used to receive Stokes parameters, calculate the surface dielectric constant distribution based on Fresnel reflection law, and identify the material type of the workpiece; 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 diagram with a pressure gradient; Multi-field-of-view fusion module: used to receive vacuum adsorption force field distribution maps from at least three azimuth industrial cameras, and generate anti-slip three-dimensional grasping trajectories through parallax compensation algorithm; Sub-pixel compensation module: Based on the microscopic vision unit, it captures the microscopic deformation characteristics of the adsorption contact surface, and then performs sub-pixel path correction on the three-dimensional grasping trajectory generated by the multi-field fusion module; Electromagnetic shielding coordination module: used to start the annular electromagnetic barrier when the vacuum adsorption force field is activated, suppress the influence of electromagnetic interference of peripheral equipment on the microscopic vision unit, and synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator.

[0006] Optionally, the polarization interference imaging module includes an adjustable wavelength laser source unit, a four-directional polarizer array unit, an image acquisition unit, and a Stokes parameter calculation unit; wherein: Adjustable wavelength laser source unit: used to emit adjustable laser with wavelength 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 at a step angle of 30° to collect four sets of interference images respectively; Image acquisition unit: works synchronously with the four-directional polarizer array unit to trigger and receive the interference image in each polarization direction and obtain surface information at different angles; Stokes parameter calculation unit: used to process four sets of interference images and calculate the Stokes parameters of each pixel, including polarization angle and phase difference.

[0007] Optionally, the Stokes parameter calculation unit includes: Image preprocessing: Preprocess each set of interference images to obtain the polarization information of each pixel. Specifically, the light intensity value in each polarization direction is obtained by performing image difference processing on the four interference images, which is recorded as ; Calculate Stokes parameters: Calculate Stokes parameters based on the light intensity values ​​in each polarization direction , the specific formula includes: ,in, Represents the total light intensity, which is the sum of the light intensities in all polarization directions; ,in, A first component representing the difference in light intensity, used to represent the difference in horizontal and vertical polarization; ,in, The second component representing the light intensity difference 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.

[0008] Optionally, the dielectric characteristic analysis module includes a Stokes parameter receiving unit, a dielectric constant calculation unit and a material identification unit; wherein: Stokes parameter receiving unit: used to receive the Stokes parameters provided by the polarization interference imaging module ; Dielectric constant calculation unit: used to calculate the surface dielectric constant distribution based on the Fresnel reflection law, and specifically calculate the polarization angle of each pixel point according to the Stokes parameters received first and phase difference , and then calculate the surface dielectric constant by Fresnel's reflection law, the formula is: ,in, is the dielectric constant of the workpiece surface, is the polarization angle, is the phase difference; Material identification unit: used to convert the calculated dielectric constant Compare the dielectric constant range of various materials with the preset material database; based on the comparison results, identify the material type of the workpiece and output the corresponding material label.

[0009] 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: 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 material type from the preset material database according to the received workpiece material type. The stress-deformation curve expression is: ,in, is the stress of the material, is the Young's modulus of the material, is the strain of the material; Pressure gradient generation unit: used to generate a vacuum adsorption force field distribution diagram with a pressure gradient according to the matched stress-deformation curve and material type. In this process, the adsorption pressure Calculated according to the following formula: ,in, represents the adsorption pressure, is the safety factor, is the maximum tolerable stress of the material, The dielectric constant calculated for the workpiece surface.

[0010] Optionally, the multi-field fusion module includes a vacuum adsorption force field distribution map receiving unit, a parallax compensation unit and a grasping trajectory generating unit; wherein: A vacuum adsorption force field distribution map receiving unit: used for receiving vacuum adsorption force field distribution maps taken 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, and eliminate image distortion caused by different camera angles by calculating the parallax between each pair of adjacent view images; Grasping trajectory generation unit: Based on the compensated multi-view vacuum adsorption force field distribution map, the iterative closest point algorithm is used to fuse multi-view data to generate an anti-slip three-dimensional grasping trajectory.

[0011] Optionally, the grasping trajectory generating unit includes: Data preparation: Receive compensated vacuum adsorption force field distribution maps from multiple industrial cameras, 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, recorded as ; Preliminary registration: By performing preliminary registration on each viewpoint data, the 3D coordinate point sets of multiple viewpoints are Projected into a unified coordinate system; at this time, define the initial transformation matrix , used to roughly align data from different viewpoints to obtain a preliminary aligned point set ; Iterative matching: Use the iterative closest point algorithm to match by minimizing the distance between each pair of corresponding point sets. The goal of the iterative closest point algorithm is to solve the optimal rigid transformation matrix. , the minimization function is defined as: ,in, is the target point of the current viewpoint, is the point to be aligned, is the transformation matrix of the current iteration step, represents the Euclidean distance, is the total number of point sets; Trajectory generation: After completing the iterative matching, the mechanical constraints of the adsorption force field are combined to generate a three-dimensional grasping trajectory with anti-slip function.

[0012] Optionally, 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 characteristics of the adsorption contact surface. The microscopic vision unit is equipped with a high-speed microscopic 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; Deformation feature extraction unit: Use Hough transform algorithm to extract the concentric circle distortion features on the adsorption contact surface, and then determine the microscopic deformation on the contact surface ; Path correction unit: Perform sub-pixel path correction on the 3D grasping trajectory generated by the multi-field fusion module based on the deformation variable to obtain the corrected 3D grasping trajectory; the path correction calculation formula is: ,in, To correct the displacement, is the deformation amount obtained by the deformation feature extraction unit.

[0013] Optionally, 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, and detect the magnetic field strength in the working area through the Hall sensor , when the magnetic field strength is detected When the preset threshold is exceeded, the electromagnetic interference level is judged to be high, and electromagnetic shielding measures need to be initiated; Electromagnetic shielding activation unit: When the electromagnetic interference detection unit feedbacks that the electromagnetic interference level in the working area 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, is the reverse magnetic field, The raw magnetic field strength detected for the working area; Trajectory synchronization unit: used to synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator to perform the grasping operation.

[0014] A method for intelligent control of an industrial robot based on machine vision is implemented by the above-mentioned intelligent control system of an industrial robot based on machine vision, and comprises the following steps: S1: Obtain the polarization characteristic spectrum of the workpiece surface, specifically by collecting the polarization information of each pixel point on the workpiece surface by rotating the polarizer array, and calculating the Stokes parameter of each pixel point; S2: Based on the Stokes parameters obtained in S1, the Jones matrix method is used to calculate the Stokes parameters of each pixel point, and then the dielectric constant distribution of the workpiece surface is obtained; S3: According to the dielectric constant distribution obtained in S2, the material type of the workpiece is identified, and a matching material type is selected according to a preset material library; S4: According to the workpiece material type identified in S3, a preset stress-deformation curve is matched, and a vacuum adsorption force field distribution diagram with a pressure gradient is generated; S5: receiving vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generating an anti-slip three-dimensional grasping trajectory based on three-dimensional parallax fusion through a parallax compensation algorithm; S6: The microscopic deformation characteristics of the adsorption contact surface are captured by the microscopic vision unit, the concentric circle distortion characteristics on the contact surface are extracted by the Hough transform algorithm, and the microscopic deformation amount is calculated; S7: Based on the microscopic deformation obtained in S6, the three-dimensional grasping trajectory generated in S5 is corrected at the sub-pixel level; S8: Synchronize the corrected three-dimensional grasping trajectory in S7 to the actuator to ensure that the industrial robot performs the grasping task according to the corrected trajectory.

[0015] Beneficial effects of the present invention: The present invention introduces polarization interference imaging technology, combines precise dielectric feature analysis and multi-field fusion methods, and the system can obtain workpiece surface information in real time, accurately identify workpieces of different materials, and then generate an adaptive grasping trajectory based on the recognition results; the innovation of this technology lies in its ability to adapt to different workpiece surface morphologies, especially when processing smooth, transparent or irregularly shaped workpieces, to ensure grasping accuracy and stability.

[0016] The present invention captures microscopic deformation features through a microscopic vision unit, realizes sub-pixel path correction of the three-dimensional grasping trajectory, and effectively avoids grasping failure caused by tiny errors; at the same time, when the vacuum adsorption force field is activated, the electromagnetic shielding collaborative module effectively suppresses the electromagnetic interference of peripheral equipment through the annular electromagnetic barrier, ensuring high-quality imaging and path calculation of the microscopic vision unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of an industrial robot intelligent control system according to an embodiment of the present invention; Figure 2 Schematic diagram of an industrial robot intelligent control method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0020] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0021] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0022] like Figure 1 As shown, an intelligent control system for an industrial robot based on machine vision includes a polarization interference imaging module, a dielectric characteristic 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; wherein: Polarization interference imaging module: used to obtain the polarization characteristic spectrum of the workpiece surface by rotating the polarizer array and extract the Stokes parameters of each pixel point; Dielectric characteristic analysis module: used to receive Stokes parameters, calculate the surface dielectric constant distribution based on Fresnel reflection law, and identify the material type of the workpiece; 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 diagram with a pressure gradient; Multi-field-of-view fusion module: used to receive vacuum adsorption force field distribution maps from at least three azimuth industrial cameras, and generate anti-slip three-dimensional grasping trajectories through parallax compensation algorithm; Sub-pixel compensation module: Based on the microscopic vision unit, it captures the microscopic deformation characteristics of the adsorption contact surface, and then performs sub-pixel path correction on the three-dimensional grasping trajectory generated by the multi-field fusion module; Electromagnetic shielding coordination module: used to start the annular electromagnetic barrier when the vacuum adsorption force field is activated, suppress the influence of electromagnetic interference of peripheral equipment on the microscopic vision unit, and synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator.

[0023] The polarization interference imaging module includes an adjustable wavelength laser source unit, a four-directional polarizer array unit, an image acquisition unit, and a Stokes parameter calculation unit; wherein: Adjustable wavelength laser source unit: used to emit adjustable laser with wavelength 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 at a step angle of 30° to collect four sets of interference images respectively; Image acquisition unit: works synchronously with the four-directional polarizer array unit to trigger and receive the interference image in each polarization direction and obtain surface information at different angles; Stokes parameter calculation unit: used to process four groups of interference images and calculate the Stokes parameters of each pixel point, including polarization angle and phase difference; specifically, the four-way polarizer array unit realizes imaging of different polarization directions by rotating different angles. After each rotation of the polarizer array, the image acquisition unit triggers the CCD camera to perform imaging and obtains the interference image at each angle. The four groups of interference images are input into the Stokes parameter calculation unit, and the Stokes parameters of each pixel point are calculated by the Jones matrix method. These parameters can be further used for the inversion of surface dielectric constant and the identification of material type; through the design of the above-mentioned polarization interference imaging module, the polarization feature map of the workpiece surface can be extracted with high precision, and the Stokes parameters of each pixel point can be accurately obtained, which provides reliable data support for subsequent material analysis and clamping force field generation.

[0024] The Stokes parameter calculation unit includes: Image preprocessing: Preprocess each set of interference images to obtain the polarization information of each pixel. Specifically, the light intensity value in each polarization direction is obtained by performing image difference processing on the four interference images, which is recorded as ,in, is the light intensity value when the polarizer array is in the first direction; is the light intensity value when the polarizer array is in the second direction; is the light intensity value when the polarizer array is in the third direction; is the light intensity value when the polarizer array is in the fourth direction; Calculate Stokes parameters: Calculate Stokes parameters based on the light intensity values ​​in each polarization direction , the specific formula includes: ,in, Represents the total light intensity, which is the sum of the light intensities in all polarization directions; ,in, A first component representing the difference in light intensity, used to represent the difference in horizontal and vertical polarization; ,in, The second component representing the light intensity difference is used to represent and Differences in polarization; ,in, The third component representing the light intensity difference 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 point, providing high-precision optical information for subsequent workpiece surface material identification.

[0025] The dielectric characteristic analysis module includes a Stokes parameter receiving unit, a dielectric constant calculation unit and a material identification unit; wherein: Stokes parameter receiving unit: used to receive the Stokes parameters provided by the polarization interference imaging module ; Dielectric constant calculation unit: used to calculate the surface dielectric constant distribution based on the Fresnel reflection law, and specifically calculate the polarization angle of each pixel point according to the Stokes parameters received first and phase difference , the formula is: ,in, is the polarization angle, which indicates the angle of the polarization direction of the light wave; ,in, is the phase difference, which indicates the phase difference of light waves in different polarization directions; represents the inverse tangent function, defined as , that is, given an input value When , output the corresponding angle value , so that ,in The value range is , and then calculate the surface dielectric constant by Fresnel reflection law, the formula is: ,in, is the dielectric constant of the workpiece surface, is the polarization angle, is the phase difference; this formula inverts the dielectric constant of the workpiece surface based on the polarization angle and phase difference; Material identification unit: used to convert the calculated dielectric constant Compare the dielectric constant range of various materials in the preset material database; based on the comparison results, identify the material type of the workpiece and output the corresponding material label; for example, if If it is less than 3.0, it is identified as a transparent material; if If it 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 realize 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.

[0026] 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 material type from the preset material database according to the received workpiece material type. The stress-deformation curve describes the relationship between stress and deformation of the material under the action of external force. The stress-deformation curve expression is: ,in, is the stress of the material, is the Young's modulus of the material, is the strain of the material; according to the selected stress-deformation curve, the adsorption pressure distribution in different regions is determined; Pressure gradient generation unit: used to generate a vacuum adsorption force field distribution diagram with a pressure gradient according to the matched stress-deformation curve and material type. In this process, the adsorption pressure Calculated according to the following formula: ,in, represents the adsorption pressure, is the safety factor, is the maximum tolerable stress of the material, The dielectric constant is calculated for the workpiece surface. Through the above technical solution, the dynamic clamping force field generation module can accurately match the stress-deformation curve according to the different workpiece material types, and then generate a vacuum adsorption force field distribution map with a pressure gradient. This force field distribution map provides accurate adsorption force support for the industrial robot, ensuring that the workpiece surface can be evenly stressed during the grasping process, avoiding stability problems caused by uneven adsorption.

[0027] 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 taken by industrial cameras from at least three directions. The industrial cameras are located at different angles, and can simultaneously obtain force field information from different viewing angles of the workpiece surface; Parallax compensation unit: used to perform parallax compensation based on the received multi-view vacuum adsorption force field distribution map, and eliminate image distortion caused by different camera angles by calculating the parallax between each pair of adjacent view images; parallax compensation adopts the formula: ,in, is the target coordinate after compensation, is the original coordinate, is the refraction compensation factor, is the current viewpoint distance, is the reference distance, The dielectric constant calculated for the workpiece surface; this formula is used to calculate the camera parallax at different angles and obtain more accurate image data after compensation; Grasping trajectory generation unit: According to the compensated multi-field vacuum adsorption force field distribution map, the iterative closest point algorithm is used to fuse multi-viewpoint data to generate an anti-slip three-dimensional grasping trajectory. The iterative closest point algorithm continuously optimizes the point-to-point matching relationship and finally obtains a three-dimensional grasping trajectory that conforms to the surface morphology of the workpiece, ensuring that no slip or offset occurs during the grasping process; through the above technical solution, the vacuum adsorption force field distribution map obtained by the multi-view camera is combined with the parallax compensation algorithm to effectively eliminate the image errors caused by different viewpoints, thereby generating an accurate anti-slip three-dimensional grasping trajectory. This solution ensures that the robot can maintain stability when grasping in a complex environment, avoiding grasping failures due to parallax or uneven adsorption force field.

[0028] The grasping trajectory generation unit includes: Data preparation: Receive compensated vacuum adsorption force field distribution images from multiple industrial cameras (420, 430, 440), 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, recorded as ; Preliminary registration: By performing preliminary registration on each viewpoint data, the 3D coordinate point sets of multiple viewpoints are Projected into a unified coordinate system; at this time, define the initial transformation matrix , used to roughly align data from different viewpoints to obtain a preliminary aligned point set ; Iterative matching: Use the iterative closest point algorithm to match by minimizing the distance between each pair of corresponding point sets. The goal of the iterative closest point algorithm is to solve the optimal rigid transformation matrix. , the minimization function is defined as: ,in, is the target point of the current viewpoint, is the point to be aligned, is the transformation matrix of the current iteration step, represents the Euclidean distance, is the total number of point sets, and the transformation matrix is ​​updated in each iteration , so that the distance error between each pair of points continues to decrease; Trajectory generation: After iterative matching is completed, a three-dimensional grasping trajectory with anti-slip function is generated in combination with 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 the force vector vary according to the characteristics of the workpiece surface and material; in order to prevent slippage, the direction of the contact force is required to be consistent with the trajectory path when the grasping trajectory is generated; combined with the mechanical constraints, the generation formula of the anti-slip three-dimensional grasping trajectory is: ,in, Indicates at point The adsorption force at is the adjustment coefficient of adsorption force, is the material property at that point (such as hardness, elasticity, etc.), is the displacement gradient at the point, indicating the degree of deformation at the point; when generating the trajectory, the force vector of the contact point between the adsorption force field and the workpiece surface is adjusted , ensuring that the force field provides sufficient friction to prevent slipping during the grasping process; 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 slipping during the grasping process.

[0029] 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 characteristics of the adsorption contact surface. The microscopic vision unit is equipped with a high-speed microscopic 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; Deformation feature extraction unit: Use Hough transform algorithm to extract the concentric circle distortion features on the adsorption contact surface, and then determine the microscopic deformation on the contact surface ; The calculation formula of Hough transform is: ,in, are the coordinates of the point on the contact surface, is the angle parameter of the circle, is the distance parameter from the circle; by detecting the distortion features on the contact surface, the deformation of the concentric circles on the contact surface and the deformation on the contact surface are obtained. Calculated by the following formula: ,in, is the radius of the deformed circle, is the radius of the circle before deformation, Indicates the microscopic deformation on the contact surface, indicating the magnitude of the deformation; Path correction unit: Perform sub-pixel path correction on the 3D grasping trajectory generated by the multi-field fusion module based on the deformation variable to obtain the corrected 3D grasping trajectory; the path correction calculation formula is: ,in, To correct the displacement, is the deformation amount obtained by the deformation feature extraction unit; according to the corrected displacement, the two-dimensional grasping trajectory is corrected; the above steps capture the microscopic deformation features through the microscopic vision unit, extract the concentric circle distortion features and calculate the microscopic deformation amount in combination with the Hough transform, and 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 contact surface deformation during the grasping process, and significantly improves the grasping accuracy and stability.

[0030] 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, and detect the magnetic field strength in the working area through the Hall sensor , when the magnetic field strength is detected Exceeding a preset threshold (e.g. ), the electromagnetic interference level is judged to be high, and electromagnetic shielding measures need to be initiated; Electromagnetic shielding activation unit: When the electromagnetic interference detection unit feedbacks that the electromagnetic interference level in the working area 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, is the reverse magnetic field, The original magnetic field strength detected in the working area. The annular electromagnetic barrier suppresses external electromagnetic interference by generating a reverse magnetic field, thereby ensuring that the microscopic vision unit can still work stably in a high electromagnetic interference environment and guaranteeing the quality of image acquisition. Trajectory synchronization unit: used to synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator to perform the grasping operation; the above unit monitors the electromagnetic environment in real time through the electromagnetic interference detection unit, and the electromagnetic shielding activation unit effectively suppresses the electromagnetic interference of surrounding equipment by generating a reverse magnetic field, thereby ensuring the efficient and stable operation of the microscopic vision unit in a complex electromagnetic environment; in addition, the trajectory synchronization unit can accurately synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator, ensuring that the robot can perform precise grasping tasks and avoiding operational errors caused by electromagnetic interference or trajectory deviation.

[0031] like Figure 2 As shown, an industrial robot intelligent control method based on machine vision is implemented by the above-mentioned industrial robot intelligent control system based on machine vision, and includes the following steps: S1: Obtain the polarization characteristic spectrum of the workpiece surface, specifically by collecting the polarization information of each pixel point on the workpiece surface by rotating the polarizer array, and calculating the Stokes parameter of each pixel point; S2: Based on the Stokes parameters obtained in S1, the Jones matrix method is used to calculate the Stokes parameters of each pixel point, and then the dielectric constant distribution of the workpiece surface is obtained; S3: According to the dielectric constant distribution obtained in S2, the material type of the workpiece is identified, and a matching material type is selected according to a preset material library; S4: According to the workpiece material type identified in S3, a preset stress-deformation curve is matched, and a vacuum adsorption force field distribution diagram with a pressure gradient is generated; S5: receiving vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generating an anti-slip three-dimensional grasping trajectory based on three-dimensional parallax fusion through a parallax compensation algorithm; S6: The microscopic deformation characteristics of the adsorption contact surface are captured by the microscopic vision unit, the concentric circle distortion characteristics on the contact surface are extracted by the Hough transform algorithm, and the microscopic deformation amount is calculated; S7: Based on the microscopic deformation obtained in S6, the three-dimensional grasping trajectory generated in S5 is corrected at the sub-pixel level; S8: Synchronize the corrected three-dimensional grasping trajectory in S7 to the actuator to ensure that the industrial robot performs the grasping task according to the corrected trajectory.

[0032] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0033] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An industrial robot intelligent control system based on machine vision, characterized in that: It includes polarization interference imaging module, dielectric characteristic analysis module, dynamic clamping force field generation module, multi-field fusion module, sub-pixel compensation module and electromagnetic shielding coordination module; among which: Polarization interference imaging module: used to obtain the polarization characteristic spectrum of the workpiece surface by rotating the polarizer array and extract the Stokes parameters of each pixel point; Dielectric characteristic analysis module: used to receive Stokes parameters, calculate the surface dielectric constant distribution based on Fresnel reflection law, and identify the material type of the workpiece; 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 diagram with a pressure gradient; Multi-field-of-view fusion module: used to receive vacuum adsorption force field distribution maps from at least three azimuth industrial cameras, and generate anti-slip three-dimensional grasping trajectories through parallax compensation algorithm; Sub-pixel compensation module: Based on the microscopic vision unit, it captures the microscopic deformation characteristics of the adsorption contact surface, and then performs sub-pixel path correction on the three-dimensional grasping trajectory generated by the multi-field fusion module; Electromagnetic shielding coordination module: used to start the annular electromagnetic barrier when the vacuum adsorption force field is activated, suppress the influence of electromagnetic interference of peripheral equipment on the microscopic vision unit, and synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator.

2. According to claim 1, an industrial robot intelligent control system based on machine vision is characterized in that: The polarization interference imaging module includes an adjustable wavelength laser source unit, a four-directional polarizer array unit, an image acquisition unit and a Stokes parameter calculation unit; wherein: Adjustable wavelength laser source unit: used to emit adjustable laser with wavelength 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 at a step angle of 30° to collect four sets of interference images respectively; Image acquisition unit: works synchronously with the four-directional polarizer array unit to trigger and receive the interference image in each polarization direction and obtain surface information at different angles; Stokes parameter calculation unit: used to process four sets of interference images and calculate the Stokes parameters of each pixel, including polarization angle and phase difference.

3. The intelligent control system of an industrial robot based on machine vision according to claim 2, characterized in that: The Stokes parameter calculation unit comprises: Image preprocessing: Preprocess each set of interference images to obtain the polarization information of each pixel. Specifically, the light intensity value in each polarization direction is obtained by performing image difference processing on the four interference images, which is recorded as ; Calculate Stokes parameters: Calculate Stokes parameters based on the light intensity values ​​in each polarization direction , the specific formula includes: ,in, Represents the total light intensity, which is the sum of the light intensities in all polarization directions; ,in, A first component representing the difference in light intensity, used to represent the difference in horizontal and vertical polarization; ,in, The second component representing the light intensity difference 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 of an industrial robot based on machine vision according to claim 1, characterized in that: The dielectric characteristic analysis module includes a Stokes parameter receiving unit, a dielectric constant calculation unit and a material identification unit; wherein: Stokes parameter receiving unit: used to receive the Stokes parameters provided by the polarization interference imaging module ; Dielectric constant calculation unit: used to calculate the surface dielectric constant distribution based on the Fresnel reflection law, and specifically calculate the polarization angle of each pixel point according to the Stokes parameters received first and phase difference , and then calculate the surface dielectric constant by Fresnel reflection law, the formula is: ,in, is the dielectric constant of the workpiece surface, is the polarization angle, is the phase difference; Material identification unit: used to convert the calculated dielectric constant Compare the dielectric constant range of various materials with the preset material database; based on the comparison results, identify the material type of the workpiece and output the corresponding material label.

5. The intelligent control system of an industrial robot 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 material type from the preset material database according to the received workpiece material type. The stress-deformation curve expression is: ,in, is the stress of the material, is the Young's modulus of the material, is the strain of the material; Pressure gradient generation unit: used to generate a vacuum adsorption force field distribution diagram with a pressure gradient according to the matched stress-deformation curve and material type. In this process, the adsorption pressure Calculated according to the following formula: ,in, represents the adsorption pressure, is the safety factor, is the maximum tolerable stress of the material, The dielectric constant calculated for the workpiece surface.

6. The intelligent control system of an industrial robot based on machine vision according to claim 1, characterized in that: The multi-view fusion module includes a vacuum adsorption force field distribution map receiving unit, a parallax compensation unit and a grasping trajectory generating unit; wherein: A vacuum adsorption force field distribution map receiving unit: used for receiving vacuum adsorption force field distribution maps taken 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, and eliminate image distortion caused by different camera angles by calculating the parallax between each pair of adjacent view images; Grasping trajectory generation unit: Based on the compensated multi-view vacuum adsorption force field distribution map, the iterative closest point algorithm is used to fuse multi-view data to generate an anti-slip three-dimensional grasping trajectory.

7. The intelligent control system of an industrial robot based on machine vision according to claim 6, characterized in that: The grasping trajectory generating unit comprises: Data preparation: Receive compensated vacuum adsorption force field distribution maps from multiple industrial cameras, 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, recorded as ; Preliminary registration: By performing preliminary registration on each viewpoint data, the 3D coordinate point sets of multiple viewpoints are Projected into a unified coordinate system; at this time, define the initial transformation matrix , used to roughly align data from different viewpoints to obtain a preliminary aligned point set ; Iterative matching: Use the iterative closest point algorithm to match by minimizing the distance between each pair of corresponding point sets. The goal of the iterative closest point algorithm is to solve the optimal rigid transformation matrix. , the minimization function is defined as: ,in, is the target point of the current viewpoint, is the point to be aligned, is the transformation matrix of the current iteration step, represents the Euclidean distance, is the total number of point sets; Trajectory generation: After completing the iterative matching, the mechanical constraints of the adsorption force field are combined to generate a three-dimensional grasping trajectory with anti-slip function.

8. The intelligent control system of an industrial robot 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 characteristics of the adsorption contact surface. The microscopic vision unit is equipped with a high-speed microscopic 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; Deformation feature extraction unit: Use Hough transform algorithm to extract the concentric circle distortion features on the adsorption contact surface, and then determine the microscopic deformation on the contact surface ; Path correction unit: Perform sub-pixel path correction on the 3D grasping trajectory generated by the multi-field fusion module based on the deformation variable to obtain the corrected 3D grasping trajectory; the path correction calculation formula is: ,in, To correct the displacement, is the deformation amount obtained by the deformation feature extraction unit.

9. The intelligent control system of an industrial robot 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, and detect the magnetic field strength in the working area through the Hall sensor , when the magnetic field strength is detected When the preset threshold is exceeded, the electromagnetic interference level is judged to be high, and electromagnetic shielding measures need to be initiated; Electromagnetic shielding activation unit: When the electromagnetic interference detection unit feedbacks that the electromagnetic interference level in the working area 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, is the reverse magnetic field, The raw magnetic field strength detected for the working area; Trajectory synchronization unit: used to synchronize the three-dimensional grasping trajectory corrected by the sub-pixel compensation module to the actuator to perform the grasping operation.

10. An industrial robot intelligent control method based on machine vision, implemented by an industrial robot intelligent control system based on machine vision according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Obtain the polarization characteristic spectrum of the workpiece surface, specifically by collecting the polarization information of each pixel point on the workpiece surface by rotating the polarizer array, and calculating the Stokes parameter of each pixel point; S2: Based on the Stokes parameters obtained in S1, the Jones matrix method is used to calculate the Stokes parameters of each pixel point, and then the dielectric constant distribution of the workpiece surface is obtained; S3: According to the dielectric constant distribution obtained in S2, the material type of the workpiece is identified, and a matching material type is selected according to a preset material library; S4: According to the workpiece material type identified in S3, a preset stress-deformation curve is matched, and a vacuum adsorption force field distribution diagram with a pressure gradient is generated; S5: receiving vacuum adsorption force field distribution maps from industrial cameras in at least three directions, and generating an anti-slip three-dimensional grasping trajectory based on three-dimensional parallax fusion through a parallax compensation algorithm; S6: The microscopic deformation characteristics of the adsorption contact surface are captured by the microscopic vision unit, the concentric circle distortion characteristics on the contact surface are extracted by the Hough transform algorithm, and the microscopic deformation amount is calculated; S7: Based on the microscopic deformation obtained in S6, the three-dimensional grasping trajectory generated in S5 is corrected at the sub-pixel level; S8: Synchronize the corrected three-dimensional grasping trajectory in S7 to the actuator to ensure that the industrial robot performs the grasping task according to the corrected trajectory.

Images