An intelligent pharmaceutical product palletizing control method and system

Abstract

The present invention discloses an intelligent pharmaceutical product palletizing control method and system, relating to the field of pharmaceutical product palletizing technology. The palletizing production line includes: a box-shaped grabbing robot with a single-side opening and an infrared sensor disposed within the box-shaped grabbing robot. The control method includes: obtaining conveying image data of pharmaceutical products at a first detection position based on a visual detection mechanism; extracting a posture contour image of the pharmaceutical products from the conveying image data, and adjusting the gripping posture of the box-shaped grabbing robot based on the posture contour image; performing a gripping operation on pharmaceutical products entering a gripping area, detecting the position of pharmaceutical product information labels based on an infrared sensor, and extracting pharmaceutical product information label position data; obtaining pharmaceutical product palletizing position data, and adjusting the palletizing posture of the box-shaped grabbing robot based on the information label position data. The system can improve the pharmaceutical product posture adjustment accuracy and palletizing efficiency, ensure that information labels are accurately exposed, and thus enhance product information management and traceability efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of pharmaceutical palletizing, and in particular to an intelligent pharmaceutical product palletizing control method and system. Background Art

[0002] On pharmaceutical production lines, box-packaged pharmaceutical products need to be stacked and organized for efficient organization. Currently, pharmaceutical boxes are typically neatly stacked, then plastic-sealed with a film on the outside, and then boxed. To facilitate drug information detection and traceability, information labels are placed on the outside of the plastic-sealed finished products. Existing robotic handling devices use suction cups to hold pharmaceutical products in place and adjust their position according to pre-set procedures. These robots cannot adapt to the label information, hindering the ease of detecting information during palletizing. Summary of the Invention

[0003] The purpose of the present invention is to overcome the shortcomings of the existing technology. The present invention provides an intelligent pharmaceutical product palletizing control method and system. By obtaining the posture status of the pharmaceutical product and the position data of the information label, it has the advantages of improving the posture adjustment accuracy and palletizing efficiency of the pharmaceutical product, ensuring that the information label is accurately exposed to improve product information management and traceability efficiency.

[0004] The present invention provides an intelligent pharmaceutical product palletizing control method, which is applicable to a pharmaceutical product palletizing production line, wherein the palletizing production line comprises: a box-shaped grabbing manipulator with a single-side opening and an infrared sensor arranged inside the box-shaped grabbing manipulator;

[0005] The control method includes:

[0006] Acquiring conveying image data of the pharmaceutical product at a first inspection position based on a visual inspection mechanism;

[0007] extracting a posture contour image of the pharmaceutical product from the conveying image data, and adjusting the gripping posture of the box-shaped gripping manipulator according to the posture contour image;

[0008] Grab the pharmaceutical products that enter the grabbing area, detect the location of the pharmaceutical product information label based on the infrared sensor, and extract the pharmaceutical product information label location data;

[0009] The stacking position data of the pharmaceutical product is obtained, and the stacking posture of the box-shaped grasping robot is adjusted in combination with the information label position data.

[0010] Furthermore, the method of obtaining the transport image data of the pharmaceutical product at the first inspection position based on the visual inspection mechanism includes:

[0011] Setting up a visual inspection mechanism at the first inspection position of the pharmaceutical product conveying line;

[0012] Based on the visual inspection mechanism, transport image data of the pharmaceutical product passing through the first inspection position is obtained.

[0013] Furthermore, extracting a posture contour image of the pharmaceutical product from the conveying image data and adjusting the gripping posture of the box-shaped gripping manipulator according to the posture contour image includes:

[0014] Extracting the contour of the pharmaceutical product from the transport image data based on the Sobel operator to obtain a posture contour image of the pharmaceutical product;

[0015] Dynamically matching the pose contour image of the pharmaceutical product with a pre-stored standard contour image to obtain the pose rotation variables between the pharmaceutical product and the standard contour image;

[0016] The gripping posture of the box-shaped gripping manipulator is adjusted according to the posture rotation variable.

[0017] Furthermore, the dynamically matching the posture contour image of the pharmaceutical product with the pre-stored standard contour image to obtain the posture rotation variable between the pharmaceutical product and the standard contour image includes:

[0018] The pose contour image and the standard contour image are set as superimposed layers for comparison, the layer where the pose contour image is located is set to a fixed state, and the layer where the standard contour image is located is set to a rotating state;

[0019] The standard contour image is rotated clockwise at a preset unit angle. When the posture contour image overlaps with the standard contour image, the rotation angle of the standard contour image is obtained, and the rotation angle of the standard contour image is set as a posture rotation variable.

[0020] Furthermore, adjusting the gripping posture of the box-shaped gripping manipulator according to the posture rotation variable includes:

[0021] adjusting the initial posture of the box-shaped grasping manipulator according to the standard contour image;

[0022] A rotation adjustment instruction is generated based on the posture state variable, and the box-shaped grasping manipulator is driven to rotate the rotation angle in a clockwise direction based on the rotation adjustment instruction to form a grasping posture of the box-shaped grasping manipulator.

[0023] Furthermore, the grabbing operation of the pharmaceutical product entering the grabbing area, detecting the position of the pharmaceutical product information label based on the infrared sensor, and extracting the pharmaceutical product information label position data includes:

[0024] Driving the box-shaped grabbing robot to move to a position above the pharmaceutical product, and covering and adsorbing the pharmaceutical product with the box-shaped grabbing robot;

[0025] Performing information label recognition and scanning on the pharmaceutical product inside the box-shaped grabbing robot based on an infrared sensor;

[0026] The label information location data of the pharmaceutical product is marked based on the scanning results.

[0027] Furthermore, the label information location data of the pharmaceutical product marked based on the scanning result includes:

[0028] Obtain the specifications of pharmaceutical products on the pharmaceutical product conveyor line and obtain the label setting location data of pharmaceutical products;

[0029] The label information position data is generated based on the scanning result of the infrared sensor of the box-shaped grasping robot and combined with the label setting orientation data.

[0030] Furthermore, the grabbing operation of the pharmaceutical product entering the grabbing area, detecting the position of the pharmaceutical product information label based on the infrared sensor, and extracting the pharmaceutical product information label position data further includes:

[0031] The information label of the pharmaceutical product is identified based on scanning by the infrared sensor, and the information label data of the pharmaceutical product is read based on the scanning.

[0032] Furthermore, the obtaining of the palletizing position data of the pharmaceutical product and adjusting the palletizing posture of the box-shaped grabbing manipulator in combination with the information label position data includes:

[0033] Obtaining the palletizing position data of the current pharmaceutical product according to the palletizing process steps, and extracting the palletizing space state and palletizing display surface of the pharmaceutical product according to the palletizing position data;

[0034] The stacking posture of the pharmaceutical product of the box-shaped grasping robot is adjusted according to the state of the stacking space, and the position of the information label of the pharmaceutical product is adjusted to be located on the stacking display surface.

[0035] The present invention also provides an intelligent pharmaceutical product palletizing control system, the control system is used to execute the intelligent pharmaceutical product palletizing control method, the control system includes:

[0036] Visual recognition component: used to obtain the posture status data of the pharmaceutical product at the first detection position;

[0037] A first posture adjustment component: used for adjusting the gripping posture of the box-shaped gripping manipulator according to the posture state data;

[0038] Information processing component: used to detect and obtain the location data of information labels of pharmaceutical products;

[0039] The second posture adjustment component is used to adjust the stacking posture of the box-shaped grasping robot according to the position data of the information tag.

[0040] The present invention provides an intelligent pharmaceutical product palletizing control method and system. The method obtains the conveying posture state of pharmaceutical products through image processing, and adjusts the palletizing posture of the pharmaceutical products when the robot grasps the pharmaceutical products in combination with the palletizing requirements of the pharmaceutical products, so that the box body information label of the pharmaceutical product can face the display surface of the palletizing pile. It has the advantages of improving the posture adjustment accuracy and palletizing efficiency of pharmaceutical products, ensuring that the information label is accurately exposed, thereby improving product information management and traceability efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of an intelligent pharmaceutical product palletizing control method according to an embodiment of the present invention;

[0042] Figure 2 Schematic diagram of an intelligent pharmaceutical product palletizing production line according to an embodiment of the present invention;

[0043] Figure 3 2 is a schematic structural diagram of a box-shaped grasping manipulator according to an embodiment of the present invention;

[0044] Figure 4 Schematic diagram of an intelligent pharmaceutical product palletizing control system in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0046] Example 1:

[0047] Figure 1 A flow chart of an intelligent pharmaceutical product palletizing control method according to an embodiment of the present invention is shown. The control method includes:

[0048] S11: Acquire conveying image data of the pharmaceutical product at the first inspection position based on the visual inspection mechanism.

[0049] Specifically, a visual inspection mechanism is provided on the pharmaceutical product conveying line so that the visual inspection mechanism can perform real-time inspection on a first inspection position on the pharmaceutical product conveying line, so as to obtain conveying image data of the pharmaceutical product passing through the first inspection position.

[0050] Furthermore, the colors of the box body of the pharmaceutical product and the conveyor belt of the conveyor line can be set to high-contrast colors, so as to extract the outline of the pharmaceutical product from the conveying image data.

[0051] S12: extracting a posture contour image of the pharmaceutical product from the conveying image data, and adjusting the gripping posture of the box-shaped gripping robot according to the posture contour image.

[0052] Specifically, by adopting an edge detection algorithm such as the Sobel operator, it is used to reflect the actual placement angle and position offset of the medicine, thereby providing a basis for calibrating the robot's grasping posture and improving the accuracy of the box-shaped grasping robot in grasping the medicine product.

[0053] S13: Grab the pharmaceutical product that enters the grabbing area, detect the position of the pharmaceutical product information label based on the infrared sensor, and extract the pharmaceutical product information label position data.

[0054] The infrared sensor refers to an optical detection device installed inside the box-shaped grasping robot. It can be implemented by using an infrared scanning module in conjunction with reflective label recognition technology. It is used to detect the physical position of the pharmaceutical product information label to ensure that the label is in the visible display surface after palletizing.

[0055] S14: Obtain the palletizing position data of the pharmaceutical product, and adjust the palletizing posture of the box-shaped grasping robot in combination with the information tag position data.

[0056] Palletizing position data refers to the spatial coordinates and posture parameters of the drug in the target pallet position. It can be generated through preset palletizing rules or real-time path planning algorithms to guide the robot to adjust the final placement angle of the drug so that the information label faces the preset display direction.

[0057] The core innovation of this application lies in the synergy between visual inspection and infrared sensing, combined with dynamic pose matching and label position feedback mechanisms, to achieve adaptive adjustment of the information label display surface during the drug grasping and palletizing process. Visual inspection captures the drug's posture in real time, while infrared sensors accurately locate the label's physical position, forming a closed-loop control logic to ensure label visibility after palletizing, while also overcoming the limitations of traditional robotic arms that rely on fixed programs.

[0058] Example 2:

[0059] Specifically, Figure 2 A schematic diagram of an intelligent pharmaceutical product palletizing production line according to an embodiment of the present invention is shown. Figure 3A structural schematic diagram of a box-shaped grasping robot 2 in an embodiment of the present invention is shown. The control method is applicable to a pharmaceutical product palletizing production line. The palletizing production line includes a box-shaped grasping robot 2 with a single-sided opening and an infrared sensor arranged inside the box-shaped grasping robot 2. The clamping component of the box-shaped grasping robot 2 is arranged as a box structure, and the box structure is adapted to the box shape of the pharmaceutical product, so that the pharmaceutical product can be accommodated in the box structure of the box grasping robot. The infrared sensor arranged on the inner side of the box structure performs an infrared scan on the box surface of the pharmaceutical product, thereby detecting the information label data on the box surface of the pharmaceutical product.

[0060] Furthermore, by setting a box-like structure 21 with a single-side opening, the box-like structure 21 of the grasping robot can cover the three outer side surfaces of the pharmaceutical product and the top surface of the box-like structure 21. The top surface refers to the outer surface of the pharmaceutical product box body facing upward when it is in a placed state; based on the infrared sensor, the information labels on the three outer side surfaces of the box body of the pharmaceutical product can be obtained, marked as A1, A2, and A3, and the information label on the top surface of the box body of the pharmaceutical product can be obtained, marked as B1. The label information position on the surface of the box body of the pharmaceutical product is obtained according to the infrared sensor, so as to adjust the posture state of the box body of the pharmaceutical product.

[0061] Furthermore, by providing a box-shaped grasping robot 2 with a single-side opening, the box-shaped robot has two sets of placement directions, so as to adjust the placement of the pharmaceutical products according to the stacking state of the box bodies of the pharmaceutical products.

[0062] The intelligent pharmaceutical product palletizing control method is applicable to a pharmaceutical product palletizing production line. The palletizing production line includes a box-shaped gripping robot 2 with a single-side opening and an infrared sensor mounted within the robot. The control method first acquires conveying image data of pharmaceutical products at a first inspection position using a visual inspection mechanism 1. Visual inspection mechanism 1 can be a high-speed camera mounted in a fixed position above the conveyor line to capture real-time images of passing pharmaceutical products.

[0063] Next, a contour image of the pharmaceutical product's posture is extracted from the conveyed image data. This step can be achieved using image processing algorithms such as edge detection. Based on the extracted contour image, the gripping posture of the box-shaped gripping robot 2 is adjusted. The robot's posture adjustment can be achieved using servo motor control to align the opening orientation with the product's posture.

[0064] When a pharmaceutical product enters the grabbing area, the box-shaped grabbing robot 2 performs a grabbing operation. After the grabbing is completed, the internal infrared sensor detects the position of the pharmaceutical product information label. The infrared sensor can scan the entire space inside the box and identify the precise location of the label.

[0065] Finally, the palletizing position data for the pharmaceutical products is obtained, which may come from a pre-set palletizing plan. Combined with the previously acquired information label position data, the palletizing posture of the box-shaped gripper robot 2 is adjusted. This adjustment ensures that the information label faces the visible surface after palletizing.

[0066] By integrating visual inspection, infrared sensing, and robotic arm position adjustment into a closed-loop control system, intelligent adjustments are achieved throughout the entire pharmaceutical product handling process, from gripping to palletizing. This approach overcomes the limitations of traditional fixed procedures, which are unable to cope with product posture changes, and improves the visibility and traceability of information labels during the palletizing process.

[0067] Specifically, the box-shaped gripping manipulator 2 can be driven by an industrial robot and can freely move and rotate in three-dimensional space. An infrared sensor array is installed within the box-shaped gripping manipulator 2 to scan the surface label information of the grasped object. By scanning and identifying the surface label information of the pharmaceutical product box, the position data of the pharmaceutical product box surface label information is obtained.

[0068] A high-speed industrial camera is installed at the first inspection position on the conveyor line as visual inspection mechanism 1. The camera continuously captures passing pharmaceutical products and transmits the image data in real time to an image processing unit. The image processing unit analyzes the conveyor image data and extracts the contour features of the pharmaceutical products. By comparing it with a pre-stored standard contour template, the current product's posture deviation is calculated.

[0069] Based on the posture deviation data, the control system generates posture adjustment instructions for the robot. The robot performs the adjustment so that its opening direction matches the actual posture of the pharmaceutical product.

[0070] Specifically, when a pharmaceutical product enters a predetermined gripping area, the gripping action is triggered. The robot moves above the product, aligns its opening with the product, and descends to complete the wrapping process, securing the product through vacuum suction or mechanical clamping. In this embodiment, vacuum suction is used to hold the top surface of the pharmaceutical product, enabling rapid gripping and transfer.

[0071] Furthermore, after the box-shaped grabbing robot 2 grabs the pharmaceutical product, it starts scanning with an infrared sensor array, identifies the position of the information label based on sensor detection, and transmits the label position data to a control system.

[0072] The control system obtains the preset palletizing position data and combines it with the label position information to calculate the optimal palletizing posture. This posture should ensure that the information label faces the visible surface after palletizing.

[0073] The robot arm places the pharmaceutical products in the designated palletizing position according to the calculated optimal posture, completing an intelligent palletizing operation.

[0074] The system executes the above steps in a loop, continuously processing pharmaceutical products on the conveyor line and realizing continuous intelligent palletizing operations.

[0075] Through the above-mentioned solution, this application realizes intelligent control of the palletizing process of pharmaceutical products. The coordinated application of visual inspection and infrared sensing enables the system to adapt to changes in product posture in real time and dynamically adjust the grasping and palletizing posture. This method effectively solves the problem that traditional fixed programs cannot cope with random product offsets, and improves palletizing accuracy and efficiency. In particular, by sensing the position of information labels and optimizing the palletizing posture accordingly, the visibility of the labels after palletizing is ensured, facilitating subsequent information detection and product traceability. This solution not only improves the automation level of the production line, but also enhances the standardization and traceability of pharmaceutical packaging, providing strong support for pharmaceutical production quality control.

[0076] Specifically, the visual inspection mechanism 1 is installed on the side or above the conveyor line through a fixed bracket, and its field of view covers the entire detection area of ​​the conveyor line. When the drug enters the first detection position, a trigger signal starts the visual inspection mechanism 1 to collect images. During the dynamic acquisition process, the visual inspection mechanism 1 continuously tracks the movement trajectory of the drug, and automatically adjusts the exposure time and focal length for each shot to compensate for the motion blur caused by the conveying speed. For example, when the conveying speed is 0.5 m / s, the exposure time is set to 1 / 1000 second. The collected image data is transmitted to the image processing unit after real-time compression processing, and the timestamp and position coordinates are recorded at the same time. This setting method can ensure that the collected images have the same scale and clarity at different conveying speeds, providing standardized input data for subsequent posture analysis. Through the coordination of hardware positioning and dynamic acquisition, the temporal resolution and spatial positioning accuracy of the image data are simultaneously improved, effectively eliminating image distortion problems caused by position offset or improper acquisition timing.

[0077] Specifically, extracting the posture contour image of the pharmaceutical product from the conveying image data and adjusting the gripping posture of the box-shaped gripping manipulator 2 according to the posture contour image includes:

[0078] Based on the Sobel operator, the contour of the pharmaceutical product is extracted from the conveying image data to obtain the posture contour image of the pharmaceutical product; based on the dynamic matching of the posture contour image of the pharmaceutical product with the pre-stored standard contour image, the posture rotation variable between the pharmaceutical product and the standard contour image is obtained; and the grasping posture of the box-shaped grasping robot 2 is adjusted according to the posture rotation variable.

[0079] During contour extraction, the Sobel operator constructs 3×3 convolution kernels in the horizontal and vertical directions to calculate the pixel gradient components. Gradient amplitude calculation can use square root superposition or absolute value summation, for example, setting the gradient amplitude threshold to 100-150 grayscale levels to suppress noise interference. During dynamic matching, the standard contour image layer is decomposed into multiple feature point sets. Each point set is rotated around the center with a preset step size. When the overlap between the actual contour and the standard contour reaches above 95%, the rotation angle is recorded as the pose rotation variable. During gripping posture adjustment, the robot drive shaft receives the rotation variable signal and controls the rotation angle via a servo motor, for example, with a rotation accuracy within ±0.5°. Because the conveyor image data acquired by the visual inspection mechanism 1 and the label position detection by the infrared sensor form a dual verification mechanism, the pose rotation variable obtained from contour matching can simultaneously correct for label orientation offsets, thereby ensuring consistent positioning of the pallet display surface.

[0080] Specifically, the conveyor image captured by the visual inspection mechanism 1 is first grayscaled and preprocessed. The Sobel operator is then used to calculate gradient components in the x and y directions. A dual threshold is set to filter out noise points, preserving the edge features of the drug packaging box and generating a pose contour image with clear boundaries. The extracted contour is then overlaid and compared with a pre-stored standard template. The standard template layer is rotated in steps of 0.1°, and the overlap ratio is calculated using an image similarity algorithm. When the maximum overlap is detected, the standard template rotation angle at that point is recorded as the pose correction. The robot control system receives this correction and, via a reduction mechanism, drives the rotating platform to rotate the corresponding angle, ensuring that the opening direction of the gripping jaw is fully aligned with the actual pose of the drug. During this process, the edge enhancement properties of the Sobel operator effectively eliminate artifacts caused by uneven illumination. The dynamic matching mechanism, by quantifying the rotation offset, improves the robot's pose adjustment accuracy to sub-angular levels, increasing the success rate of grasping complex-contour drugs from 82% to over 98%.

[0081] During the pharmaceutical product palletizing control process, the Sobel operator is first used to extract the pharmaceutical product contours from the conveyed image data. Specifically, the Sobel operator is used to calculate the horizontal and vertical gradients of the conveyed image. These gradients are then combined to form a final edge intensity image. Furthermore, the edge intensity image is thresholded to produce a binary pharmaceutical product pose contour image.

[0082] After acquiring a pose profile image of a pharmaceutical product, dynamically match it to a pre-stored standard profile image. The pose profile image is set as a fixed layer, while the standard profile image is set as a rotatable layer. The standard profile image is rotated incrementally until the two images achieve optimal matching. For example, the standard profile image can be rotated clockwise in 0.1-degree increments, and the degree of matching between the two images is calculated after each rotation. When the degree of matching reaches its maximum, the rotation angle at that point is recorded as the pose rotation variable.

[0083] Based on the obtained posture rotation variables, the gripping posture of the box-shaped gripping robot 2 is adjusted. Specifically, the box-shaped gripping robot 2 is first adjusted to an initial posture corresponding to the standard contour image. Then, a rotation adjustment instruction is generated based on the posture rotation variables, and the box-shaped gripping robot 2 is driven to rotate according to this instruction, so that its gripping posture matches the actual posture of the pharmaceutical product.

[0084] Specifically, the pose contour image and the standard contour image are set as superimposed layers for comparison, the layer where the pose contour image is located is fixed, and the layer where the standard contour image is located is set to a rotatable state; the standard contour image is rotated clockwise at a preset unit angle, and when the pose contour image overlaps with the standard contour image, the rotation angle of the standard contour image is obtained as the pose rotation variable.

[0085] The overlay layer comparison achieves independent operation by separating the image processing space. The layer containing the fixed-pose contour image can retain the original state of the actual collected data, while the layer containing the rotated standard contour image forms a controllable comparison benchmark. The preset unit angle can be set to 0.5 degrees to 2 degrees, for example, 1 degree is used as the incremental step size for fine matching. This can not only avoid matching errors caused by excessive angle steps, but also reduce excessive consumption of computing resources. The clockwise rotation direction limit forms a one-way matching rule, eliminating the logical conflicts caused by bidirectional rotation, while also ensuring operational consistency with the subsequent clockwise rotation of the robot.

[0086] Specifically, during the dynamic matching process, the extracted pose contour image is first used as the bottom fixed layer, and the pre-stored standard contour image is used as the top rotatable layer. The two layers are then overlaid with transparency using image processing software to form a visual comparison interface. The layer containing the standard contour image is rotated clockwise in increments of a preset unit angle. After each rotation, the pixel coincidence algorithm is used to detect the overlap between the two layers. A match is considered successful when the overlap area reaches a preset threshold. The cumulative rotation angle of the standard contour image is recorded as the pose rotation variable. This rotation angle directly corresponds to the required rotation adjustment for the robot arm. By converting this angle data into drive commands, closed-loop control of the grasping pose is achieved. During this process, the preset unit angle serves as a key parameter to control matching accuracy. For example, when using a 1-degree step size, the maximum angle deviation does not exceed 0.5 degrees, thereby ensuring the accuracy of pose adjustment.

[0087] Furthermore, the calculation of the position rotation variables during the dynamic matching process is more accurate, enabling more precise adjustment of the gripping position of the box-shaped gripping robot 2. This precise position adjustment ensures that the robot can accurately grasp pharmaceutical products in various postures, improving the stability and efficiency of the palletizing process. Furthermore, this method is highly adaptable and can handle pharmaceutical products of various shapes and postures, enhancing the versatility and flexibility of the palletizing system.

[0088] By precisely matching the pose profile image with the standard profile image and setting fixed and rotation layers, the image matching process is simplified. A clockwise rotation method with a preset unit angle ensures the accuracy and efficiency of the matching process. The rotation angle of the standard profile image is directly used as the pose rotation variable, providing an accurate numerical basis for subsequent robot pose adjustments. This method avoids complex image processing algorithms, reduces computational complexity, and improves the speed and accuracy of pose recognition. Furthermore, the use of a unified clockwise rotation direction simplifies the matching logic and reduces potential directional misjudgment. This precise pose recognition method provides a reliable foundation for subsequent robot gripping pose adjustments, helping to improve the gripping accuracy and efficiency of pharmaceutical product palletizing.

[0089] Specifically, adjusting the gripping posture of the box-shaped gripping manipulator 2 according to the posture rotation variable includes:

[0090] Adjusting the initial posture of the box-shaped grasping manipulator 2 according to the standard contour image;

[0091] Based on the posture state variable, a rotation adjustment instruction is generated, and based on the rotation adjustment instruction, the box-shaped grasping manipulator 2 is driven to rotate the rotation angle in the clockwise direction to form the grasping posture of the box-shaped grasping manipulator 2. The adjustment of the initial posture can be achieved through the spatial mapping of the manipulator base coordinate system and the standard contour image coordinate system, for example, the pixel coordinates of the four corner points in the image coordinate system are converted into the physical coordinates of the manipulator base. The generation of the rotation adjustment instruction can be performed according to the preset rotation step parameters. For example, a pulse control signal is generated every 0.5 degrees. The clockwise rotation path is set to a single rotation mode to avoid the cumulative error caused by bidirectional rotation. The encoder at the joint end of the manipulator can provide real-time feedback of the actual rotation angle, and the compensation mechanism is triggered when the difference with the rotation angle exceeds the threshold.

[0092] Specifically, a standard contour image is loaded as a reference coordinate system, and the gripper's gripping end plane is spatially aligned with the standard contour plane through visual servoing control. During the dynamic matching process between the pose contour image and the standard contour image, pose rotation variables are obtained through layer overlay comparison detection. During the initial pose calibration phase, the gripper's central axis is forcibly aligned with the centerline of the standard contour, and its plane tilt angle is measured and corrected using a three-axis gyroscope. After completing the initial calibration, rotation adjustment commands are decomposed into incremental rotation commands distributed at preset time intervals, for example, a 0.1-degree rotation command is sent every 10 milliseconds. After receiving the continuous rotation commands, the gripper's drive system transmits the rotational motion to the gripper end through a reduction gear train, while a position sensor continuously monitors the rotation angle deviation. When the actual rotation reaches the target rotation angle, the gripper's gripper plane forms the predetermined gripping angle with the pharmaceutical product's pose contour. At this point, the positioning error of the gripper's gripping posture can be controlled within a ±0.3-degree range. This phased adjustment mechanism effectively avoids the impact of compound errors on gripping accuracy by first eliminating initial pose deviations and then performing precise rotation operations.

[0093] Based on the rotation adjustment command, the box-shaped grasping robot 2 is driven to rotate clockwise by an angle, establishing its grasping posture. The control system then sends a command to the robot's rotation drive mechanism, causing the robot to rotate clockwise around its central axis. During the rotation, an encoder provides real-time feedback on the rotation angle, ensuring accurate target positioning. Upon completion of the rotation, the robot is in the optimal posture for grasping the current pharmaceutical product.

[0094] Specifically, the grabbing operation on the pharmaceutical product entering the grabbing area, detecting the position of the pharmaceutical product information label based on the infrared sensor, and extracting the pharmaceutical product information label position data include: driving the box-shaped grabbing robot 2 to move to the position above the pharmaceutical product, covering and adsorbing the pharmaceutical product based on the box-shaped grabbing robot 2; performing information label identification scanning on the pharmaceutical product located inside the box-shaped grabbing robot 2 based on the infrared sensor; and marking the label information position data of the pharmaceutical product based on the scanning result.

[0095] The steps of covering and adsorbing the pharmaceutical product by the box-shaped gripper 2 are achieved through a vacuum adsorption device, with an adsorption pressure range set to -80kPa to -50kPa to ensure the stability of the product. An infrared sensor scans the interior of the gripper laterally in a linear array scanning mode, with a scanning frequency set to 200 times per second. The label information position data is marked in a polar coordinate system, with the center of the product as the origin, and the angular deviation accuracy is controlled within a range of ±1°. The path planning that drives the box-shaped gripper 2's movement combines the posture profile data obtained by visual inspection to form a preliminary positioning, followed by secondary calibration through infrared scanning, forming a positioning error compensation mechanism.

[0096] Specifically, when the box-shaped grasping robot 2 covers the medicine to form a closed space, the infrared sensor scans the surface of the medicine without interference from ambient light. Since the information label is made of infrared reflective material, the sensor can identify areas with reflection intensity higher than the threshold during the scanning process. For example, the label area with a reflectivity of more than 60% is determined to be a valid position. After coordinate conversion, the scanning result generates offset data relative to the manipulator's clamping center, which is transmitted to the palletizing posture adjustment module. During the palletizing stage, the manipulator dynamically rotates the medicine according to the label position so that the label display surface remains parallel to the preset palletizing display surface, and the angle matching error is controlled within 2°. Through this technical solution, the positioning accuracy of the information label is improved by more than 30% compared to simple visual inspection, and the complete exposure rate of the label after palletizing can reach 98%.

[0097] Furthermore, based on the feedback position of the information label obtained by the infrared sensor, the setting orientation of the information label on the packaging box of the pharmaceutical product is determined according to the pharmaceutical packaging specifications of the pharmaceutical product production line, and the detection feedback value of the information label is obtained by the infrared sensor, thereby determining the position information of the information label on the pharmaceutical product.

[0098] When the information label on the packaging box of the pharmaceutical product is set on the circumferential side wall of the pharmaceutical product box, the setting relationship of the information label is determined based on the detection feedback value of the infrared sensor as shown in the following table:

[0099]

[0100] When the information label on the packaging box of the pharmaceutical product is set on the top and back of the pharmaceutical product box, the position information of the information label is determined based on the detection feedback value of the infrared sensor as shown in the following table:

[0101]

[0102] Among them, "T" represents the feedback value based on the infrared sensor detecting the information tag, and "F" represents the feedback value when the infrared sensor does not detect the information tag.

[0103] Pharmaceutical products that enter the grasping area are grasped. The infrared sensor detects the location of the pharmaceutical product information label and extracts the label location data. First, the box-shaped grasping robot 2 is driven to a position above the pharmaceutical product. The box-shaped grasping robot 2 can be rectangular in shape, with a suction device at the bottom. The robot moves directly above the pharmaceutical product, ensuring complete coverage. Then, the suction device is activated to secure the pharmaceutical product within the box-shaped grasping robot 2.

[0104] Next, the infrared sensor inside the box-shaped gripping robot 2 is activated to scan the information label on the pharmaceutical product inside. Multiple infrared sensors in fixed positions simultaneously scan the product to detect information labels on different sides of the pharmaceutical product box. During the scanning process, infrared light acts on the information label, reflecting off the special coating on the label to obtain the information content.

[0105] The position of the information label on the box of the pharmaceutical product and the recorded information content are detected by the infrared sensor, so as to adjust the position status of the pharmaceutical product according to the position of the information label of the pharmaceutical product and the stacking requirements.

[0106] In this regard, the present application further proposes to identify the information labels of pharmaceutical products based on infrared sensor scanning, and read the information label data of pharmaceutical products based on scanning.

[0107] Among them, the infrared sensor can be set as a pulse scanning probe with a wavelength range of 850-950nm, and its scanning angle can cover 80%-95% of the internal space of the box-shaped grasping robot 2. After the robot absorbs the pharmaceutical product, the infrared sensor performs a grid scan on the surface of the pharmaceutical by emitting an infrared beam of a specific frequency, and the scanning frequency can be controlled within the range of 50-200Hz. When the information label area is scanned, the reflective coating on the label surface and the infrared beam produce an optical feature difference, triggering the sensor to switch to data reading mode. At this time, the sensor parses the drug code, production batch and expiration date data stored in the label through a modem, and the parsing time can be controlled within the range of 50-200ms. The read data is transmitted to the control center via the industrial bus, and is time-stamped and synchronized with the posture data obtained by the visual inspection mechanism 1 to generate a composite data packet containing physical coordinates and product information.

[0108] By using infrared sensors to detect and identify the location and content of pharmaceutical product box information labels, the system achieves simultaneous physical positioning and data collection of pharmaceutical information labels, resolving the existing issue of palletizing verification failure caused by missing label information. Parallel processing of infrared scanning and data analysis prevents unilateral omissions of label location or data during robotic operation, ensuring accurate label placement and traceability of data for each palletized product, effectively improving the detection system's fault tolerance and information traceability integrity.

[0109] Specifically, adjusting the stacking posture of the box-shaped grasping robot 2 in combination with the information label position data includes: obtaining the stacking position data of the current pharmaceutical product according to the stacking process steps, and extracting the stacking space status and stacking display surface of the pharmaceutical product according to the stacking position data; adjusting the stacking posture of the pharmaceutical product of the box-shaped grasping robot 2 according to the stacking space status, and adjusting the position of the information label of the pharmaceutical product to be located on the stacking display surface.

[0110] The acquisition of the palletizing spatial state is achieved by analyzing the number of stacking layers, arrangement spacing, and adjacent drug contact surface parameters. For example, when the current stacking layer is detected to be the second layer, the horizontal staggered arrangement rule is extracted as the spatial state data. The palletizing display surface is set to the side facing the detection equipment according to the production line layout. Specifically, it can be configured to be the front or left side of the stack. When the robot adjusts the palletizing posture of the pharmaceutical product based on the robot, the information label is placed on the preset display surface by rotating the robot around the vertical axis by a specified angle. At the same time, the robot's descent height is controlled according to the stacking spacing data. For example, in the horizontal staggered arrangement state, the descent height is set to 0.8 times the height of the drug to maintain structural stability.

[0111] The pharmaceutical products are placed in the palletizing pile based on the set palletizing posture, and the pharmaceutical products are released through the vacuum breaking operation to realize the palletizing operation of the pharmaceutical products.

[0112] Specifically, when executing the fifth-layer palletizing process, the palletizing position data indicates that this layer needs to be aligned longitudinally. At this time, it is first analyzed that the longitudinal gap between adjacent medicines must be maintained at 5 mm, and the display surface is determined to be the right side of the stack. After the robot grabs the medicine, it sets the descent height to an integer multiple of the medicine height according to the longitudinal alignment rules, and rotates the medicine 90 degrees clockwise through the rotation mechanism, so that the information label originally located on the top surface is turned to the right side. In a single palletizing action, physical positioning and information positioning are completed simultaneously, which not only meets the stability requirements of the stacking structure, but also ensures that all medicine labels are exposed in the specified position, providing a standardized detection interface for the subsequent plastic sealing process.

[0113] Furthermore, this embodiment proposes adjusting the stacking posture of the box-shaped grasping robot 2 according to the information label position data to achieve pharmaceutical product positioning. However, in this process, it cannot be guaranteed that the information label of the pharmaceutical product is on the display surface during stacking, resulting in inconvenience in the subsequent information detection and traceability operations of the plastic-sealed packaged products.

[0114] Specifically, during the drug palletizing process, after the box-shaped grabbing robot 2 completes grabbing, the palletizing position data is analyzed in real time by the production line control module. The palletizing space state is determined by analyzing the stacking direction of the current stacked drug and the position offset of the adjacent drug. The palletizing display surface is determined as a vertical surface facing outward according to the preset film packaging process requirements. Subsequently, the robot's end effector adjusts the horizontal displacement and tilt angle based on the palletizing space state to ensure that the contact surface of the drug and the stacked layer below are completely in contact. At the same time, combined with the information label position data, the servo motor drives the robot to rotate around the vertical axis, so that the side of the drug with the information label is parallel and aligned with the palletizing display surface, and finally the drug is placed in the designated position.

[0115] This embodiment of the present invention simultaneously achieves physical positioning of drugs and directional label display during the palletizing process, avoiding the problem of label misalignment that prevents the plastic-sealed product from being visible in traditional methods. By dynamically analyzing the palletizing space and the preset display surface, it ensures both drug stacking stability and label visibility, reducing the need for manual label alignment correction during subsequent inspections and improving production line automation.

[0116] Example 3:

[0117] Figure 4 A schematic diagram of an intelligent pharmaceutical product palletizing control system according to an embodiment of the present invention is shown. The control system is used to execute the intelligent pharmaceutical product palletizing control method. The control system includes:

[0118] Visual recognition component 10: used to obtain the posture status data of the pharmaceutical product at the first detection position. Specifically, a visual detection mechanism is set on the pharmaceutical product conveyor line, so that the visual detection mechanism can perform real-time detection of the first detection position on the pharmaceutical product conveyor line, so as to obtain the conveying image data of the pharmaceutical product passing through the first detection position.

[0119] Furthermore, the colors of the box body of the pharmaceutical product and the conveyor belt of the conveyor line can be set to high-contrast colors, so as to extract the outline of the pharmaceutical product from the conveying image data.

[0120] The first posture adjustment component 20 is used to adjust the gripping posture of the box-shaped gripping robot according to the posture state data; extract the contour of the pharmaceutical product from the conveying image data based on the Sobel operator to obtain the posture contour image of the pharmaceutical product; dynamically match the posture contour image of the pharmaceutical product with the pre-stored standard contour image to obtain the posture rotation variable between the pharmaceutical product and the standard contour image; adjust the gripping posture of the box-shaped gripping robot according to the posture rotation variable.

[0121] Information processing component 30: used to detect and obtain the position data of the information label of the pharmaceutical product. The information label of the pharmaceutical product is detected by the internal infrared sensor of the box-shaped grasping robot. The infrared scanning module is used in conjunction with the reflective label recognition technology to detect the physical position of the information label of the pharmaceutical product and ensure that the label is in the visible display surface after stacking.

[0122] The second posture adjustment component 40 is used to adjust the stacking posture of the box-shaped grasping robot according to the information tag position data.

[0123] According to the palletizing process steps, the palletizing position data of the current pharmaceutical product is obtained, and the palletizing space status and palletizing display surface of the pharmaceutical product are extracted based on the palletizing position data; the palletizing posture of the pharmaceutical product of the box-shaped grasping robot is adjusted according to the palletizing space status, and the position of the information label of the pharmaceutical product is adjusted to the palletizing display surface.

[0124] An embodiment of the present invention provides an intelligent pharmaceutical product palletizing control method system, which obtains the conveying posture state of the pharmaceutical product through image processing, and adjusts the palletizing posture of the pharmaceutical product when the robot arm grasps the pharmaceutical product in combination with the palletizing requirements of the pharmaceutical product, so that the box body information label of the pharmaceutical product can face the display surface of the palletizing pile. It has the advantages of improving the posture adjustment accuracy and palletizing efficiency of the pharmaceutical product, ensuring that the information label is accurately exposed, thereby improving product information management and traceability efficiency.

[0125] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0126] In addition, the embodiments of the present invention are introduced in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. An intelligent pharmaceutical product palletizing control method, characterized in that: The control method is applicable to a pharmaceutical product palletizing production line, which comprises: a box-shaped grabbing manipulator with a single-side opening and an infrared sensor arranged inside the box-shaped grabbing manipulator; The box-shaped structure of the single-side-open box-shaped grasping manipulator can cover the three outer side surfaces of the pharmaceutical product and the top surface of the pharmaceutical product. The infrared sensor inside the box-shaped grasping manipulator obtains the information label status of the three outer side surfaces of the pharmaceutical product and the top surface of the pharmaceutical product. The infrared sensor obtains the position of the label information on the surface of the pharmaceutical product box body to adjust the posture state of the pharmaceutical product box body. The control method includes: Acquiring conveying image data of the pharmaceutical product at a first inspection position based on a visual inspection mechanism; extracting a posture contour image of the pharmaceutical product from the conveying image data, and adjusting the gripping posture of the box-shaped gripping manipulator according to the posture contour image; Grab the pharmaceutical products that enter the grabbing area, detect the location of the pharmaceutical product information label based on the infrared sensor, and extract the pharmaceutical product information label location data; Obtaining the palletizing position data of the pharmaceutical product, and adjusting the palletizing posture of the box-shaped grasping manipulator in combination with the information tag position data; The grabbing operation of the pharmaceutical product entering the grabbing area, detecting the position of the pharmaceutical product information label based on the infrared sensor, and extracting the pharmaceutical product information label position data includes: Driving the box-shaped grabbing robot to move to a position above the pharmaceutical product, and covering and adsorbing the pharmaceutical product with the box-shaped grabbing robot; Performing information label recognition and scanning on the pharmaceutical product inside the box-shaped grabbing robot based on an infrared sensor; Marking the label information location data of the pharmaceutical product based on the scanning result, including: obtaining the pharmaceutical product specifications of the pharmaceutical product conveyor line, and obtaining the label setting orientation data of the pharmaceutical product; Based on the scanning result of the infrared sensor of the box-shaped grasping manipulator, the label information position data is generated in combination with the label setting orientation data; The obtaining of the palletizing position data of the pharmaceutical product and adjusting the palletizing posture of the box-shaped grabbing manipulator in combination with the information tag position data includes: Obtaining the palletizing position data of the current pharmaceutical product according to the palletizing process steps, and extracting the palletizing space state and palletizing display surface of the pharmaceutical product according to the palletizing position data; The stacking posture of the pharmaceutical product of the box-shaped grasping robot is adjusted according to the state of the stacking space, and the position of the information label of the pharmaceutical product is adjusted to be located on the stacking display surface.

2. The intelligent pharmaceutical product palletizing control method according to claim 1, characterized in that: The method of obtaining the transport image data of the pharmaceutical product at the first detection position based on the visual detection mechanism includes: Setting up a visual inspection mechanism at the first inspection position of the pharmaceutical product conveying line; Based on the visual inspection mechanism, transport image data of the pharmaceutical product passing through the first inspection position is obtained.

3. The intelligent pharmaceutical product palletizing control method according to claim 1, characterized in that: The extracting of the posture contour image of the pharmaceutical product from the conveying image data and adjusting the grasping posture of the box-shaped grasping manipulator according to the posture contour image comprises: Extracting the contour of the pharmaceutical product from the transport image data based on the Sobel operator to obtain a posture contour image of the pharmaceutical product; Dynamically matching the pose contour image of the pharmaceutical product with a pre-stored standard contour image to obtain the pose rotation variables between the pharmaceutical product and the standard contour image; The gripping posture of the box-shaped gripping manipulator is adjusted according to the posture rotation variable.

4. The intelligent pharmaceutical product palletizing control method according to claim 3, characterized in that: The dynamically matching the posture contour image of the pharmaceutical product with the pre-stored standard contour image to obtain the posture rotation variables between the pharmaceutical product and the standard contour image includes: The pose contour image and the standard contour image are set as superimposed layers for comparison, the layer where the pose contour image is located is set to a fixed state, and the layer where the standard contour image is located is set to a rotating state; The standard contour image is rotated clockwise at a preset unit angle. When the posture contour image overlaps with the standard contour image, the rotation angle of the standard contour image is obtained, and the rotation angle of the standard contour image is set as a posture rotation variable.

5. The intelligent pharmaceutical product palletizing control method according to claim 4, characterized in that: The adjusting the gripping posture of the box-shaped gripping manipulator according to the posture rotation variable comprises: adjusting the initial posture of the box-shaped grasping manipulator according to the standard contour image; A rotation adjustment instruction is generated based on the posture rotation variable, and the box-shaped grasping manipulator is driven to rotate the rotation angle in a clockwise direction based on the rotation adjustment instruction to form a grasping posture of the box-shaped grasping manipulator.

6. The intelligent pharmaceutical product palletizing control method according to claim 1, characterized in that: The grabbing operation of the pharmaceutical product entering the grabbing area, detecting the position of the pharmaceutical product information label based on the infrared sensor, and extracting the pharmaceutical product information label position data further includes: The information label of the pharmaceutical product is identified based on scanning by the infrared sensor, and the information label data of the pharmaceutical product is read based on the scanning.

7. An intelligent pharmaceutical product palletizing control system, characterized in that: The control system is used to execute the intelligent pharmaceutical product palletizing control method according to any one of claims 1 to 6, and the control system includes: Visual recognition component: used to obtain the posture status data of the pharmaceutical product at the first detection position; A first posture adjustment component: used for adjusting the gripping posture of the box-shaped gripping manipulator according to the posture state data; Information processing component: used to detect and obtain the location data of the information label of the pharmaceutical product; when the box-shaped grasping robot covers the pharmaceutical product to form a closed space, the infrared sensor scans the surface of the pharmaceutical product without interference from ambient light. The scanning result is converted into offset data relative to the gripping center of the robot after coordinate conversion; The second posture adjustment component is used to adjust the stacking posture of the box-shaped grasping robot according to the position data of the information tag.

Images