Dynamic calibration method for cosmetic intelligent manufacturing robot

Abstract

This invention relates to the field of dynamic calibration of incoming materials, and more particularly to a dynamic calibration method for a cosmetic intelligent manufacturing robot. This invention acquires real-time image data of incoming materials for corresponding batches on the production line to extract the contact state characteristics of the incoming materials; combines the contact state characteristics with the material spacing deviation value to calculate the placement deviation characterization value of the incoming materials, thus marking the incoming materials; in response to the presence of materials marked as deviated, the operating conditions of the production line are adaptively evaluated, analyzed, and adjusted, thereby controlling the robot to perform position calibration and adjustment operations on the deviated materials. This invention can achieve automation, precision, and efficiency in incoming material deviation calibration of cosmetic production lines, balancing calibration accuracy and production efficiency, and improving the stability of production line operating conditions and robot adaptability.

Description

Technical Field

[0001] This invention relates to the field of dynamic calibration of incoming materials, and more particularly to a dynamic calibration method for a cosmetic intelligent manufacturing robot. Background Technology

[0002] Against the backdrop of the rapid development of the intelligent manufacturing industry in the cosmetics sector, automated production lines have become the core carrier for improving quality and efficiency and achieving standardized production in the beauty industry. Embossed robots, due to their precise operational capabilities and adaptability to multiple scenarios, are widely used in core processes such as material handling and gripping in cosmetic production lines, becoming key execution units in the intelligent manufacturing system of cosmetics. Cosmetic production materials are mostly irregularly shaped lightweight parts such as bottles, cans, and tubes. To improve production efficiency, production lines generally adopt continuous and dense material handling modes. The conveying process of materials on conveyor belts and other transmission devices is easily affected by various factors such as equipment vibration and transmission speed fluctuations, resulting in problems such as placement posture deviation, stacking tilt, and uneven spacing.

[0003] Due to the existence of incoming material deviations, if robots directly grasp materials with deviations, it is highly likely to cause problems such as grasping errors, material falling and damage, and equipment jamming. This not only leads to material waste and production interruptions, but may also cause wear and tear on the robot's end effector and transmission device failure due to equipment collisions, increasing the maintenance costs of the production line and affecting overall production efficiency and operational stability. Against this backdrop, incoming material position calibration has become a necessary process in the automated production line of cosmetic intelligent manufacturing to ensure the normal operation of robots and maintain production continuity.

[0004] Meanwhile, with the integration of industrial automation and machine vision technology, machine vision-based incoming material inspection and calibration technology is gradually being applied to various manufacturing fields, providing technical support for incoming material deviation calibration in cosmetic production lines. The combination of embodied robots and machine vision systems enables the visual acquisition of incoming material status and precise control of robot operations, making automated incoming material calibration possible. Machine vision technology can acquire image data, extract features, and identify deviations from incoming materials on the production line, providing data for robot calibration operations; while the multi-degree-of-freedom motion and precise pose adjustment capabilities of embodied robots provide the execution basis for position calibration of deviation-prone incoming materials. The synergistic application of these two technologies has become the mainstream technical direction for dynamic incoming material calibration in the cosmetic intelligent manufacturing field. Against this industrial technology background, machine vision-based embodied robot dynamic calibration technology has become a key research direction for solving the problem of incoming material deviation in cosmetic production lines and improving the level of intelligent manufacturing. Its core lies in achieving accurate identification of incoming material deviations, real-time evaluation of production line conditions, and dynamic adaptation of robot calibration operations to meet the actual production needs of cosmetic intelligent manufacturing.

[0005] Chinese Patent Application Publication No. CN118875824A discloses a method and system for correcting board alignment based on robot vision. The method includes: a vision correction system taking a picture of the target board to obtain a label image; a robot system sending theoretical information about the board to the vision correction system; the vision correction system calculating the target barcode coordinates based on the target feature information of the barcode on the label image; calculating the actual coordinates of the label in the robot coordinate system based on the transformation relationship matrix, the reference point coordinates of the calibration image in the robot coordinate system, and the target barcode coordinates; calculating the positioning offset based on the reference point coordinates of the calibration image in the robot coordinate system, the actual coordinates of the label in the robot coordinate system, the theoretical coordinates of the board, and the theoretical coordinates of the label; calculating the center coordinates of the target board based on the positioning offset; and sending the center coordinates of the target board to the robot system so that the robot system can control the robot to adsorb the target board. This method can improve the accuracy and production efficiency of automated operations and can be widely applied in the field of automation control technology.

[0006] However, the following problems still exist in the existing technology. Most automated calibration methods focus on fixed calibrations for single types of incoming material deviations, lacking adaptability to the diverse nature of cosmetic incoming materials and the complexity of deviation types. Furthermore, the lack of dynamic adjustments to calibration decisions based on the actual operating conditions of the production line easily leads to problems such as insufficient calibration accuracy and mismatch between calibration operations and the production line's transmission rhythm, making it difficult to balance the dual requirements of calibration accuracy and production efficiency. Summary of the Invention

[0007] To address this, the present invention provides a dynamic calibration method for a cosmetic intelligent manufacturing robot. This method overcomes the limitations of existing automated calibration methods, which mostly perform fixed calibrations for single types of incoming material deviations, lacking adaptability to the diversity of cosmetic incoming materials and the complexity of deviation types. Furthermore, the lack of dynamic adjustment of calibration decisions based on the actual operating conditions of the production line easily leads to problems such as insufficient calibration accuracy and mismatch between calibration operations and the production line's transmission rhythm, making it difficult to balance the dual requirements of calibration accuracy and production efficiency.

[0008] To achieve the above objectives, the present invention provides a dynamic calibration method for a cosmetic intelligent manufacturing robot, comprising: Real-time acquisition of incoming material image data for corresponding batches on the production line to extract the contact state features of the incoming materials, including the horizontal deviation angle of the contact material and the contact area of ​​the contact material. Based on the aforementioned contact state characteristics and material spacing deviation values, the placement deviation characterization value of the incoming material is calculated to mark the incoming material. In response to the presence of incoming materials marked as deviating from the standard, the operating conditions of the production line are evaluated, analyzed, and adjusted, including: Determine the nearest distance between the deviated incoming material and the gripping position, as well as the material distribution characteristics within the predetermined transmission range, analyze the overall deviation characterization coefficient to determine whether the placement deviation threshold is met, and control the robot to perform position calibration and adjustment operations on the deviated incoming material. If the placement deviation threshold is not met, the current posture image of the robot is obtained, the deviation concentration area is locked, and based on the space margin for the end effector to operate on the deviation concentration area, it is determined whether to pause the feeding transmission device. Based on the judgment result, the sequence of position calibration and adjustment operations performed by the end effector for the deviated incoming material is determined; If the placement deviation threshold is met, the materials with deviation are sorted in descending order based on the positional distance between the materials with deviation and the corresponding end effector of the robot, and the end effector is controlled to perform position calibration and adjustment operations in sequence according to the sorting. The incoming material distribution characteristics include the quantity of materials with deviations and the percentage of the stacked area of ​​incoming materials.

[0009] Further, the process of calculating the placement deviation characterization value of the incoming material includes: The sum of the ratio of the horizontal deviation angle of the incoming material to the horizontal deviation angle threshold and the ratio of the bonding area of ​​the incoming material to the bonding area threshold is used as the first placement deviation characteristic value. The ratio of the material spacing deviation value to the material spacing deviation threshold is used as the second placement deviation characteristic value; The first placement deviation feature value and the second placement deviation feature value are weighted and summed to determine the placement deviation characterization value.

[0010] Furthermore, the incoming materials are labeled, including: If the placement deviation characterization value of the incoming material is greater than or equal to the placement deviation characterization threshold, the incoming material is marked as a deviation incoming material.

[0011] Furthermore, the process of analyzing the overall deviation characterization coefficient includes: The ratio of the nearest distance threshold to the nearest distance from the material to the gripping position is used as the first overall deviation feature value. The sum of the ratio of the deviation quantity of incoming materials to the threshold of the deviation quantity of incoming materials and the ratio of the proportion of incoming material stacking area to the threshold of the proportion of incoming material stacking area is used as the second overall deviation characteristic value. The weighted sum of the first overall deviation characteristic value and the second overall deviation characteristic value is determined as the overall deviation characterization coefficient.

[0012] Further, determining whether the placement deviation threshold is met includes: If the overall deviation characterization coefficient is less than the overall deviation characterization coefficient threshold, then the placement deviation threshold is satisfied.

[0013] Furthermore, the process of identifying the concentrated area of ​​deviation includes: The predetermined transmission range is divided into several material distribution areas; Determine the positional dispersion of the incoming materials within each of the aforementioned incoming material distribution areas; The material distribution area corresponding to the minimum positional dispersion is locked as the deviation concentration area; Specifically, the standard deviation of the gripping distance of the end effector for the biased incoming material in each of the material distribution areas is calculated, and the standard deviation of the gripping distance is used as the positional dispersion.

[0014] Further, determining whether to pause the feeding transmission device includes: Determine the area of ​​empty material location in the deviation concentration region and the gap width between the end effector and the deviation material; The sum of the ratio of the area of ​​the empty material location to the threshold value of the empty material location and the ratio of the gap width to the threshold value of the gap width is used as the space margin. If the space margin is greater than or equal to the space margin threshold, it is determined that there is no need to pause the feeding transmission device.

[0015] Further, determining the sequence of position calibration and adjustment operations performed by the end effector for the deviated incoming material includes: If the determination result is to pause the feeding transmission device, then based on the positional distance between the deviation material and the corresponding end effector of the robot, the deviation material is sorted in descending order, and the end effector is controlled to perform position calibration and adjustment operations in sequence according to the sorting. If the determination result is that there is no need to pause the feeding transmission device, then based on the distribution deviation characteristics and the same degree of gripping direction of the gripping position corresponding to the deviation material, the operation limitation characterization parameters are evaluated to adjust the pose correction range of the end effector. The end effector is controlled to preferentially perform position calibration and adjustment operations on the incoming material with deviation within the deviation concentration area; The distribution deviation characteristics include the tilt angle of the incoming material and the position deviation angle.

[0016] Furthermore, the process of evaluating operationally constrained characterization parameters includes: The sum of the ratio of the placement tilt angle of the biased incoming material to the placement tilt angle threshold and the ratio of the position deviation angle to the position deviation angle threshold is used as the first operation-restricted feature value. The ratio of the grip direction consistency threshold to the grip direction consistency of the corresponding grip position of the deviation material is used as the second operation-restricted feature value. The weighted sum of the first operation-restricted feature value and the second operation-restricted feature value is used to determine the operation-restricted characterization parameter.

[0017] Further, adjusting the pose correction range of the end effector includes: Increase the pose correction range, and the increase in the pose correction range is positively correlated with the operation-limited characterization parameter; The pose correction range includes a translation compensation range and a rotation compensation range.

[0018] Compared with existing technologies, this invention extracts the contact state characteristics of incoming materials by acquiring real-time image data of corresponding batches on the production line. Combining these contact state characteristics with material spacing deviation values, it calculates the placement deviation characterization value of the incoming materials to mark them. In response to the presence of materials marked as deviated, the invention adaptively evaluates and adjusts the production line's operating conditions, thereby controlling the robot to perform position calibration and adjustment operations on the deviated materials. This invention enables automated, precise, and efficient incoming material deviation calibration in cosmetic production lines, balancing calibration accuracy and production efficiency, and improving production line stability and robot adaptability.

[0019] In particular, in the actual operation of cosmetic production lines, incoming materials are mostly bottled and jarred beauty products. Due to factors such as product shape and production line transmission, problems such as stacking tilt and uneven spacing are prone to occur. Based on this, this invention quantifies the distribution morphology and spatial arrangement characteristics of incoming materials. For the entire stack of incoming materials, the horizontal tilt deviation of the entire stack of incoming materials is quantified by the horizontal deviation angle of the materials, thereby reflecting the regularity of the geometric shape of the material being stacked, and corely characterizing the horizontal posture regularity of the material as a whole. This feature is related to the accuracy of the robot's visual positioning of the material. The larger the value of this feature, the more serious the horizontal tilt of the material, the more likely the robot's visual positioning is to misjudge the grasping center, and the material is prone to slippage due to uneven force during grasping; the smaller the value of this feature, the closer the material is to a horizontal placement, the more accurately the robot can locate the grasping reference of the material, ensuring the stability of the grasping. Furthermore, regarding the actual contact state between incoming materials, the impact of the contact area on the accuracy of the robot's grasping calibration operation and the risk of collision between materials are quantified by the contact area. This feature is the core reference for the physical contact and positioning when the robot grasps and adjusts materials that are close together. It corresponds to the size of the gap and the degree of misalignment between the materials, reflecting the actual operational impact of the contact state between the materials on the robot's grasping and adjustment operation. The larger the contact area, the tighter the contact between the materials. When the robot grasps a single material, the boundary between the operating space of the end effector and the corresponding material is blurred, making it prone to scratches and a high risk of collision. The smaller the contact area, the lower the probability of collision or scratches between the end effector and the corresponding material, and the lower the risk of collision. In addition, from the perspective of production line spatial layout, the uniformity of the spatial distribution of materials on the production line is quantified by the material spacing deviation value, characterizing the uniformity of the material distribution on the transmission path, which is related to the robot's grasping path planning efficiency and the continuous transmission efficiency of the production line. The larger the value of this feature, the more uneven the actual arrangement of incoming materials on the production line, indicating a scattered distribution of materials and problems with positional offsets. Furthermore, the placement deviation characterization value is calculated to form a comprehensive characterization of the placement distribution of cosmetic incoming materials in terms of their contact state and spatial arrangement. This invention is adapted to the high-speed transmission characteristics of intelligent cosmetic manufacturing production lines, enabling rapid screening of deviation-prone incoming materials without manual intervention, thus improving the comprehensiveness and accuracy of deviation judgment.

[0020] In particular, since materials in cosmetic production lines are continuously conveyed in batches on the transmission device, deviations in a single material may be localized or represent a batch deviation problem across the entire transmission system. Therefore, this invention extracts the spatial location characteristics of the deviated material and the batch distribution characteristics within a predetermined transmission range, analyzes the overall deviation characterization coefficient, and achieves a quantitative assessment of the overall deviation status of all materials within a specific transmission interval. This upgrades deviation identification from a "point" to a "surface" level of operational control, avoiding the neglect of the overall deviation trend of the production line due to focusing only on a single deviated material, and ensuring that subsequent calibration decisions are more aligned with the actual operating state of the production line. This invention comprehensively reflects the overall deviation characteristics within a predetermined transmission range from three dimensions: time urgency, deviation scale, and spatial impact range. In the spatial location dimension, the time urgency and operational window for deviation calibration are quantified based on the closest distance between the deviated material and the gripping position. The smaller this characteristic data, the closer the deviated material is to the gripping position, indicating a shorter time window for the robot to perform calibration operations and a higher calibration urgency. If not handled promptly, the deviated material will quickly enter the gripping stage, easily causing robot gripping errors, material damage, and equipment jamming. The larger the value of this feature, the farther the deviated material is from the gripping position, indicating more time for calibration operations. The robot can plan a more reasonable calibration path without emergency intervention. Simultaneously, this feature quantifies the immediate threat posed by the deviated material to the gripping station, serving as a core spatial indicator for determining whether the production line needs to initiate an emergency calibration operation. In terms of quantity, the number of deviated materials quantifies the scale and frequency of deviations within the predetermined transmission range. A larger value indicates more deviated materials within the predetermined transmission range, suggesting the deviation is not a localized, sporadic issue but rather a malfunction in the production line's transmission device or feeding equipment, such as conveyor belt misalignment or inaccurate positioning of the feeding robot. The overall deviation is large, with a wide impact on subsequent continuous gripping operations. A smaller value indicates that the deviated material is only an isolated, sporadic occurrence, and the overall production line is operating well. Furthermore, in terms of spatial proportion, the ratio of the total stacked area of ​​all deviated materials within the predetermined transmission range to the total effective placement area within that range quantifies the spatial proportion and distribution density of deviated materials within the predetermined transmission range, indirectly reflecting the actual impact range of the deviation. The larger this characteristic value, the higher the proportion and density of the deviated material within the predetermined transmission range. This not only compresses the calibration operation space of the robot's end effector but may also cause the deviated material to squeeze against the normal material, leading to secondary deviations, such as the normal material being squeezed out of place, misaligned, or even clogging the conveyor belt. Conversely, the smaller the value, the lower the proportion and dispersion of the deviated material within the space, resulting in a higher margin of space for robot calibration operations, less impact on the arrangement and transmission of normal material, and fewer interfering factors during calibration.This feature compensates for the deficiency of "abnormal incoming material quantity" which only reflects the absolute number, further quantifying the actual spatial impact of abnormal incoming materials, and is a key indicator for determining whether there are spatial constraints in robot calibration operations. Therefore, this invention combines the aforementioned three features to analyze the overall deviation characterization coefficient, achieving a three-dimensional and comprehensive quantification of the overall deviation conditions of the production line, characterizing the actual distribution of batches of abnormal incoming materials and their impact on production operations, providing data support for subsequent determination of whether placement deviation thresholds are met. This invention closely matches the actual operating conditions of high-speed transmission and batch distribution of incoming materials in cosmetic production lines, achieving both precise control over the overall deviation state of the production line and making robot calibration operation decisions more adaptable.

[0021] In particular, this invention constructs a logical chain of "quantification of location dispersion - locking of deviation concentration area - assessment of space redundancy". First, the calibration operation is concentrated in the core area with the highest deviation density, avoiding path redundancy caused by dispersed operations, significantly improving the calibration efficiency of batch deviation materials, and realizing "focusing on key points and intensive" calibration operations, which is suitable for the calibration efficiency requirements of high-speed transmission in cosmetic production lines. In reality, cosmetic materials are mostly irregularly shaped parts such as bottles and cans, and production lines usually adopt a dense arrangement to improve efficiency, which easily leads to the mutual compression of deviation materials and narrow gaps, placing higher demands on the robot's operating space. Based on this, before controlling the robot to perform calibration operations, it is necessary to consider whether to pause the transmission device to ensure that the robot's end effector has sufficient physical operating space in the deviation concentration area. If the space is insufficient and transmission continues, it is very easy for the end effector to collide with the material or conveyor belt, or for the calibration operation to be incomplete. This invention selects two core indicators: the area of ​​the empty material position and the width of the gap between the end effector and the material, respectively, to determine the space redundancy from the dimensions of overall space and local operating space, comprehensively quantifying the actual usable operating space of the robot in the deviation concentration area. The aforementioned two indicators directly correspond to the physical operational limits of the end effector, such as the size of the gripper and the minimum space required for rotation / translation, ensuring a high degree of match between the space margin assessment results and the robot's actual operational capabilities. This allows for continuous production line operation without interruption when there is sufficient operating space, while also enabling timely pauses when operating space is insufficient, providing a stable and safe calibration environment for the robot and mitigating production risks such as equipment collisions and material damage at the source. This invention solves the problem of difficulty in identifying concentrated deviation areas under dense material arrangement by locking the deviation concentration area; and it accurately adapts to the operating space requirements under dense arrangement of irregularly shaped materials through a two-dimensional spatial assessment of the empty material location area and the gap width between incoming materials. This effectively determines the operable space of the end effector in narrow gaps, avoiding calibration operation failures caused by dense material arrangement. Furthermore, it adapts to the production needs of different categories of cosmetic materials, possessing good scenario adaptability and scalability.

[0022] In particular, this invention designs a calibration operation planning stage for non-pause operation conditions. The aim is to adapt the robot end effector calibration operation to the production line's transmission rhythm while maintaining continuous production without interruption. This ensures the calibration operation accurately adapts to the actual constraints of the deviation conditions, is efficiently completed in continuous production, and maximizes the stability and accuracy of the calibration operation. In practice, for non-pause operation conditions, the core challenge of robot calibration operation is matching the pose correction range with the degree of deviation constraint. Therefore, this invention characterizes the actual constraints of robot calibration operation from two dimensions: the inherent deviation characteristics of the incoming material and the characteristics of the batch grasping direction. In terms of elevation geometry, the tilt deviation of a single biased incoming material is quantified by its placement tilt angle, reflecting the basic requirements for robot rotation correction. A larger value indicates a more severe material tilt, requiring a larger rotation compensation range from the robot end effector to calibrate it to a standard posture, resulting in stronger rotational constraints during operation. A smaller value indicates that the incoming material is placed close to a vertical standard posture, requiring only minor rotational correction or even no rotation at all, indicating lower constraints on rotational operation. In terms of planar orientation, the degree of orientational deviation of the incoming material's spatial position is quantified by the deviation angle of the incoming material's position, reflecting the robot's orientational adaptation requirements for translation / rotation. A larger value indicates a more significant orientational offset of the incoming material within the transmission plane, requiring the robot to adjust the planar operating orientation of the end effector, and even combining translational and rotational corrections for precise calibration, resulting in stronger orientational constraints during operation. A smaller value indicates that the spatial orientation of the incoming material is close to the baseline arrangement, requiring less adjustment of the robot's operating orientation, and exhibiting lower limitations in planar operation. These two features form a two-dimensional geometric deviation quantification of the planar and elevation dimensions, fully characterizing the overall posture deviation of a single defective incoming material, which is the core basis for robot pose correction. Simultaneously, by combining the overall quantification index for all defective incoming materials within the deviation concentration area—that is, the uniformity of the gripping direction at the corresponding gripping position of the defective incoming material—the consistency of the gripping direction of a batch of incoming materials within the deviation concentration area is quantified, reflecting the directional coordination constraints of the robot's batch operations. The lower the value of this feature, the more chaotic the gripping directions of the incoming materials within the area. The robot needs to expand the range of pose correction directions and frequently adjust the gripping direction to adapt to all incoming materials, resulting in low efficiency and strong directional constraints in continuous operation. Conversely, the higher the value of this feature, the more consistent the gripping directions of the incoming materials within the deviation concentration area. The robot does not need to frequently adjust the gripping direction; it only needs to perform small pose corrections within a fixed directional range, resulting in high efficiency and weak directional constraints in continuous operation. This feature compensates for the limitations of the distributed deviation feature, reflecting the robot's operational constraints from a batch operation perspective. It is a key indicator for adapting to continuous transmission and batch calibration conditions on production lines. Furthermore, evaluating the operational constraint characterization parameters helps to characterize the overall degree of constraint on the robot's calibration operations within the deviation concentration area under non-pause conditions.This invention achieves a perfect match between robot calibration operations and the continuous transmission rhythm of the production line through dynamic pose correction and efficient operation sequence planning. This ensures both the accuracy and efficiency of calibration operations while maintaining production continuity. Furthermore, adapting to the actual working conditions of the cosmetic production line and the technical characteristics of the robot is crucial for ensuring the efficient operation of the dynamic calibration system under non-stop conditions. Attached Figure Description

[0023] Figure 1 A schematic diagram illustrating the steps of a dynamic calibration method for a cosmetic intelligent manufacturing robot according to an embodiment of the invention; Figure 2 This is a logic diagram for labeling incoming materials according to an embodiment of the invention. Figure 3 This is a logic diagram for determining whether a placement deviation threshold is met in an embodiment of the invention. Figure 4 This is a logic diagram for determining whether to pause the feeding transmission device according to an embodiment of the invention. Detailed Implementation

[0024] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0025] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0026] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0027] Please see Figure 1 The diagram illustrates the steps of a dynamic calibration method for a cosmetic intelligent manufacturing robot according to an embodiment of the present invention. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to an embodiment of the present invention includes: Step S1: Real-time acquisition of incoming material image data for the corresponding batch on the production line to extract the contact state features of the incoming material. The contact state features include the horizontal deviation angle of the contact material and the contact area of ​​the contact material. Step S2: Combine the contact state characteristics and the material spacing deviation value to calculate the placement deviation characterization value of the incoming material, so as to mark the incoming material. Step S3, in response to the presence of incoming materials marked as deviating materials, evaluate, analyze, and adjust the operating conditions of the production line, including, Determine the nearest distance between the deviated incoming material and the gripping position, as well as the material distribution characteristics within the predetermined transmission range, analyze the overall deviation characterization coefficient to determine whether the placement deviation threshold is met, and control the robot to perform position calibration and adjustment operations on the deviated incoming material. If the placement deviation threshold is not met, the current posture image of the robot is obtained, the deviation concentration area is locked, and based on the space margin for the end effector to operate on the deviation concentration area, it is determined whether to pause the feeding transmission device. Based on the judgment result, the sequence of position calibration and adjustment operations performed by the end effector for the deviated incoming material is determined; If the placement deviation threshold is met, the materials with deviation are sorted in descending order based on the positional distance between the materials with deviation and the corresponding end effector of the robot, and the end effector is controlled to perform position calibration and adjustment operations in sequence according to the sorting. The incoming material distribution characteristics include the quantity of materials with deviations and the percentage of the stacked area of ​​incoming materials.

[0028] Specifically, the incoming material image data includes features such as contact state characteristics, material spacing deviation value, the closest distance between the deviated incoming material and the gripping position, material distribution characteristics, position dispersion, spatial margin, and the position distance between the deviated incoming material and the robot's corresponding end effector.

[0029] The acquisition of incoming material image data can be achieved by relying on mature machine vision technology combined with image recognition algorithms. The relevant technologies are existing conventional technologies in the field of industrial automation. The specific implementation methods will not be elaborated here.

[0030] Specifically, the process of calculating the placement deviation characterization value of the incoming material includes: The sum of the ratio of the horizontal deviation angle of the incoming material to the horizontal deviation angle threshold and the ratio of the bonding area of ​​the incoming material to the bonding area threshold is used as the first placement deviation characteristic value. The ratio of the material spacing deviation value to the material spacing deviation threshold is used as the second placement deviation characteristic value; The first placement deviation feature value and the second placement deviation feature value are weighted and summed to determine the placement deviation characterization value.

[0031] Specifically, the two features involved in the contact state characteristics directly affect the safety and feasibility of robot grasping. In practice, material tilting can lead to robot visual positioning errors, uneven force during grasping, and easy slippage, dropping, or even damage to the material. Furthermore, tight contact between materials can compress the operating space of the end effector, increasing the risk of scratches and collisions. These two features together determine whether the robot can safely and stably complete the calibration operation, belonging to the highest priority "safety-type" indicators. Uneven material spacing mainly affects the robot's grasping path planning and cycle time, and generally does not directly lead to grasping failure or equipment damage. It can be adapted by adjusting the grasping order or trajectory, belonging to the "efficiency-type" indicator. Under the premise of ensuring safety, uneven arrangement can be tolerated to a certain extent, but severe stacking must be dealt with first. Therefore, in the weighted summation, the first placement deviation feature value is given a higher weight coefficient, which can be selected in the range of 0.6 to 0.8, and is set to 0.7 in this embodiment. Correspondingly, the weight coefficient of the second placement deviation feature value can be selected in the range of 0.2 to 0.4, and is set to 0.3 in this embodiment. Those skilled in the art can adaptively fine-tune the balance between safety and efficiency based on the actual production line's emphasis on these two aspects.

[0032] In this embodiment, the purpose of setting the horizontal deviation angle threshold, the bonding area threshold, and the material spacing deviation threshold is to characterize situations where the incoming material stacking is severe and the overall spatial arrangement is highly discrete. This is achieved by acquiring historical data of incoming material images corresponding to several batches of material conveying processes on the production line, and by calling historical data on the horizontal deviation angle of the incoming material, the bonding area of ​​the incoming material, the material spacing deviation value, and the percentage of the incoming material stacking area. The average horizontal deviation angle, the average bonding area, and the average material spacing deviation are calculated and used as benchmark values ​​under normal conditions. Based on the purpose of setting the above three thresholds, the horizontal deviation angle threshold is determined as the product of the average horizontal deviation angle and the horizontal deviation coefficient; the bonding area threshold is determined as the product of the average bonding area and the bonding deviation coefficient; and the material spacing deviation threshold is determined as the product of the average material spacing deviation and the spacing deviation coefficient.

[0033] Specifically, considering that in actual production, due to factors such as equipment vibration and individual differences in incoming materials, there are normal random fluctuations in horizontal deviation angles, bonding areas, and material spacing deviations, if the thresholds are set too low, a large number of normal incoming materials will be misjudged as deviation materials, leading to unnecessary robot calibration operations and affecting production efficiency. Therefore, each threshold is set to an appropriate multiple of the historical average of the corresponding characteristic quantity, reserving an error tolerance for normal fluctuations. The horizontal deviation coefficient is selected within the range [1.1, 1.2], with a coefficient greater than 1, allowing for slight fluctuations in the horizontal angle. It is only considered abnormal when the incoming material is significantly tilted overall. In practice, a coefficient of 1.1 is preferred.

[0034] The bonding deviation coefficient is selected within the range [1.2, 1.3]. A larger coefficient is suitable for the normal fluctuation range of a large bonding area. Severe stacking is only considered when the incoming materials are tightly bonded together without gaps. In practice, a coefficient of 1.2 is preferred.

[0035] The arrangement deviation coefficient is selected within the interval [1.1, 1.2], with a coefficient slightly greater than 1, allowing for random and slight differences in material spacing. It is only considered discrete when the arrangement is significantly uneven, such as locally dense or locally sparse. In practice, a coefficient of 1.1 is preferred.

[0036] The selection of the above-mentioned preferred values ​​achieves a good balance between effectively identifying true deviations and maintaining production smoothness. Of course, those skilled in the art can readjust the value range according to the actual stability of the production line and the consistency of incoming materials, or they can adjust the coefficients within a given range.

[0037] Specifically, the horizontal deviation angle of the incoming material refers to the tilt angle of the horizontal plane on which the incoming material is actually placed relative to the horizontal reference plane of the production line. In this embodiment, the horizontal bearing surface on which the incoming material is normally conveyed by the production line is taken as the horizontal reference plane.

[0038] In this embodiment, the standard deviation of the spacing between the geometric centers of the same batch of incoming materials is used as the material spacing deviation value, which will not be elaborated further.

[0039] Specifically, in the actual operation of cosmetic production lines, incoming materials are mostly bottles and jars of beauty products. Due to factors such as product shape and production line transmission, problems such as stacking tilt and uneven spacing are prone to occur. Based on this, this invention quantifies and analyzes the distribution morphology and spatial arrangement characteristics of incoming materials. For incoming materials that are pressed together, the horizontal tilt deviation of the entire group of incoming materials is quantified by the horizontal deviation angle of the incoming materials, thereby reflecting the regularity of the geometric shape of the pressed materials and corely characterizing the horizontal posture regularity of the pressed materials as a whole. This feature is related to the accuracy of the robot's visual positioning of the pressed materials. The larger the value of this feature, the more serious the horizontal tilt of the pressed materials, the more likely the robot's visual positioning is to misjudge the grasping center, and the pressed materials are prone to slippage due to uneven force during grasping; the smaller the value of this feature, the closer the pressed materials are to a horizontal placement state, the more accurately the robot can locate the grasping reference of the incoming materials, and the more stable the grasping can be. Furthermore, regarding the actual contact state between incoming materials, the impact of the contact area on the accuracy of the robot's grasping calibration operation and the risk of collision between materials are quantified by the contact area. This feature is the core reference for the physical contact and positioning when the robot grasps and adjusts materials that are close together. It corresponds to the size of the gap and the degree of misalignment between the materials, reflecting the actual operational impact of the contact state between the materials on the robot's grasping and adjustment operation. The larger the contact area, the tighter the contact between the materials. When the robot grasps a single material, the boundary between the operating space of the end effector and the corresponding material is blurred, making it easy to scratch and increasing the risk of collision. The smaller the contact area, the lower the probability of collision or scratching between the end effector and the corresponding material, and the lower the risk of collision. In addition, from the perspective of production line spatial layout, the uniformity of the spatial distribution of materials on the production line is quantified by the material spacing deviation value, characterizing the uniformity of the material distribution on the transmission path, which is related to the robot's grasping path planning efficiency and the continuous transmission efficiency of the production line. The larger the value of this feature, the more uneven the actual arrangement of incoming materials on the production line, indicating a scattered distribution of materials and problems with positional offsets. Furthermore, the placement deviation characterization value is calculated to form a comprehensive characterization of the placement distribution of cosmetic incoming materials in terms of their contact state and spatial arrangement. This invention is adapted to the high-speed transmission characteristics of intelligent cosmetic manufacturing production lines, enabling rapid screening of deviation-prone incoming materials without manual intervention, thus improving the comprehensiveness and accuracy of deviation judgment.

[0040] Specifically, please refer to Figure 2 As shown, this is a logic decision diagram for marking incoming materials according to an embodiment of the present invention. Marking the incoming materials includes: If the placement deviation characterization value of the incoming material is greater than or equal to the placement deviation characterization threshold, the incoming material is marked as a deviation incoming material. If the placement deviation characterization value of the incoming material is less than the placement deviation characterization threshold, then there is no need to label the incoming material.

[0041] To determine the placement deviation characterization threshold, samples of incoming materials confirmed to be in normal placement status can be selected from historical production data of similar incoming material specifications on the same production line. The placement deviation characterization value of each sample is calculated, and the mean 'a' and standard deviation 'b' of the sequence are obtained. The sum of the product of the standard deviation 'b' and the sensitivity coefficient 'k1' and the mean 'a' is used as the placement deviation characterization threshold. Furthermore, the placement deviation characterization threshold should be dynamically updated with each production batch to adapt to changes in incoming material specifications and production line status drift. Simultaneously, the placement deviation characterization threshold should be coordinated with the horizontal deviation angle threshold, the bonding area threshold, and the material spacing deviation threshold, and can be iteratively optimized by verifying the mislabeling rate of normal incoming materials.

[0042] The sensitivity coefficient k1 can be selected within the range [1.5, 3.0] based on the tolerance for missed detections of placement deviations; for example, k1 = 2.0. If sufficient historical data is available, the 95th percentile can also be used as the threshold.

[0043] Specifically, the process of analyzing the overall deviation characterization coefficient includes: The ratio of the nearest distance threshold to the nearest distance from the material to the gripping position is used as the first overall deviation feature value. The sum of the ratio of the deviation quantity of incoming materials to the threshold of the deviation quantity of incoming materials and the ratio of the proportion of incoming material stacking area to the threshold of the proportion of incoming material stacking area is used as the second overall deviation characteristic value. The weighted sum of the first overall deviation characteristic value and the second overall deviation characteristic value is determined as the overall deviation characterization coefficient.

[0044] Specifically, the closest distance between the deviated incoming material and the gripping position directly reflects the immediate threat posed by the deviated material to the gripping station, which is related to production safety and continuity. If not handled promptly, the robot may directly grip the deviated material, leading to gripping failure, material damage, or even equipment jamming, causing production interruption. This feature quantifies the time window for immediate intervention and belongs to the highest priority "safety-type" indicator. The two features involved in the material distribution characteristics mainly reflect the scale and spatial impact of batch deviations, which can be gradually mitigated through scheduling. Even if the number of deviations is large and the stacked area is large, as long as the distance from the gripping position is still far, the robot can still plan and calibrate the operation in batches in an orderly manner, which may not necessarily cause an accident immediately. It has more of an impact on workload and efficiency than an emergency risk. Therefore, in the weighted summation, the first overall deviation characteristic value is given a higher weight coefficient, which can be selected in the range of 0.6 to 0.7. In this embodiment, it is set to 0.65. Correspondingly, the weight coefficient of the second overall deviation characteristic value can be selected in the range of 0.3 to 0.4. In this embodiment, it is set to 0.35. Those skilled in the art can adaptively fine-tune it according to the sensitivity of the production line to urgency.

[0045] In this embodiment, the purpose of setting the nearest distance threshold, the deviation in incoming material quantity threshold, and the incoming material stacking area ratio threshold is to characterize the situation where the actual distribution of batch deviation incoming materials is relatively chaotic and has a significant impact on production operations. This is achieved by acquiring historical image data of incoming materials corresponding to several batches of material transfer processes on the production line, and calling historical data on the nearest distance of the deviation incoming materials from the capture position, historical data on the deviation in incoming material quantity, and historical data on the incoming material stacking area ratio. The average nearest distance, average deviation in incoming material quantity, and average incoming material stacking area ratio are calculated and used as benchmark values ​​under normal conditions. Based on the purpose of setting the above three thresholds, the nearest distance threshold is determined as the product of the average nearest distance and the distance deviation coefficient; the deviation in incoming material quantity threshold is determined as the product of the average deviation in incoming material quantity and the quantity deviation coefficient; and the incoming material stacking area ratio threshold is determined as the product of the average incoming material stacking area ratio and the stacking deviation coefficient.

[0046] Specifically, considering that in actual production, the distance between the incoming material with deviation and the nearest gripping position may fluctuate randomly, and the quantity and stacking area ratio of the material with deviation may also vary slightly due to batch differences, if the threshold is set improperly, normal operating conditions may be misjudged as not meeting the placement deviation threshold, leading to unnecessary pauses and calibration operations. Therefore, based on the direction of each characteristic quantity, the threshold is shifted towards the direction that "does not easily increase the overall deviation characterization coefficient": The distance deviation coefficient is selected within the range [0.8, 0.9], with a coefficient less than 1, allowing for normal distance fluctuations. The coefficient only significantly affects the material deviation when it is very close to the gripping position and requires emergency handling. In practice, a value of 0.85 is preferred.

[0047] The quantity deviation coefficient is selected within the range [1.2, 1.4], with a coefficient greater than 1, allowing for occasional small deviations. A chaotic state is only determined when the quantity significantly exceeds the norm, such as in cases of batch anomalies. In practice, a coefficient of 1.3 is preferred.

[0048] The stacking deviation coefficient is selected within the range [1.1, 1.3]. A coefficient greater than 1 allows for minor stacking, and a significant impact is only considered when the proportion is significantly high, such as when there is dense compression in space. In practice, a coefficient of 1.2 is preferred.

[0049] The selection of the above-mentioned preferred values ​​achieves a good balance between effectively identifying batch deviations and avoiding overreaction. Of course, those skilled in the art can readjust the value range based on the actual stability of the production line and historical deviation data, or adjust the coefficients within a given range.

[0050] Specifically, since materials in cosmetic production lines are continuously conveyed in batches on the transmission device, deviations in a single material may be localized or represent a batch deviation problem across the entire transmission system. Therefore, this invention extracts the spatial location characteristics of the deviated material and the batch distribution characteristics within a predetermined transmission range, analyzes the overall deviation characterization coefficient, and achieves a quantitative assessment of the overall deviation status of all materials within a specific transmission interval. This upgrades deviation identification from a "point" to a "surface" level of operational control, avoiding the neglect of the overall deviation trend of the production line due to focusing only on a single deviated material, and ensuring that subsequent calibration decisions are more aligned with the actual operating state of the production line. This invention comprehensively reflects the overall deviation characteristics within a predetermined transmission range from three dimensions: time urgency, deviation scale, and spatial impact range. In the spatial location dimension, the time urgency and operational window for deviation calibration are quantified based on the closest distance between the deviated material and the gripping position. The smaller this characteristic data, the closer the deviated material is to the gripping position, indicating a shorter time window for the robot to perform calibration operations and a higher calibration urgency. If not handled promptly, the deviated material will quickly enter the gripping stage, easily causing robot gripping errors, material damage, and equipment jamming. The larger the value of this feature, the farther the deviated material is from the gripping position, indicating more time for calibration operations. The robot can plan a more reasonable calibration path without emergency intervention. Simultaneously, this feature quantifies the immediate threat posed by the deviated material to the gripping station, serving as a core spatial indicator for determining whether the production line needs to initiate an emergency calibration operation. In terms of quantity, the number of deviated materials quantifies the scale and frequency of deviations within the predetermined transmission range. A larger value indicates more deviated materials within the predetermined transmission range, suggesting the deviation is not a localized, sporadic issue but rather a malfunction in the production line's transmission device or feeding equipment. The overall deviation is large, impacting subsequent continuous gripping operations. A smaller value indicates that the deviated material is only an isolated, sporadic occurrence, and the overall production line is operating well. Furthermore, in terms of spatial proportion, the ratio of the total stacked area of ​​all deviated materials within the predetermined transmission range to the total effective placement area within that range quantifies the spatial proportion and distribution density of deviated materials within the predetermined transmission range, indirectly reflecting the actual impact range of the deviation. A higher value for this characteristic indicates a higher proportion and denser distribution of the biased incoming material within the predetermined transmission range. This not only compresses the calibration operation space of the robot's end effector but may also cause the biased material to squeeze against the normal material, leading to secondary deviations. Conversely, a lower value indicates a lower proportion and more dispersed distribution of the biased material, resulting in greater space leeway for robot calibration operations, less impact on the arrangement and transmission of normal material, and fewer interfering factors in the calibration operation. This characteristic compensates for the insufficiency of "number of biased incoming materials" which only reflects the absolute number, further quantifying the actual spatial impact of the biased material. It is a key indicator for determining whether there are space constraints in robot calibration operations.Therefore, this invention combines the aforementioned three features to analyze the overall deviation characterization coefficient, achieving a three-dimensional and comprehensive quantification of the overall deviation conditions of the production line. It characterizes the actual distribution of batch-deviation incoming materials and their impact on production operations, providing data support for subsequent determination of whether the placement deviation threshold is met. This invention closely aligns with the actual operating conditions of high-speed transmission and batch-distribution of incoming materials in cosmetic production lines, achieving precise control over the overall deviation state of the production line and making robot calibration operations more adaptable.

[0051] Specifically, please refer to Figure 3 As shown, this is a logic diagram for determining whether the placement deviation threshold is met according to an embodiment of the present invention. Determining whether the placement deviation threshold is met includes: If the overall deviation characterization coefficient is less than the overall deviation characterization coefficient threshold, then the placement deviation threshold is satisfied. If the overall deviation characterization coefficient is greater than or equal to the overall deviation characterization coefficient threshold, then the placement deviation threshold is not met.

[0052] To determine the threshold for the overall deviation characterization coefficient, several time-domain segments confirmed as having good overall operating conditions and without complex interventions due to batch deviations are selected from historical production data of the same production line and similar incoming material specifications. The overall deviation characterization coefficient for each time-domain segment is calculated, and the mean q and standard deviation r of the sequence are obtained. The sum of the product of the standard deviation r and the sensitivity coefficient k2 and the mean q is used as the threshold for the overall deviation characterization coefficient. Furthermore, the threshold for the overall deviation characterization coefficient should be dynamically updated with each production batch to adapt to production line status drift. Simultaneously, the threshold for the overall deviation characterization coefficient should be coordinated with the nearest distance threshold, the threshold for the quantity of defective incoming materials, and the stacking area ratio threshold, and can be iteratively optimized by verifying normal operating conditions and the proportion of misjudgments.

[0053] The sensitivity coefficient k2 can be selected within the range [1.0, 2.0] based on the tolerance for missed detections of batch deviations; for example, k2 = 1.5. If sufficient historical data is available, the 90th percentile can also be used as the threshold.

[0054] Specifically, the process of identifying the concentrated area of ​​deviation includes: The predetermined transmission range is divided into several material distribution areas; Determine the positional dispersion of the incoming materials within each of the aforementioned incoming material distribution areas; The material distribution area corresponding to the minimum positional dispersion is locked as the deviation concentration area; Specifically, the standard deviation of the gripping distance of the end effector for the biased incoming material in each of the material distribution areas is calculated, and the standard deviation of the gripping distance is used as the positional dispersion.

[0055] Specifically, there is no specific limitation on the way the material distribution area is divided. It can be divided into several equal areas based on the device length of a single conveying cycle of the feeding conveyor, which can be used as the material distribution area.

[0056] In this embodiment, the maximum operating coverage area where the robot end effector can perform incoming material calibration operation is taken as the predetermined transmission range, and then the predetermined transmission range is equally divided into several sub-regions, each of which is the corresponding incoming material distribution area.

[0057] Specifically, please refer to Figure 4 As shown, this is a logic diagram for determining whether to pause the feeding transmission device according to an embodiment of the present invention. Determining whether to pause the feeding transmission device includes: Determine the area of ​​empty material location in the deviation concentration region and the gap width between the end effector and the deviation material; The sum of the ratio of the area of ​​the empty material location to the threshold value of the empty material location and the ratio of the gap width to the threshold value of the gap width is used as the space margin. If the space margin is greater than or equal to the space margin threshold, it is determined that there is no need to pause the feeding transmission device. If the space margin is less than the space margin threshold, the feeding transmission device will be paused.

[0058] In this embodiment, the purpose of setting the empty material location area threshold and the gap width threshold is to characterize the sufficiency of the operating space of the end effector. This is achieved by acquiring historical data of incoming material images corresponding to several batches of material conveying processes on the production line, and by calling historical data of the empty material location area and the historical data of the gap width when the end effector is positioned between the deviating incoming materials. The average empty material location area and the average gap width are calculated and used as the baseline values ​​under normal conditions. Based on the purpose of setting the above two thresholds, the empty material location area threshold is determined as the product of the average empty material location area and the area deviation coefficient, and the gap width threshold is determined as the product of the average gap width and the width deviation coefficient.

[0059] Specifically, considering that the area of ​​the empty material location and the width of the gap may fluctuate randomly in actual production, such as slight displacement of incoming material or measurement error, in order to avoid unnecessary pauses in operation due to normal fluctuations and to ensure the continuous transmission efficiency of the production line, this embodiment shifts the threshold towards the direction of "not being able to determine insufficient operating space", that is, the threshold is lower than the historical average, thereby reducing the sensitivity to space constraints and prioritizing the maintenance of production continuity.

[0060] The area deviation coefficient is selected within the range [0.85, 0.95]. A coefficient less than 1 ensures that the threshold is below the average level, allowing for a certain degree of reduction in the empty material area while still considering sufficient space, thus avoiding misjudgment of space shortage due to occasional gap changes. In practice, a coefficient of 0.9 is preferred.

[0061] The width deviation coefficient is selected within the range [0.85, 0.95], with a coefficient less than 1, so that the threshold is slightly lower than the mean, accommodating normal width fluctuations, and only triggering the insufficient space judgment when the gap becomes significantly narrower. In practice, a coefficient of 0.9 is preferred.

[0062] The selection of the above-mentioned preferred values ​​takes into account both operational safety and production efficiency. Of course, those skilled in the art can readjust the value range according to the regularity of incoming materials, the size of the end effector, and the tolerance for production interruption, or they can adjust the coefficients within a given range.

[0063] The space margin threshold is predetermined. The space margin calculated is determined as the space margin threshold when the area of ​​the empty material position is equal to the empty material position area threshold, and the gap width between the end effector and the deviation material is equal to the gap width threshold.

[0064] Specifically, this invention constructs a logical chain of "quantification of dispersed location - locking of concentrated deviation areas - assessment of spatial redundancy". First, calibration operations are concentrated in the core area with the highest concentration of deviations, avoiding path redundancy caused by dispersed operations, significantly improving the calibration efficiency of batches of defective materials, and achieving "focused and intensive" calibration operations, adapting to the calibration efficiency requirements of high-speed transmission in cosmetic production lines. In reality, cosmetic materials are mostly irregularly shaped parts such as bottles and cans, and production lines often adopt a dense arrangement to improve efficiency, which easily leads to situations where defective materials are squeezed together and the gaps are narrow, placing higher demands on the robot's operating space. Therefore, before controlling the robot to perform calibration operations, it is considered whether to pause the transmission device to ensure that the robot's end effector has sufficient physical operating space in the concentrated deviation area. If transmission continues despite insufficient space, the end effector is prone to collisions with the material or conveyor belt, or incomplete calibration operations. This invention selects two core indicators: the area of ​​the empty material location and the width of the gap between the end effector and the material, respectively, to determine the spatial redundancy from the dimensions of overall space and local operating space, comprehensively quantifying the actual usable operating space of the robot in the concentrated deviation area. The aforementioned two indicators directly correspond to the physical operational limits of the end effector, such as the size of the gripper and the minimum space required for rotation / translation, ensuring a high degree of match between the space margin assessment results and the robot's actual operational capabilities. This allows for continuous production line operation without interruption when there is sufficient operating space, while also enabling timely pauses when operating space is insufficient, providing a stable and safe calibration environment for the robot and mitigating production risks such as equipment collisions and material damage at the source. This invention solves the problem of difficulty in identifying concentrated deviation areas under dense material arrangement by locking the deviation concentration area; and it accurately adapts to the operating space requirements under dense arrangement of irregularly shaped materials through a two-dimensional spatial assessment of the empty material location area and the gap width between incoming materials. This effectively determines the operable space of the end effector in narrow gaps, avoiding calibration operation failures caused by dense material arrangement. Furthermore, it adapts to the production needs of different categories of cosmetic materials, possessing good scenario adaptability and scalability.

[0065] Specifically, determining the sequence of position calibration and adjustment operations performed by the end effector for the deviated incoming material includes: If the determination result is to pause the feeding transmission device, then based on the positional distance between the deviation material and the corresponding end effector of the robot, the deviation material is sorted in descending order, and the end effector is controlled to perform position calibration and adjustment operations in sequence according to the sorting. If the determination result is that there is no need to pause the feeding transmission device, then based on the distribution deviation characteristics and the same degree of gripping direction of the gripping position corresponding to the deviation material, the operation limitation characterization parameters are evaluated to adjust the pose correction range of the end effector. The end effector is controlled to preferentially perform position calibration and adjustment operations on the incoming material with deviation within the deviation concentration area; The distribution deviation characteristics include the tilt angle of the incoming material and the positional deviation angle. Specifically, the process of evaluating the operationally constrained characterization parameters includes: The sum of the ratio of the placement tilt angle of the biased incoming material to the placement tilt angle threshold and the ratio of the position deviation angle to the position deviation angle threshold is used as the first operation-restricted feature value. The ratio of the grip direction consistency threshold to the grip direction consistency of the corresponding grip position of the deviation material is used as the second operation-restricted feature value. The weighted sum of the first operation-restricted feature value and the second operation-restricted feature value is used to determine the operation-restricted characterization parameter.

[0066] Specifically, the uniformity of the gripping direction reflects the directional consistency of incoming materials in a batch, directly affecting the efficiency and collision risk of continuous robot calibration operations. Low uniformity of the gripping direction means that the optimal gripping direction for each incoming material within the area differs; for example, some require gripping from the left, while others require gripping from the front. During continuous calibration, the robot must frequently change the posture of its end effector, such as rotating or translating, leading to a significant increase in calibration time, more complex motion paths, and increased susceptibility to interference with surrounding incoming materials or transmission devices. This characteristic reflects the collaborative constraints of batch operations and is crucial in determining calibration efficiency under non-pause conditions. The placement tilt angle and positional deviation angle of individual incoming materials can be locally adapted by expanding the pose correction range, resulting in less overall restriction on continuous operations. Therefore, in the weighted summation, the second operation-restricted characteristic value is assigned a higher weight coefficient, which can be selected within the range of 0.6 to 0.7; in this embodiment, it is set to 0.65. Correspondingly, the weight coefficient of the first operation-restricted characteristic value can be selected within the range of 0.3 to 0.4; in this embodiment, it is set to 0.35. Those skilled in the art can adaptively fine-tune this according to the production line's sensitivity to batch directional consistency.

[0067] In this embodiment, the purpose of setting the placement tilt angle threshold, position deviation angle threshold, and grasping direction consistency threshold is to characterize situations where the robot calibration operation is severely restricted and requires a large range of pose correction. This is achieved by acquiring historical material image data corresponding to several batches of incoming materials being conveyed on the production line, and by calling historical data on the placement tilt angle, position deviation angle, and grasping direction consistency of the corresponding grasping positions of the deviated materials. The average placement tilt angle, average position deviation angle, and average grasping direction consistency are calculated and used as the baseline values ​​under normal conditions. Based on the purpose of setting the above three thresholds, the placement tilt angle threshold is determined as the product of the average placement tilt angle and the tilt deviation coefficient; the position deviation angle threshold is determined as the product of the average position deviation angle and the position deviation coefficient; and the grasping direction consistency threshold is determined as the product of the average grasping direction consistency and the direction deviation coefficient.

[0068] Specifically, considering that in actual production, there are normal slight fluctuations in the tilt and positional deviation of the incoming material, and the sameness of the gripping direction is usually kept at a high level, if the threshold is set improperly, normal working conditions will be misjudged as severely restricted, causing the pose correction range to be unnecessarily expanded and reducing calibration efficiency. Therefore, based on the index direction of each characteristic quantity, the threshold is shifted towards the direction that "does not easily increase the characteristic parameters of operation restriction": The tilt deviation coefficient is selected within the range [1.2, 1.4]. A coefficient greater than 1 allows for normal, slight tilting. Only when the tilt is significant, such as when the bottle is noticeably skewed, is a large-scale rotation correction considered necessary. In practice, a coefficient of 1.3 is preferred.

[0069] The position deviation coefficient is selected within the range [1.2, 1.4]. A coefficient greater than 1 allows for normal orientation fluctuations. Only when the deviation is significant, such as when the incoming material deviates from the standard layout path, is a large-scale orientation adjustment considered necessary. In practice, a coefficient of 1.3 is preferred.

[0070] The directional deviation coefficient is selected within the range [0.8, 0.9]. A coefficient less than 1 ensures that the threshold is below the average level, allowing for some dispersion in the gripping direction. Only when the same degree is significantly low, such as when the gripping direction of incoming materials in a chaotic area, does it reflect severe operational limitations. In practice, a coefficient of 0.85 is preferred.

[0071] The selection of the above-mentioned preferred values ​​balances the timeliness of restricted identification with calibration efficiency. Of course, those skilled in the art can readjust the value range according to the consistency of incoming material specifications and the accuracy of robot operation, or they can adjust the coefficients within a given range.

[0072] Specifically, the placement tilt angle refers to the angle between the vertical center axis of the incoming material and the vertical baseline of the production line.

[0073] The vertical reference line is the normal direction perpendicular to the horizontal reference plane of the production line. This angle quantifies the degree of tilt deviation of the incoming material on the vertical surface, reflecting whether the incoming material maintains a standard vertical placement posture. The larger the value, the more severe the tilt of the incoming material's vertical surface, and the higher the rotational compensation required for the robot to calibrate its posture.

[0074] Specifically, the positional deviation angle refers to the angle between the line connecting the center reference point of the deviation material and the standard layout path of the production line, and the direction line of the corresponding preset point on the standard layout path.

[0075] The central reference point is the geometric center of the incoming cosmetic material, such as the intersection of the central axis of the bottle / can and the bottom contact surface. The standard layout path is the conveying trajectory of the incoming material on the transmission device. The preset point refers to the point on the standard layout path of the production line that corresponds directly to the actual transmission position of the incoming material. This angle quantifies the degree of planar orientation deviation of the incoming material within the horizontal reference plane of the production line, reflecting the orientation deviation between the actual layout position and the standard layout position. The larger this value, the more significant the orientation deviation of the incoming material in the horizontal plane, and the higher the planar pose adjustment requirement for robot position calibration.

[0076] Specifically, the uniformity of the gripping direction of the corresponding gripping position of the deviation material refers to the degree of consistency of the standard gripping direction of all deviation materials at their corresponding gripping positions within the same deviation concentration area.

[0077] In this embodiment, by determining the standard gripping direction of each deviated material within the deviation concentration area, each standard gripping direction is converted into a unit direction vector in a unified coordinate system, and the average direction vector of each unit direction vector is calculated and normalized to a unit average direction vector; the cosine value of the angle between the unit direction vector of each deviated material and the unit average direction vector is calculated, and the ratio of the sum of the cosine values ​​to the total number of deviated materials is used as the gripping direction uniformity of the gripping position corresponding to the deviated material, which will not be elaborated further.

[0078] Specifically, adjusting the pose correction range of the end effector includes: Increase the pose correction range, and the increase in the pose correction range is positively correlated with the operation-limited characterization parameter; The pose correction range includes a translation compensation range and a rotation compensation range.

[0079] In this embodiment, optionally, The operation-restricted characterization parameters are compared with preset first operation-restricted characterization parameter comparison thresholds and second operation-restricted characterization parameter comparison thresholds. When the operation-restricted representation parameter is greater than the second operation-restricted representation parameter comparison threshold, the increase in pose correction range is determined as the first increase, which is set to be 0.4 times the upper and lower limits of the reference pose range, respectively. When the operation-restricted representation parameter is greater than or equal to the first operation-restricted representation parameter comparison threshold and less than or equal to the second operation-restricted representation parameter comparison threshold, the increment of the pose correction range is determined to be the second increment. The second increment is set to be 0.3 times the upper limit and lower limit of the reference pose range, respectively. When the operation-restricted representation parameter is less than the first operation-restricted representation parameter comparison threshold, the increase in pose correction range is determined to be the third increase. The third increase is set to be 0.2 times the upper limit and lower limit of the reference pose range, respectively. The first operationally restricted characterization parameter comparison threshold is 1.1 times the operationally restricted characterization parameter threshold, and the second operationally restricted characterization parameter comparison threshold is 1.3 times the operationally restricted characterization parameter threshold.

[0080] To determine the threshold for the operational limitation characterization parameter, several time-domain segments are selected from historical production data of the same production line and similar incoming material specifications. These segments are identified as having slightly limited robot calibration operation and are capable of completing calibration within the initial pose range. The operational limitation characterization parameter for each time-domain segment is calculated, and the mean v and standard deviation x of the sequence are obtained. The sum of the product of the standard deviation x and the sensitivity coefficient k3 and the mean v is used as the threshold for the operational limitation characterization parameter. Furthermore, the threshold for the operational limitation characterization parameter should be dynamically updated with each production batch to adapt to equipment state drift. Simultaneously, the threshold for the operational limitation characterization parameter should be coordinated with the placement tilt angle threshold, position deviation angle threshold, and grasping direction consistency threshold, and can be iteratively optimized by verifying normal operating conditions and the proportion of misjudgments.

[0081] The sensitivity coefficient k3 can be selected within the range [1.0, 2.0] based on the tolerance for missed detections due to operational limitations; for example, k3 = 1.5. If sufficient historical data is available, the 90th percentile can also be used as the threshold.

[0082] Understandably, the purpose of increasing the pose correction range is to achieve a precise and hierarchical dynamic match between the pose correction range of the robot's end effector and the actual operational constraints of the deviation concentration area. This is achieved by quantifying the increment to adapt to different levels of operational constraints. This ensures sufficient adjustment space under highly constrained conditions, guaranteeing effective calibration within a limited window; while under less constrained conditions, it compresses the adjustment range, improving motion efficiency and preventing materials from moving out of the operating area due to excessively long motion times. Ultimately, this achieves a high degree of matching between the robot's calibration operation and the continuous drive rhythm of the production line, ensuring calibration effectiveness without interrupting production, balancing production continuity and calibration accuracy, and adapting to the continuous drive operation characteristics of cosmetic production lines. Furthermore, the increment of the pose correction range under different conditions can also be adjusted by those skilled in the art.

[0083] Specifically, the initial pose range of the robot end effector is the basic operating pose range set after the production line is debugged, for standard unbiased incoming materials and preset standard layout conditions. Therefore, in implementation, the initial pose range can be used as the reference pose range, or other methods can be used to determine the reference pose range, which will not be elaborated here.

[0084] Specifically, this invention designs a calibration operation planning stage for non-pause operation. The aim is to adapt the robot end effector calibration operation to the production line's transmission rhythm while maintaining continuous transmission without interruption. This ensures the calibration operation accurately adapts to the actual constraints of the deviation conditions, is efficiently completed in continuous production, and maximizes the stability and accuracy of the calibration operation. In practice, for non-pause operation, the core challenge of robot calibration is matching the pose correction range with the degree of deviation constraint. For example, a correction range that is too small will result in incomplete calibration of the deviated incoming material, while a correction range that is too large will increase actuator redundancy, reduce calibration efficiency, and even cause interference with surrounding incoming material / transmission devices due to excessive movement amplitude. Therefore, this invention characterizes the actual constraints of robot calibration operation from two dimensions: the inherent deviation characteristics of the incoming material and the characteristics of the batch grasping direction. In terms of elevation geometry, the tilt deviation of a single deviated incoming material is quantified by the placement tilt angle, reflecting the basic requirements for robot rotation correction. The larger the value of this feature, the more severe the material tilting. For example, if a cosmetic bottle is tilted to one side, the robot's end effector needs a larger rotational compensation range to calibrate it to a standard posture, resulting in stronger rotational constraints during operation. The smaller the value, the closer the material is to a vertical standard posture, requiring only minor rotational correction or even no rotation at all, resulting in lower restrictions on rotational operations. In terms of planar orientation, the degree of orientational deviation of the material's spatial position is quantified by the deviation angle of the material's position, reflecting the robot's orientational adaptation requirements for translation / rotation. The larger the value, the more significant the orientational offset of the material within the transmission plane. The robot needs to adjust the planar operating orientation of the end effector, and may even need to combine translational and rotational corrections for accurate calibration, resulting in stronger orientational constraints during operation. The smaller the value, the closer the spatial orientation of the material is to a baseline arrangement, requiring the robot to make significant adjustments to the operating orientation, resulting in lower restrictions on planar operations. The aforementioned two features form a two-dimensional geometric deviation quantification of the planar and vertical dimensions, fully characterizing the overall posture deviation of a single material with deviation, which is the core basis for robot pose correction. Simultaneously, by combining the overall quantitative index for all incoming materials with deviations within the deviation concentration area—that is, the consistency of the gripping direction at the corresponding gripping position of the incoming materials with deviations—the consistency of the gripping direction of batch incoming materials within the deviation concentration area is quantified, reflecting the directional coordination constraints of the robot's batch operations. The lower this characteristic value, the more chaotic the gripping directions of the incoming materials within the area. For example, some materials need to be gripped from the left, while others need to be gripped from the front or right. The robot needs to expand the directional range of pose correction and frequently adjust the gripping direction to adapt to all incoming materials, resulting in low efficiency and strong directional constraints in continuous operation. Conversely, the higher this characteristic value, the more consistent the gripping directions of the incoming materials within the deviation concentration area. The robot does not need to frequently adjust the gripping direction, only requiring small pose corrections within a fixed directional range, resulting in high efficiency and weak directional constraints in continuous operation.This feature overcomes the limitations of distribution deviation characteristics, reflecting the robot's operational constraints from a batch operation perspective. It is a key indicator for adapting to continuous transmission and batch calibration conditions on production lines. Furthermore, the operational constraint characterization parameters are evaluated to represent the overall degree of constraint on the robot's calibration operations within the deviation concentration area under non-pause conditions. This invention achieves a perfect match between robot calibration operations and the continuous transmission rhythm of the production line through dynamically adapted pose correction and efficient operation sequence planning, ensuring both the accuracy and efficiency of calibration operations while maintaining production continuity. Simultaneously, adapting to the actual operating conditions of cosmetic production lines and the technical characteristics of the embodied robot is crucial for ensuring the efficient operation of the dynamic calibration system under non-pause conditions.

[0085] If the dynamic calibration method of the cosmetic intelligent manufacturing robot of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0086] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A dynamic calibration method for a cosmetic intelligent manufacturing robot, characterized in that, include: Real-time acquisition of incoming material image data for corresponding batches on the production line to extract the contact state features of the incoming materials, including the horizontal deviation angle of the contact material and the contact area of ​​the contact material. Based on the aforementioned contact state characteristics and material spacing deviation values, the placement deviation characterization value of the incoming material is calculated to mark the incoming material. In response to the presence of incoming materials marked as deviating from the standard, the operating conditions of the production line are evaluated, analyzed, and adjusted, including: Determine the nearest distance between the deviated incoming material and the gripping position, as well as the material distribution characteristics within the predetermined transmission range, analyze the overall deviation characterization coefficient to determine whether the placement deviation threshold is met, and control the robot to perform position calibration and adjustment operations on the deviated incoming material. If the placement deviation threshold is not met, the current posture image of the robot is obtained, the deviation concentration area is locked, and based on the space margin for the end effector to operate on the deviation concentration area, it is determined whether to pause the feeding transmission device. Based on the judgment result, the sequence of position calibration and adjustment operations performed by the end effector for the deviated incoming material is determined; If the placement deviation threshold is met, the materials with deviation are sorted in descending order based on the positional distance between the materials with deviation and the corresponding end effector of the robot, and the end effector is controlled to perform position calibration and adjustment operations in sequence according to the sorting. The incoming material distribution characteristics include the quantity of materials with deviations and the percentage of the stacked area of ​​incoming materials.

2. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 1, characterized in that, The process of calculating the placement deviation characterization value of the incoming material includes: The sum of the ratio of the horizontal deviation angle of the incoming material to the horizontal deviation angle threshold and the ratio of the bonding area of ​​the incoming material to the bonding area threshold is used as the first placement deviation characteristic value. The ratio of the material spacing deviation value to the material spacing deviation threshold is used as the second placement deviation characteristic value; The first placement deviation feature value and the second placement deviation feature value are weighted and summed to determine the placement deviation characterization value.

3. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 2, characterized in that, The incoming materials are labeled, including: If the placement deviation characterization value of the incoming material is greater than or equal to the placement deviation characterization threshold, the incoming material is marked as a deviation incoming material.

4. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 1, characterized in that, The process of analyzing the overall deviation characterization coefficient includes: The ratio of the nearest distance threshold to the nearest distance from the material to the gripping position is used as the first overall deviation feature value. The sum of the ratio of the deviation quantity of incoming materials to the threshold of the deviation quantity of incoming materials and the ratio of the proportion of incoming material stacking area to the threshold of the proportion of incoming material stacking area is used as the second overall deviation characteristic value. The weighted sum of the first overall deviation characteristic value and the second overall deviation characteristic value is determined as the overall deviation characterization coefficient.

5. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 4, characterized in that, Determining whether the placement deviation threshold is met includes: If the overall deviation characterization coefficient is less than the overall deviation characterization coefficient threshold, then the placement deviation threshold is satisfied.

6. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 1, characterized in that, The process of identifying the concentrated area of ​​deviation includes: The predetermined transmission range is divided into several material distribution areas; Determine the positional dispersion of the incoming materials within each of the aforementioned incoming material distribution areas; The material distribution area corresponding to the minimum positional dispersion is locked as the deviation concentration area; Specifically, the standard deviation of the gripping distance of the end effector for the biased incoming material in each of the material distribution areas is calculated, and the standard deviation of the gripping distance is used as the positional dispersion.

7. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 1, characterized in that, Determining whether to pause the feeding transmission device includes: Determine the area of ​​empty material location in the deviation concentration region and the gap width between the end effector and the deviation material; The sum of the ratio of the area of ​​the empty material location to the threshold value of the empty material location and the ratio of the gap width to the threshold value of the gap width is used as the space margin. If the space margin is greater than or equal to the space margin threshold, it is determined that there is no need to pause the feeding transmission device.

8. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 7, characterized in that, Determining the sequence of position calibration and adjustment operations performed by the end effector for the deviated incoming material includes: If the determination result is to pause the feeding transmission device, then based on the positional distance between the deviation material and the corresponding end effector of the robot, the deviation material is sorted in descending order, and the end effector is controlled to perform position calibration and adjustment operations in sequence according to the sorting. If the determination result is that there is no need to pause the feeding transmission device, then based on the distribution deviation characteristics and the same degree of gripping direction of the gripping position corresponding to the deviation material, the operation limitation characterization parameters are evaluated to adjust the pose correction range of the end effector. The end effector is controlled to preferentially perform position calibration and adjustment operations on the incoming material with deviation within the deviation concentration area; The distribution deviation characteristics include the tilt angle of the incoming material and the position deviation angle.

9. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 8, characterized in that, The process of evaluating operationally constrained characterization parameters includes: The sum of the ratio of the placement tilt angle of the biased incoming material to the placement tilt angle threshold and the ratio of the position deviation angle to the position deviation angle threshold is used as the first operation-restricted feature value. The ratio of the grip direction consistency threshold to the grip direction consistency of the corresponding grip position of the deviation material is used as the second operation-restricted feature value. The weighted sum of the first operation-restricted feature value and the second operation-restricted feature value is used to determine the operation-restricted characterization parameter.

10. The dynamic calibration method for the cosmetic intelligent manufacturing robot according to claim 9, characterized in that, Adjusting the pose correction range of the end effector includes: Increase the pose correction range, and the increase in the pose correction range is positively correlated with the operation-limited characterization parameter; The pose correction range includes a translation compensation range and a rotation compensation range.

Images