Robotic vision coordinated control soldering system and method

Abstract

The application discloses a kind of robot vision collaborative control soldering system and method.The system includes: the first acquisition module forms the molten pool trajectory data when welding trajectory and the molten state data at each welding point in molten pool trajectory;Second acquisition module obtains the spatial posture data of robot through multi-axis sensor, determines the welding response data of welding torch under current spatial posture data when forming welding point;Data processing module determines the welding quality data at each welding point, determines process parameter data;Data correction module determines the first deviation value between welding quality data and preset welding quality, simultaneously determines the second deviation value between process parameter data and preset process parameter, establishes the data mapping between first deviation value and second deviation value, and feeds back to robot to generate correction parameter.The application improves welding quality stability, system adaptability and automation level by multi-module cooperation, dynamic regulation and control and process learning.

Description

Technical Field

[0001] This invention relates to the field of robot vision collaborative control technology, specifically to a robot vision collaborative control soldering system and method. Background Technology

[0002] In modern industrial production, robotic soldering technology is widely used in many fields such as electronics, electrical appliances, and machinery manufacturing. However, existing robotic soldering systems struggle to precisely control soldering quality during the soldering process, leading to frequent soldering defects such as cold solder joints, excessive solder, and incomplete solder joints. The main reason for this is that existing systems cannot comprehensively acquire key information during the soldering process, accurately assess the relationship between soldering quality and current process parameters, or make real-time and effective adjustments to process parameters based on actual soldering conditions. For example, traditional systems rely solely on a single sensor or visual information, failing to comprehensively consider the impact of multiple factors such as the molten pool state, robot posture, and soldering torch movement on soldering quality. This makes it difficult to achieve high-precision, high-quality soldering operations, severely affecting product performance and reliability, and hindering the improvement of industrial production efficiency and product quality. Summary of the Invention

[0003] To address the above problems, this invention provides a robot vision-based collaborative control soldering system and method.

[0004] A first aspect of the present invention provides a robot vision-coordinated soldering system, comprising:

[0005] The first acquisition module is configured to acquire the molten pool trajectory data when the welding trajectory is formed and the melting state data at each weld point in the molten pool trajectory.

[0006] The second acquisition module is configured to acquire the robot’s spatial posture data via a multi-axis sensor and determine the welding response data when the welding torch forms a weld point under the current spatial posture data.

[0007] The data processing module is configured to acquire the molten pool trajectory data and the molten state data to determine the welding quality data at each weld point, and is configured to determine the process parameter data based on the spatial attitude data and the welding response data.

[0008] The data correction module is configured to determine a first deviation value between the welding quality data and the preset welding quality, and simultaneously determine a second deviation value between the process parameter data and the preset process parameters, establish a data mapping between the first deviation value and the second deviation value, and feed back the data mapping to the robot to generate correction parameters.

[0009] As a preferred embodiment, the first acquisition module includes:

[0010] The first imaging unit is configured to acquire visible light images of each solder joint in the molten pool trajectory;

[0011] The second imaging unit is configured to acquire thermal imaging images of each solder joint in the molten pool trajectory via infrared.

[0012] The first acquisition module acquires multiple molten state boundaries at different temperatures in the thermal imaging image, maps the molten state boundaries to the corresponding visible light image, and generates molten state data including molten state boundaries and physical morphological boundaries;

[0013] The first acquisition module acquires molten state data from multiple consecutive weld points in the thermal imaging image to generate molten pool trajectory data.

[0014] As a preferred embodiment, the second acquisition module includes:

[0015] The first sensing unit is configured to acquire the acceleration and velocity parameters of each axis and each arm of the robot;

[0016] The second sensing unit is configured to acquire the acceleration, velocity parameters, and surface pressure parameters of the welding torch;

[0017] The arc sensing unit is configured to determine the current and voltage parameters of the surface-induced arc when the welding torch performs the operation;

[0018] The second acquisition module obtains the spatial attitude data by determining the relative position and attitude between each arm and each axis of the robot at its current position based on the first sensing unit.

[0019] The second acquisition module determines the arc intensity and pressure contact parameters on the surface of the welding torch when the welding torch forms the current weld point, as well as the first motion parameters of the robot and the second motion parameters of the welding torch, based on the second sensing unit and the arc sensing unit. The first motion parameters include the acceleration and velocity parameters of each axis and each arm of the robot, and the second motion parameters are the acceleration and velocity parameters of the welding torch. The module determines the transmission error between the first motion parameters and the second motion parameters under the current spatial posture data, and obtains the welding response data.

[0020] As a preferred embodiment, when the data processing module determines the welding quality data, it is configured to perform the following steps:

[0021] Acquire multiple consecutive frames of visible light images and corresponding thermal imaging data at the first solder joint;

[0022] Acquire the first frame that forms the first physical shape boundary from multiple consecutive frames, and determine the first melt shape boundary corresponding to the first physical shape boundary;

[0023] The first thermal deformation error is determined by comparing the distance between the boundary of the first physical state and the boundary of the first molten state.

[0024] Acquire the second frame that forms the second physical deformation boundary from multiple consecutive frames, determine the second molten state boundary corresponding to the second physical state boundary, determine the distance between the first molten state boundary and the second molten state boundary, determine the first heat conduction rate, determine the distance between the second physical deformation boundary and the second molten state boundary, determine the second thermal deformation error, and establish a mapping relationship between the first heat conduction rate and the second thermal deformation error;

[0025] The third frame in a series of consecutive frames is used to form the third physical form boundary. The third melting form boundary at this time is determined, the distance from the third melting deformation boundary to the third physical form boundary is determined, and the third thermal deformation error is determined. The third physical form boundary is the limit boundary when no new physical form boundary is formed.

[0026] Acquire visible light images and corresponding thermal imaging data of multiple consecutive frames of adjacent solder joints of the first solder joint;

[0027] When obtaining the first physical boundary formed by the adjacent weld points and calculating the first thermal deformation error;

[0028] The molten state boundary of the first weld point is determined, and the molten state boundary is used as the interference interval of the first thermal deformation error of the adjacent weld points to obtain the fourth thermal deformation error.

[0029] The welding quality data are obtained based on the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error.

[0030] As a preferred embodiment, the data correction module is configured to perform the following steps:

[0031] Extract the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error to construct a welding quality feature group;

[0032] After assigning weights to each item in the welding quality feature group and the preset welding quality benchmark group, normalization is performed, and the difference between the two is calculated to obtain the offset of each item, generating the first deviation value.

[0033] After assigning weights to the shaft acceleration, welding torch speed and arc current of the process parameter data and the preset process parameters, the data is normalized, and the difference between the two is calculated to generate a second deviation value.

[0034] Establish a correlation mapping between the first deviation value and the second deviation value. The correlation mapping includes a first component and a second component. The first component is the overall influence rate of the process parameters on the welding quality, and the second component is the individual influence rate of each process parameter on each parameter in the welding quality feature group.

[0035] As a preferred method, the construction of the association mapping includes the following steps:

[0036] The first thermal deformation error is mapped to a spatial coordinate system to generate a first position weighting factor characterizing the stability of the solder joint position.

[0037] Based on the correlation between the first heat conduction rate and the second heat deformation error, a dynamic correction matrix for heat influence with two control dimensions is constructed. The second heat deformation error is linearly converted into a welding torch moving speed compensation value, and the first heat conduction rate is mapped into an arc voltage gradient adjustment amount.

[0038] By integrating the first position weighting factor and the dynamic correction matrix of thermal influence, the overall influence rate of process parameters on welding quality is generated.

[0039] The third thermal deformation error is normalized to generate a boundary solidification factor that characterizes the degree of solidification at the weld pool boundary. The fourth thermal deformation error is converted into an interference compensation coefficient between adjacent weld points. The interference compensation coefficient is correlated with the boundary solidification factor to quantify the single influence rate of process parameters on individual welding defects.

[0040] As a preferred embodiment, the data correction module is further configured to perform the following steps:

[0041] The first positional weighting factor is spatially coupled with the output of the thermal effect dynamic correction matrix to generate the global influence intensity of process parameters on welding quality.

[0042] The global influence intensity is calibrated by combining the degree of solidification at the molten pool boundary to obtain the overall influence rate value within a preset range;

[0043] Based on the influence of the real-time adjustment of the welding torch speed compensation value and the arc voltage adjustment on each of the aforementioned thermal deformation errors, dynamic influence weights are assigned to them respectively.

[0044] Each thermal deformation error is weighted and superimposed with the interference compensation coefficient to quantify the targeted influence intensity of each process parameter on characteristic welding defects.

[0045] When the overall impact rate exceeds the first preset value, the arc voltage gradient adjustment amount is increased proportionally to the overall impact rate value, and the compensation range of the welding torch moving speed is expanded to the preset range of each thermal deformation error based on the third thermal deformation error value.

[0046] When any single influence rate exceeds the first preset value, compensation of the corresponding process parameters is performed until the corresponding thermal deformation error is within the preset range.

[0047] A second aspect of the present invention provides a robot vision-coordinated soldering method, comprising the following steps:

[0048] S1. Obtain molten pool trajectory data and solder joint melting state data, including:

[0049] Melt state boundary data is generated by simultaneously acquiring visible light and infrared imaging;

[0050] Merging molten state data from continuous weld points generates molten pool trajectory data;

[0051] S2. The robot's spatial posture data is acquired through multi-axis sensors, and the transmission error is calculated by combining the welding torch surface pressure, arc parameters and motion parameters to obtain welding response data.

[0052] S3. Calculate four thermal deformation errors based on the molten pool trajectory data and molten state data to generate welding quality data;

[0053] S4. Based on spatial attitude data and welding response data, analyze the axial acceleration fluctuation, welding torch speed change rate and arc current deviation to generate process parameter data;

[0054] S5. Perform double deviation value calculation:

[0055] The first deviation value is generated by weighted normalization of the thermal deformation error;

[0056] The process parameters are weighted and normalized to generate a second deviation value;

[0057] S6. Construct association mapping:

[0058] The first thermal deformation error is converted into a position weighting factor;

[0059] A thermal effect correction matrix is ​​generated based on the thermal conduction rate and the second thermal deformation error, and the output speed compensation value and voltage regulation amount are then used.

[0060] The overall influence rate is generated by fusing the location weighting factor and the correction matrix;

[0061] The third and fourth thermal deformation errors are converted into boundary solidification factors and interference compensation coefficients to generate a single influence rate.

[0062] S7, Response Correction Control:

[0063] When the overall impact rate is greater than the first preset value, the voltage regulation intensity is increased and the speed compensation range is expanded.

[0064] When the single influence rate is greater than the second preset value, compensation is performed for the target thermal deformation error.

[0065] Compared with the prior art, the present invention has the following advantages:

[0066] Through the first acquisition module and the second acquisition module, this invention can comprehensively acquire key information such as molten pool trajectory data, molten state data, robot spatial posture data and welding response data during the welding process, providing rich data support for accurately evaluating welding quality and adjusting process parameters.

[0067] The data correction module establishes a correlation mapping between welding quality data and process parameter data, and corrects the deviation value. It can adjust the process parameters in real time, keeping the welding process in the optimal state at all times, effectively improving welding quality, reducing welding defects, and improving product performance and reliability. Attached Figure Description

[0068] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0069] Figure 1 This is a structural block diagram of the system provided in the embodiments of the present invention. Detailed Implementation

[0070] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] This disclosure provides a robot vision-assisted collaborative control system for soldering, such as... Figure 1 As shown, it includes a first acquisition module, a second acquisition module, a data processing module, and a data correction module.

[0072] The first acquisition module is configured to acquire the molten pool trajectory data during the formation of the welding trajectory, as well as the molten state data at each weld point within the molten pool trajectory. This module specifically includes:

[0073] First imaging unit: configured to acquire visible light images of each solder joint in the molten pool trajectory;

[0074] Second imaging unit: configured to acquire thermal imaging images of each solder joint in the molten pool trajectory via infrared;

[0075] The first acquisition module acquires multiple molten state boundaries at different temperatures in the thermal imaging image, maps the molten state boundaries to the corresponding visible light image, and generates molten state data including molten state boundaries and physical morphological boundaries; at the same time, it acquires molten state data in multiple consecutive weld points in the thermal imaging image to generate molten pool trajectory data.

[0076] The second acquisition module is configured to acquire the robot's spatial posture data via multi-axis sensors and determine the welding response data when the welding torch forms a weld point under the current spatial posture data. This module specifically includes:

[0077] First sensing unit: configured to acquire acceleration and velocity parameters of each axis and each arm of the robot;

[0078] The second sensing unit is configured to acquire the acceleration, velocity parameters, and surface pressure parameters of the welding torch.

[0079] Arc sensing unit: configured to determine the current and voltage parameters of the surface-generated arc when the welding torch performs the operation;

[0080] The second acquisition module obtains the spatial attitude data by determining the relative position and attitude between each arm and axis of the robot at its current position based on the first sensing unit; it also obtains the arc intensity and pressure contact parameters on the surface of the welding torch when the welding torch forms the current weld point based on the second sensing unit and the arc sensing unit, as well as the robot's first motion parameters and the welding torch's second motion parameters. The first motion parameters include the acceleration and velocity parameters of each axis and arm of the robot, and the second motion parameters are the acceleration and velocity parameters of the welding torch. The module then determines the transmission error between the first motion parameters and the second motion parameters under the current spatial attitude data to obtain the welding response data.

[0081] The data processing module is configured to acquire the molten pool trajectory data and the molten state data to determine the welding quality data at each weld point, and is configured to determine process parameter data based on the spatial attitude data and the welding response data. When determining the welding quality data, the following steps are performed:

[0082] Acquire multiple consecutive frames of visible light images and corresponding thermal imaging data at the first solder joint;

[0083] Acquire the first frame that forms the first physical shape boundary from multiple consecutive frames, and determine the first melt shape boundary corresponding to the first physical shape boundary;

[0084] The first thermal deformation error is determined by comparing the distance between the boundary of the first physical state and the boundary of the first molten state.

[0085] Acquire the second frame that forms the second physical deformation boundary from multiple consecutive frames, determine the second molten state boundary corresponding to the second physical state boundary, determine the distance between the first molten state boundary and the second molten state boundary, determine the first heat conduction rate, determine the distance between the second physical deformation boundary and the second molten state boundary, determine the second thermal deformation error, and establish a mapping relationship between the first heat conduction rate and the second thermal deformation error;

[0086] The third frame in a series of consecutive frames is used to form the third physical form boundary. The third melting form boundary at this time is determined, the distance from the third melting deformation boundary to the third physical form boundary is determined, and the third thermal deformation error is determined. The third physical form boundary is the limit boundary when no new physical form boundary is formed.

[0087] Acquire visible light images and corresponding thermal imaging data of multiple consecutive frames of adjacent solder joints of the first solder joint;

[0088] When obtaining the first physical boundary formed by the adjacent weld points and calculating the first thermal deformation error;

[0089] The molten state boundary of the first weld point is determined, and the molten state boundary is used as the interference interval of the first thermal deformation error of the adjacent weld points to obtain the fourth thermal deformation error.

[0090] The welding quality data are obtained based on the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error.

[0091] For the first thermal deformation error, assuming that the coordinates of a point on the first physical shape boundary in the image coordinate system are... The coordinates of the corresponding points on the boundary of the first molten state are: Then the first thermal deformation error It can be determined using Euclidean geometry formulas:

[0092]

[0093] In the second calculation of heat deformation error, the distance between the boundary of the first molten state and the boundary of the second molten state is determined. and the distance from the first physical deformation boundary to the second physical deformation boundary. Suppose that during a certain period of time The change from the boundary of the first physical state to the boundary of the second physical state is completed internally, and the thermal conductivity is... It can be done through the formula:

[0094]

[0095] Calculate the second thermal deformation error. Similarly, the calculation is based on the distance between corresponding points of the physical boundary and the molten boundary, using a method similar to the first thermal deformation error. When calculating the fourth thermal deformation error, after determining the molten boundary of the first solder joint, for adjacent solder joints forming the first physical boundary, the thermal deformation within the interference area between the adjacent solder joint and the molten boundary of the first solder joint is calculated. It is assumed that the shortest distance between a point on the first physical boundary of an adjacent solder joint and the molten boundary of the first solder joint is... Considering the influence of heat conduction within the interference region, the fourth thermal deformation error It can be calculated using a weighted average method, with the weights based on distance. And the temperature gradient within that region is determined, if the distance The smaller the value, the greater the weight. The specific formula can be expressed as:

[0096]

[0097] in, As weight, This represents the thermal deformation at the corresponding point.

[0098] In this embodiment of the disclosure, the determination of the third physical morphological boundary includes: when the rate of change of the physical deformation boundary is less than 0.05 mm / frame in three consecutive frames, and the temperature gradient of the molten state decreases to within ±10% of the solidus temperature of the material, it is determined that the limit boundary has been reached. In addition, the calculation of the third thermal deformation error is similar to that of the first and second thermal deformation errors, and will not be repeated in this embodiment of the disclosure.

[0099] When the data processing module determines process parameters based on spatial attitude data and welding response data, it takes the calculation of axis acceleration fluctuation as an example. Assume that during a certain motion phase of the robot, the theoretical accelerations of each axis are as follows: The acceleration actually obtained through the first sensing unit is Then the axial acceleration fluctuation is:

[0100]

[0101] In the calculation of the rate of change of welding torch speed, it is assumed that the welding torch changes speed at two adjacent sampling times. and The speeds are respectively and The rate of change of welding torch speed is _____.

[0102] .

[0103] When calculating the arc current deviation value, if the preset arc current is... The actual arc current obtained through the arc sensing unit is The arc current deviation value is:

[0104] .

[0105] The data correction module is configured to determine a first deviation value between the welding quality data and the preset welding quality, and simultaneously determine a second deviation value between the process parameter data and the preset process parameters. A data mapping between the first and second deviation values ​​is established, and the data mapping is fed back to the robot to generate correction parameters. Specifically, the following steps are performed:

[0106] Extract the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error to construct a welding quality feature group;

[0107] After assigning weights to each item in the welding quality feature group and the preset welding quality benchmark group, normalization is performed, and the difference between the two is calculated to obtain the offset of each item, generating the first deviation value.

[0108] After assigning weights to the shaft acceleration, welding torch speed and arc current of the process parameter data and the preset process parameters, the data is normalized, and the difference between the two is calculated to generate a second deviation value.

[0109] Establish a correlation mapping between the first deviation value and the second deviation value. The construction of the correlation mapping includes the following steps:

[0110] When mapping the first thermal deformation error to a spatial coordinate system to generate the first position weighting factor, a three-dimensional spatial coordinate system is established, with the center of the solder joint as the origin. shaft and The axes correspond to two directions of the welding plane, respectively. The axis is perpendicular to the welding plane. Each component of the first thermal deformation error is mapped onto this coordinate system. By calculating the magnitude and direction of the error vector and combining it with a preset weighting function, a first position weighting factor is generated. For example, the weighting function can be non-linearly adjusted according to the magnitude of the error; if the error is small, the weight increases less; if the error is large, the weight increases more. When constructing the dynamic correction matrix for thermal effects, it is based on the first thermal conduction rate. With the second thermal deformation error The correlation relationship, let the correction matrix be... The elements are determined based on the changing trends of both. For example, when the heat conduction rate increases, in order to reduce the second thermal deformation error, the welding torch moving speed compensation value needs to be increased accordingly. At this time, the value of the corresponding element in the matrix will increase. At the same time, according to the relationship between the heat conduction rate and the arc voltage, the arc voltage gradient adjustment amount is adjusted, causing the relevant elements in the matrix to change. When quantifying the single influence rate of process parameters on a single welding defect, for the boundary solidification factor and the interference compensation coefficient, a regression model between the two and the process parameters is established through experimental data. For example, for the boundary solidification factor transformed from the third thermal deformation error, the boundary solidification degree data under different combinations of axial acceleration, welding torch speed, and arc current are obtained through experiments. Using multiple linear regression or nonlinear regression methods, a functional relationship between the boundary solidification factor and the process parameters is established, thereby quantifying the influence of process parameters on the degree of boundary solidification, i.e., the single influence rate. In this embodiment, the method of calculating the influence rate of one parameter on another using multiple linear regression or nonlinear regression methods will not be described in detail.

[0111] The first thermal deformation error is mapped to a spatial coordinate system to generate a first position weighting factor characterizing the stability of the weld point position.

[0112] Based on the correlation between the first heat conduction rate and the second heat deformation error, a dynamic correction matrix for thermal influence with two control dimensions is constructed. The second heat deformation error is linearly converted into a welding torch moving speed compensation value, and the first heat conduction rate is mapped into an arc voltage gradient adjustment amount.

[0113] By integrating the first position weighting factor and the dynamic correction matrix of thermal influence, the overall influence rate of process parameters on welding quality is generated.

[0114] The third thermal deformation error is normalized to generate a boundary solidification factor that characterizes the degree of solidification at the weld pool boundary. The fourth thermal deformation error is converted into an interference compensation coefficient between adjacent weld points. The interference compensation coefficient is correlated with the boundary solidification factor to quantify the single influence rate of process parameters on individual welding defects.

[0115] The data correction module is also configured to perform the following steps:

[0116] When generating the global influence intensity by spatially coupling the output of the first position weighting factor with the dynamic correction matrix of thermal influence, a combination of vector dot product and cross product is used. Let the first position weighting factor be a vector. The output of the thermal effect dynamic correction matrix can be expressed as a vector. The global impact strength can be expressed by the formula.

[0117] ,

[0118] Calculation, where and These are coefficients adjusted based on actual welding conditions. When assigning dynamic influence weights to the welding torch speed compensation value and arc voltage adjustment amount, the rate of change of thermal deformation error monitored in real-time during the welding process is considered. For example, if the rate of change of a certain thermal deformation error is large, it indicates that the impact of this error on welding quality is more significant at the current stage, and the dynamic influence weight of the welding torch speed compensation value or arc voltage adjustment amount related to this error is increased accordingly. When the overall influence rate exceeds a first preset value, the characteristics of the welding material and welding process requirements are considered when expanding the compensation range of the welding torch movement speed. For example, for solder materials with different melting points, the expansion range of the compensation range is determined based on their thermal conductivity and melting point temperature. If the solder material has a high melting point and poor thermal conductivity, the compensation range expansion will be relatively small to avoid over-compensation leading to welding defects. This is a common technique in the field, based on the characteristics of the soldering material and process requirements, and can be determined according to industry standards. In this embodiment, the purpose is to determine the degree of adjustment based on the overall impact rate and thus determine the compensation range. Therefore, the characteristics of the soldering material and process requirements are not the focus of this embodiment. Those skilled in the art should make appropriate adjustments based on the application of this embodiment to different soldering materials or soldering process requirements, which will not be elaborated further here.

[0119] Then, the first position weighting factor and the output of the thermal effect dynamic correction matrix are spatially coupled to generate the global influence intensity of the process parameters on the welding quality. The global influence intensity is calibrated in combination with the solidification degree of the molten pool boundary to obtain the overall influence rate value within the preset range.

[0120] Based on the influence of the real-time adjustment of the welding torch speed compensation value and the arc voltage adjustment on each of the aforementioned thermal deformation errors, dynamic influence weights are assigned to them respectively.

[0121] Each thermal deformation error is weighted and superimposed with the interference compensation coefficient to quantify the targeted influence intensity of each process parameter on characteristic welding defects.

[0122] When the overall impact rate exceeds the first preset value, the arc voltage gradient adjustment amount is increased proportionally to the overall impact rate value, and the compensation range of the welding torch moving speed is expanded to the preset range of each thermal deformation error based on the third thermal deformation error value.

[0123] When any single influence rate exceeds the first preset value, compensation is performed on the corresponding process parameters until the corresponding thermal deformation error is within the preset range.

[0124] Specifically, this disclosure describes the setup of each module. In actual welding operations, the first acquisition module of this disclosure has a first imaging unit and a second imaging unit installed at appropriate positions on the robot to ensure clear acquisition of visible light and thermal images of each solder joint along the molten pool trajectory. The first imaging unit uses a high-resolution visible light camera to capture information such as the appearance and shape of the solder joint; the second imaging unit uses a high-precision infrared thermal imager to acquire the temperature distribution of the solder joint. When the robot performs soldering operations, the first and second imaging units work synchronously to acquire images of each solder joint. The first acquisition module processes the acquired thermal images, identifies multiple molten state boundaries at different temperatures, maps them to the corresponding visible light images, and generates molten state data containing molten state boundaries and physical morphological boundaries. Simultaneously, by analyzing and integrating the molten state data in the thermal images of multiple consecutive solder joints, molten pool trajectory data is generated. During image acquisition, the shooting frequency of the imaging unit can be dynamically adjusted according to the welding speed and solder joint spacing to ensure that each solder joint is clearly captured.

[0125] In this embodiment of the second acquisition module, the first sensing unit, the second sensing unit, and the arc sensing unit are respectively installed at corresponding parts of the robot. The accelerometer and velocity sensor of the first sensing unit are installed at key locations on each axis and arm of the robot, such as joint connections and arm ends, to ensure accurate acquisition of motion parameters for each axis and arm. During installation, it is necessary to ensure that the measurement direction of the sensor is consistent with the motion direction of the axis or arm, and to perform calibration to eliminate installation errors. The sampling frequency of the sensor is set to no less than 100Hz to meet the requirements for real-time acquisition of motion parameters.

[0126] When installing the accelerometer, velocity sensor, and pressure sensor on the welding torch, the structure and motion characteristics of the welding torch must be considered. The accelerometer and velocity sensors are installed near the center of gravity of the welding torch to accurately measure its overall motion. The pressure sensor is installed at the point of contact between the welding torch and the workpiece, such as the nozzle, to ensure real-time monitoring of the pressure on the torch surface. The pressure sensor's range is selected according to the welding process requirements, generally between 0 and 10 N, with an accuracy of no less than 0.1 N.

[0127] The electrical connection between the arc sensing unit and the welding torch must ensure good conductivity and stability to avoid signal interference. Its sampling frequency is synchronized with the first and second sensing units to ensure that the acquired arc current and voltage parameters match the motion parameters of the robot and the welding torch in time. During robot operation, each sensing unit collects data in real time. Based on this data, the second acquisition module calculates the robot's current spatial posture data and determines the welding response data when the welding torch forms the current weld point, including arc intensity, welding torch surface pressure contact parameters, and transmission error between the robot and welding torch motion parameters.

[0128] After the data processing module receives the molten pool trajectory data and molten state data from the first acquisition module, it calculates the welding quality data for each weld point. Taking a certain weld point as an example, it acquires multiple consecutive frames of visible light images and corresponding thermal imaging data at that weld point. From these image data, it finds the first frame image that forms the first physical shape boundary, determines the corresponding first molten shape boundary, and obtains the first thermal deformation error by calculating the distance between them. In the same way, it acquires the second frame image that forms the second physical deformation boundary, determines the second molten shape boundary, calculates the relevant distance and rate, obtains the second thermal deformation error, and establishes a mapping relationship between the first heat conduction rate and the second thermal deformation error. Then, it acquires the third frame image that forms the third physical shape boundary, determines the third molten shape boundary, and calculates the third thermal deformation error. Next, it acquires multiple consecutive frames of visible light images and corresponding thermal imaging data of adjacent weld points. When calculating the first thermal deformation error of adjacent weld points, it uses the molten shape boundary of the weld point as the interference interval of the first thermal deformation error of the adjacent weld points, thereby obtaining the fourth thermal deformation error. Finally, by combining the first, second, third, and fourth thermal deformation errors, the welding quality data of the weld point is obtained.

[0129] Simultaneously, the data processing module analyzes and calculates the shaft acceleration fluctuation, welding torch speed change rate, and arc current deviation based on the spatial attitude data and welding response data transmitted from the second acquisition module, generating process parameter data. During the calculation process, digital signal processing technology can be used to filter the data collected by the sensors, removing noise interference and improving the accuracy and reliability of the data.

[0130] The data correction module first extracts the calculated first, second, third, and fourth thermal deformation errors to construct a welding quality feature group. It then assigns weights to each item in the welding quality feature group and the preset welding quality benchmark group according to their importance to welding quality, and performs normalization. The offset of each item is calculated by determining the difference between the two groups, thus generating the first deviation value. Similarly, the process parameter data and the preset process parameters (axis acceleration, welding torch speed, and arc current) are weighted, normalized, and their differences are calculated to generate the second deviation value.

[0131] When constructing the correlation mapping, the first thermal deformation error is mapped to the spatial coordinate system to generate a first position weighting factor characterizing the stability of the weld point position. Based on the correlation between the first heat conduction rate and the second thermal deformation error, a dynamic correction matrix for thermal effects is constructed. The second thermal deformation error is linearly converted into a welding torch movement speed compensation value, and the first heat conduction rate is mapped into an arc voltage gradient adjustment amount. The first position weighting factor and the dynamic correction matrix for thermal effects are fused to obtain the overall influence rate of process parameters on welding quality. The third thermal deformation error is normalized to generate a boundary solidification factor characterizing the degree of solidification at the weld pool boundary. The fourth thermal deformation error is converted into an interference compensation coefficient between adjacent weld points. The interference compensation coefficient and the boundary solidification factor are correlated to quantify the single influence rate of process parameters on individual welding defects.

[0132] Subsequently, the output of the first position weighting factor and the dynamic correction matrix for thermal impact are spatially coupled to generate the global influence intensity of process parameters on welding quality. This global influence intensity is then calibrated based on the solidification degree of the molten pool boundary to obtain the overall influence rate value within a preset range. Dynamic influence weights are assigned to each thermal deformation error based on the real-time adjustment of the welding torch speed compensation value and the arc voltage adjustment value. Each thermal deformation error is then superimposed with its weight and the interference compensation coefficient to quantify the targeted influence intensity of each process parameter on characteristic welding defects. When the overall influence rate exceeds the first preset value, the arc voltage gradient adjustment value is increased proportionally to the overall influence rate value, and the compensation range of the welding torch movement speed is expanded to ensure that each thermal deformation error is within the preset range, using the third thermal deformation error value as a benchmark. When any single influence rate exceeds the first preset value, compensation of the corresponding process parameter is performed until the corresponding thermal deformation error is within the preset range, and the correction parameters are fed back to the robot to adjust the welding operation. In practical applications, the preset values ​​can be flexibly adjusted according to different welding materials and process requirements to achieve the best welding results.

[0133] Example 2

[0134] This disclosure provides a method suitable for implementing Embodiment 1, comprising the following steps:

[0135] Includes the following steps:

[0136] S1: Acquire molten pool trajectory data and solder joint melting state data, including: synchronously acquiring molten state boundaries through visible light and infrared imaging to generate molten state data; fusing molten state data of continuous solder joints to generate molten pool trajectory data;

[0137] S2: The robot's spatial posture data is acquired through multi-axis sensors, and the transmission error is calculated by combining the welding torch surface pressure, arc parameters and motion parameters to obtain welding response data;

[0138] S3: Calculate four thermal deformation errors based on molten pool trajectory data and molten state data to generate welding quality data;

[0139] S4: Based on spatial attitude data and welding response data, analyze the axial acceleration fluctuation, welding torch speed change rate and arc current deviation to generate process parameter data;

[0140] S5: Perform dual deviation value calculation: generate the first deviation value by weighted normalization of the hot deformation error; generate the second deviation value by weighted normalization of the process parameters;

[0141] S6: Constructing the correlation mapping: Convert the first thermal deformation error into a position weighting factor; generate a thermal influence correction matrix based on the heat conduction rate and the second thermal deformation error, and output the speed compensation value and voltage adjustment amount; fuse the position weighting factor and the correction matrix to generate the overall influence rate; convert the third and fourth thermal deformation errors into boundary solidification factors and interference compensation coefficients to generate a single influence rate;

[0142] S7: Response Correction Control: When the overall impact rate is greater than the first preset value, the voltage regulation intensity is increased and the speed compensation range is expanded; when the single impact rate is greater than the second preset value, compensation is performed for the target thermal deformation error.

[0143] Specifically, in actual implementation, each step, S1: Before the robot begins soldering, the first and second imaging units are activated and put into operation. The imaging parameters, such as resolution, frame rate, and exposure time, are set according to the requirements of the soldering task. When the robot is soldering, the two imaging units simultaneously acquire visible light and thermal images of each solder joint along the molten pool trajectory. The first acquisition module processes the acquired images in real time, identifying the molten state boundaries in the thermal images using image recognition algorithms and mapping them to the visible light images to generate molten state data containing both molten state boundaries and physical morphological boundaries. Simultaneously, the molten state data of consecutive solder joints are fused to generate molten pool trajectory data. During image acquisition and processing, parallel computing technology can be used to improve data processing speed, ensuring timely data support for subsequent steps.

[0144] S2: During robot operation, the first sensing unit, the second sensing unit, and the arc sensing unit collect data in real time. Each sensing unit transmits the collected data to the second acquisition module, which preprocesses the data, including data filtering and normalization. Then, based on this data, the robot's current spatial posture data is calculated, and the welding response data when the welding torch forms the current weld point is determined, including arc intensity, welding torch surface pressure contact parameters, and transmission error between the robot and welding torch motion parameters. During the calculation process, a Kalman filter algorithm can be used to fuse the sensor data, improving the accuracy and stability of the data.

[0145] S3: The data processing module receives the molten pool trajectory data and molten state data from the first acquisition module. For each weld point, following predetermined calculation steps, it acquires multiple consecutive frames of visible light images and thermal imaging data, calculates the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error, thereby generating welding quality data. During the calculation process, multi-threading technology can be used to calculate the welding quality data of different weld points in parallel, improving computational efficiency.

[0146] S4: The data processing module analyzes and calculates the shaft acceleration fluctuation, welding torch speed change rate, and arc current deviation based on the spatial attitude data and welding response data transmitted from the second acquisition module, generating process parameter data. To improve the accuracy of the calculation, the calculation results can be verified and corrected multiple times, for example, by comparing and analyzing with historical data to determine the rationality of the calculation results.

[0147] S5: The data correction module performs weighted normalization on the calculated thermal deformation error to generate the first deviation value; it also performs weighted normalization on the process parameter data to generate the second deviation value. During the weighting process, the weights can be determined using the Analytic Hierarchy Process (AHP). Through expert scoring or actual data analysis, the relative importance of each indicator is determined, thus obtaining reasonable weight values.

[0148] S6: The data correction module constructs an association mapping according to a specific method. It transforms the first thermal deformation error into a position weighting factor, generates a thermal influence correction matrix based on the heat conduction rate and the second thermal deformation error, outputs the speed compensation value and voltage adjustment amount, and fuses the position weighting factor and correction matrix to generate the overall influence rate. The third and fourth thermal deformation errors are transformed into boundary solidification factors and interference compensation coefficients to generate a single influence rate. During the construction process, machine learning algorithms can be used to learn and train on a large amount of welding data to optimize the model parameters of the association mapping, improving its accuracy and reliability.

[0149] S7: When the overall impact rate exceeds the first preset value, the data correction module proportionally increases the arc voltage gradient adjustment based on the overall impact rate value, and expands the compensation range of the welding torch movement speed to ensure that each heat deformation error is within the preset range, using the third heat deformation error value as a benchmark. When a single impact rate exceeds the second preset value, compensation is performed for the target heat deformation error. During the compensation process, changes in welding quality data and process parameter data are monitored in real time, and the compensation strategy is adjusted according to the actual situation to ensure that the welding quality meets the requirements. Simultaneously, the corrected process parameters are fed back to the robot control system, enabling the robot to perform welding operations according to the new process parameters.

[0150] The foregoing description and accompanying drawings fully illustrate embodiments of this disclosure to enable those skilled in the art to practice them. Other embodiments may include structural, logical, electrical, procedural, and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operations may vary. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. Moreover, the terminology used in this application is for describing embodiments only and is not intended to limit the claims. As used in the description of embodiments and claims, the singular forms “a,” “an,” and “the” are intended to equally include the plural forms unless the context clearly indicates otherwise. Similarly, the term “and / or” as used in this application means including one or more of the associated listed items and all possible combinations thereof. Additionally, when used in this application, the term "comprise" and its variations "comprises" and / or "comprising" refer to the presence of stated features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof. Without further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, or apparatus that includes said element. In this document, each embodiment may focus on the differences from other embodiments, and similar or identical parts between embodiments can be referred to mutually. For methods, products, etc., disclosed in the embodiments, if they correspond to the method section disclosed in the embodiments, the relevant parts can be referred to the description of the method section.

[0151] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented using electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods for each specific application to achieve the described functions, but such implementation should not be considered beyond the scope of the embodiments of this disclosure. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the described devices, apparatuses, and units can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0152] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, function, and operation of possible implementations of apparatus, methods, and computer program products according to embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than those disclosed in the description; sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based device that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

Claims

1. A robot vision-coordinated soldering system, characterized in that, include: The first acquisition module is configured to acquire the molten pool trajectory data when the welding trajectory is formed and the melting state data at each weld point in the molten pool trajectory. The second acquisition module is configured to acquire the robot’s spatial posture data via a multi-axis sensor and determine the welding response data when the welding torch forms a weld point under the current spatial posture data. The data processing module is configured to acquire the molten pool trajectory data and the molten state data to determine the welding quality data at each weld point, and is configured to determine the process parameter data based on the spatial attitude data and the welding response data. The data correction module is configured to determine a first deviation value between the welding quality data and the preset welding quality, and at the same time determine a second deviation value between the process parameter data and the preset process parameters, establish a data mapping between the first deviation value and the second deviation value, and feed back the data mapping to the robot to generate correction parameters. The first acquisition module includes: The first imaging unit is configured to acquire visible light images of each solder joint in the molten pool trajectory; The second imaging unit is configured to acquire thermal imaging images of each solder joint in the molten pool trajectory via infrared. The first acquisition module acquires multiple melting morphology boundaries at different temperatures in the thermal imaging image, maps the melting morphology boundaries to the corresponding visible light image, and generates melting state data including melting morphology boundaries and physical morphology boundaries; The first acquisition module acquires molten state data from multiple consecutive weld points in the thermal imaging image to generate molten pool trajectory data; The second acquisition module includes: The first sensing unit is configured to acquire the acceleration and velocity parameters of each axis and each arm of the robot; The second sensing unit is configured to acquire the acceleration, velocity parameters, and surface pressure parameters of the welding torch; The arc sensing unit is configured to determine the current and voltage parameters of the surface-induced arc when the welding torch performs the operation; The second acquisition module obtains the spatial attitude data by determining the relative position and attitude between each arm and each axis of the robot at its current position based on the first sensing unit. The second acquisition module determines the arc intensity and pressure contact parameters on the surface of the welding torch when the welding torch forms the current weld point, as well as the first motion parameters of the robot and the second motion parameters of the welding torch, based on the second sensing unit and the arc sensing unit. The first motion parameters include the acceleration and velocity parameters of each axis and each arm of the robot, and the second motion parameters are the acceleration and velocity parameters of the welding torch. The module determines the transmission error between the first motion parameters and the second motion parameters under the current spatial posture data, and obtains the welding response data. When the data processing module determines the welding quality data, it is configured to perform the following steps: Acquire multiple consecutive frames of visible light images and corresponding thermal imaging data at the first solder joint; Acquire the first frame that forms the first physical shape boundary from multiple consecutive frames, and determine the first melt shape boundary corresponding to the first physical shape boundary; The first thermal deformation error is determined by comparing the distance between the boundary of the first physical state and the boundary of the first molten state. Acquire the second frame that forms the second physical form boundary from multiple consecutive frames, determine the second molten form boundary corresponding to the second physical form boundary, determine the distance between the first molten form boundary and the second molten form boundary, determine the first heat conduction rate, determine the distance between the second physical form boundary and the second molten form boundary, determine the second thermal deformation error, and establish a mapping relationship between the first heat conduction rate and the second thermal deformation error. The third frame that forms the third physical form boundary is obtained from multiple consecutive frames. The third molten form boundary at this time is determined, the distance from the third molten form boundary to the third physical form boundary is determined, and the third thermal deformation error is determined. The third physical form boundary is the limit boundary when no new physical form boundary is formed. Acquire visible light images and corresponding thermal imaging data of multiple consecutive frames of adjacent solder joints of the first solder joint; When obtaining the first physical boundary formed by the adjacent weld points and calculating the first thermal deformation error; The molten state boundary of the first weld point is determined, and the molten state boundary is used as the interference interval of the first thermal deformation error of the adjacent weld points to obtain the fourth thermal deformation error. The welding quality data is obtained based on the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error. The data correction module is configured to perform the following steps: Extract the first thermal deformation error, the second thermal deformation error, the third thermal deformation error, and the fourth thermal deformation error to construct a welding quality feature group; After assigning weights to each item in the welding quality feature group and the preset welding quality benchmark group, normalization is performed, and the difference between the two is calculated to obtain the offset of each item, generating the first deviation value. After assigning weights to the shaft acceleration, welding torch speed and arc current of the process parameter data and the preset process parameters, the data is normalized, and the difference between the two is calculated to generate a second deviation value. Establish a correlation mapping between the first deviation value and the second deviation value. The correlation mapping includes a first component and a second component. The first component is the overall influence rate of the process parameters on the welding quality, and the second component is the individual influence rate of each process parameter on each parameter in the welding quality feature group.

2. The robot vision collaborative control soldering system according to claim 1, characterized in that, The construction of the association mapping includes the following steps: The first thermal deformation error is mapped to a spatial coordinate system to generate a first position weighting factor characterizing the stability of the weld point position. Based on the correlation between the first heat conduction rate and the second heat deformation error, a dynamic correction matrix for heat influence with two control dimensions is constructed. The second heat deformation error is linearly converted into a welding torch moving speed compensation value, and the first heat conduction rate is mapped into an arc voltage gradient adjustment amount. By integrating the first position weighting factor and the dynamic correction matrix of thermal influence, the overall influence rate of process parameters on welding quality is generated. The third thermal deformation error is normalized to generate a boundary solidification factor that characterizes the degree of solidification at the weld pool boundary. The fourth thermal deformation error is converted into an interference compensation coefficient between adjacent weld points. The interference compensation coefficient is correlated with the boundary solidification factor to quantify the single influence rate of process parameters on individual welding defects.

3. The robot vision collaborative control soldering system according to claim 2, characterized in that, The data correction module is also configured to perform the following steps: The first positional weighting factor is spatially coupled with the output of the thermal effect dynamic correction matrix to generate the global influence intensity of process parameters on welding quality. The global influence intensity is calibrated by combining the degree of solidification at the molten pool boundary to obtain the overall influence rate value within a preset range; Based on the influence of the real-time adjustment of the welding torch speed compensation value and the arc voltage adjustment value on each thermal deformation error, dynamic influence weights are assigned to them respectively. Each thermal deformation error is weighted and superimposed with the interference compensation coefficient to quantify the targeted influence intensity of each process parameter on characteristic welding defects. When the overall impact rate exceeds the first preset value, the arc voltage gradient adjustment amount is increased proportionally to the overall impact rate value, and the compensation range of the welding torch moving speed is expanded to the preset range of each thermal deformation error based on the third thermal deformation error value. When any single influence rate exceeds the second preset value, compensation of the corresponding process parameters is performed until the corresponding thermal deformation error is within the preset range.

4. A robot vision-coordinated soldering method, implemented based on the system described in claim 3, characterized in that, Includes the following steps: S1. Obtain molten pool trajectory data and solder joint melting state data, including: Melt morphology boundaries are acquired simultaneously using visible light and infrared imaging to generate melt state data; Merging molten state data from continuous weld points generates molten pool trajectory data; S2. The robot's spatial posture data is acquired through multi-axis sensors, and the transmission error is calculated by combining the welding torch surface pressure, arc parameters and motion parameters to obtain welding response data. S3. Calculate four thermal deformation errors based on the molten pool trajectory data and molten state data to generate welding quality data; S4. Based on spatial attitude data and welding response data, analyze the axial acceleration fluctuation, welding torch speed change rate and arc current deviation to generate process parameter data; S5. Perform double deviation value calculation: The first deviation value is generated by weighted normalization of the thermal deformation error; The process parameters are weighted and normalized to generate a second deviation value; S6. Construct association mapping: The first thermal deformation error is converted into a position weighting factor; A thermal effect correction matrix is ​​generated based on the thermal conduction rate and the second thermal deformation error, and the output speed compensation value and voltage regulation amount are then used. The overall influence rate is generated by fusing the location weighting factor and the correction matrix; The third and fourth thermal deformation errors are converted into boundary solidification factors and interference compensation coefficients to generate a single influence rate. S7, Response Correction Control: When the overall impact rate is greater than the first preset value, the voltage regulation intensity is increased and the speed compensation range is expanded. When the single influence rate is greater than the second preset value, compensation is performed for the target thermal deformation error.

Images