Manufacturing apparatus for marked articles, manufacturing method, manufacturing program, marked articles, and recognition program for marked articles.

The apparatus generates markers on articles to identify gripping positions, addressing the inefficiencies in gripping and transporting without database access by encoding necessary information directly on the articles.

JP2026182070APending Publication Date: 2026-11-19IHI CORP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2025077991
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2025-05-08
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Existing systems require database access to acquire cargo form information, which is not feasible in environments without network connectivity, leading to inefficiencies in gripping and transporting articles.

Method used

A manufacturing apparatus and method that uses a controller to generate markers on articles based on dimensions and installation positions, enabling identification of gripping positions without database access, utilizing geometric figures and codes to encode relevant information.

Benefits of technology

Enables efficient gripping and transporting of articles by allowing the acquisition of cargo form information and gripping positions in environments where database access is limited.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026182070000001_ABST
    Figure 2026182070000001_ABST
Patent Text Reader

Abstract

In environments where databases cannot be easily accessed, this invention provides a manufacturing apparatus, manufacturing method, manufacturing program, marked articles, and a recognition program for marked articles that can obtain information on the packaging of articles and improve the efficiency of handling and transporting those articles. [Solution] The manufacturing apparatus, manufacturing method, and manufacturing program for marker-equipped articles use a controller connected to an input unit and an installation unit. The controller acquires article information, including at least information regarding the dimensions of the article, via the input unit, generates markers based on the article information and the installation positions of the markers on the surface of the article, outputs the markers and installation positions to the installation unit, and installs the markers at the installation positions via the installation unit. Based on the installation positions and markers, the gripping position for gripping the article can be determined.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to an apparatus for manufacturing a marker-equipped article, a manufacturing method, a manufacturing program, a marker-equipped article, and a recognition program for a marker-equipped article.

Background Art

[0002] Patent Document 1 discloses a technique of acquiring box body information stored in a database based on a two-dimensional code on a form affixed to a box body stacked on a pallet, and generating coordinate data of each part of the box body as cargo form information of the box body stacked on the pallet based on the box body information.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] According to the technique described in Patent Document 1, it is necessary to refer to a database to acquire box body information. Therefore, in an environment where the database cannot be easily referenced, such as an environment not connected to an external network, there is a problem that the cargo form information of the article cannot be obtained, and the efficiency of gripping, transporting, and the like of the article cannot be improved.

[0005] The present disclosure has been made in view of the above problems. An object of the present disclosure is to provide an apparatus for manufacturing a marker-equipped article, a manufacturing method, a manufacturing program, a marker-equipped article, and a recognition program for a marker-equipped article, which are capable of obtaining cargo form information of an article and improving the efficiency of gripping, transporting, and the like of the article in an environment where a database cannot be easily referenced.

Means for Solving the Problems

[0006] The manufacturing apparatus, manufacturing method, and manufacturing program for marked articles relating to this disclosure use a controller connected to an input unit and an installation unit. The controller acquires article information, including at least information regarding the dimensions of the article, via the input unit, generates a marker based on the article information and the installation position of the marker on the surface of the article, outputs the marker and installation position to the installation unit, and installs the marker at the installation position via the installation unit. Based on the installation position and the marker, a gripping position for gripping the article can be identified.

[0007] The marker may include a geometric figure that identifies a two-dimensional coordinate axis fixed to the surface on which the marker is installed.

[0008] The marker may include a one-dimensional code or a two-dimensional code.

[0009] The marker may also be used to identify the coordinates of the surface on which the marker is installed, based on the installation location.

[0010] The marker may have three or more finder patterns, and the arrangement of the finder patterns may determine the two-dimensional coordinate axes.

[0011] The installation location may be on the same plane as the surface where the gripping position appears.

[0012] Multiple different markers may be applied to one surface of the item.

[0013] The multiple distinct markers may include a first marker and a second marker, and the second marker may be applied on one side of the article, on the opposite side of the gripping center of the article from the first marker.

[0014] The marker may identify additional information about the item.

[0015] The additional information may include at least one of the following: the gripping position of the item, the weight of the item, the position of the item based on its geometric center of gravity, the position of the item based on its physical center of gravity, the method of gripping the item, the conditions for gripping the item, the permitted stacking method for the item, the number of stacking layers, and information specified by the user for the item.

[0016] In the marked article relating to this disclosure, the marker may hold article information, which includes at least information relating to the dimensions of the article, and encoded information based on the placement of the marker on the surface of the article, and may be able to identify a gripping position for gripping the article based on the placement and encoded information.

[0017] The recognition device, recognition method, and recognition program for marked articles relating to this disclosure use a controller connected to an input unit and an output unit. The controller acquires an image of an article marked with a marker via the input unit, extracts the marker and the placement of the marker on the surface of the article from the image, calculates a gripping position for gripping the article based on the marker and placement, and outputs the gripping position via the output unit.

[0018] The controller may identify a two-dimensional coordinate axis fixed to the surface on which the marker is installed, based on the shape of the marker.

[0019] The controller may calculate the coordinates of the surface on which the marker is installed, based on the marker and the installation location.

[0020] The controller may, when multiple markers are extracted from an image, calculate a gripping position for each marker, identify a group of markers relating to the same item based on the multiple gripping positions, and calculate a single gripping position by averaging the multiple gripping positions relating to the group of markers. [Effects of the Invention]

[0021] According to the present disclosure, it is possible to provide an apparatus for manufacturing a marked article, a manufacturing method, a manufacturing program, a marked article, and a recognition program for a marked article, which are capable of acquiring packaging form information of the article and improving the efficiency of gripping, transporting, etc. of the article in an environment where a database cannot be easily referenced. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] [Figure 1] It is a block diagram showing the configuration of the apparatus for manufacturing a marked article according to the present disclosure. [Figure 2] It is a flowchart showing the method for manufacturing a marked article according to the present disclosure. [Figure 3] It is a block diagram showing the configuration of the recognition apparatus for a marked article according to the present disclosure. [Figure 4] It is a flowchart showing the recognition method for a marked article according to the present disclosure. [Figure 5] It is a diagram showing an example of a marker. [Figure 6] It is a top view showing a first example of marker arrangement. [Figure 7] It is a top view showing a second example of marker arrangement. [Figure 8] It is a top view showing a third example of marker arrangement. [Figure 9] It is a top view showing a fourth example of marker arrangement. [Figure 10] It is a top view showing an example of a case where a plurality of articles are stacked adjacent to each other. DETAILED DESCRIPTION OF EMBODIMENTS FOR CARRYING OUT THE INVENTION

[0023] Hereinafter, several exemplary embodiments will be described with reference to the drawings. Note that common parts in each drawing are denoted by the same reference numerals, and overlapping descriptions are omitted.

[0024] [Configuration of Manufacturing Apparatus] Figure 1 is a block diagram showing the configuration of a manufacturing apparatus for marker-marked articles according to this disclosure. The manufacturing apparatus 20 comprises an input unit 21, a controller 22, and a processing unit 24.

[0025] The input unit 21 is an interface for acquiring item information, which includes at least information regarding the dimensions of the item. Specifically, it may be a manual input device such as a touch panel display, keyboard, or mouse.

[0026] Furthermore, the input unit 21 may have a network connection function to a company's core system, such as ERP (Enterprise Resource Planning) or WMS (Warehouse Management System), and may automatically acquire necessary information from a database of goods information.

[0027] Furthermore, the input unit 21 may be equipped with measuring instruments such as a 3D scanner, laser measuring instrument, or stereo camera for automatically measuring the dimensions of the item. In addition to dimensional information such as the width, height, and depth of the item, the item information may also include additional information such as shape classification (rectangular prism, cylinder, etc.), weight, physical and geometric center of gravity of the contents, pressure resistance, information specifying grippable points, stacking restrictions (maximum number of layers, directional restrictions, etc.), and hazardous material markings. This information may be pre-registered in a database.

[0028] "Article" refers to an object with width, depth, and height that is subject to transport or other means. For example, "article" may be goods, materials, or cardboard boxes, containers, individual product boxes, construction materials, bags (rice bags, cement bags, etc.), or other containers used to store goods, materials, etc. "Article" may also be something that can be reasonably approximated as a rectangular prism. "Article" is not limited to the examples given herein.

[0029] Controller 22 is a general-purpose computer equipped with a CPU (Central Processing Unit), memory, and input / output units. Controller 22 has a computer program (manufacturing program) installed that enables it to function as the manufacturing equipment 20. By executing the computer program, Controller 22 functions as one of the multiple information processing circuits (221, 222, 223, 224, 225) of the manufacturing equipment 20.

[0030] This disclosure provides an example of implementing multiple information processing circuits (221, 222, 223, 224, 225) using software. However, it is also possible to configure the information processing circuits (221, 222, 223, 224, 225) by preparing dedicated hardware for each of the information processing operations described below. Alternatively, the multiple information processing circuits (221, 222, 223, 224, 225) may be configured using separate hardware.

[0031] As shown in Figure 1, the controller 22 includes a plurality of information processing circuits (221, 222, 223, 224, 225), namely an information acquisition unit 221, a gripping position setting unit 222, a construction position setting unit 223, a marker generation unit 224, and an output generation unit 225.

[0032] The information acquisition unit 221 acquires item information via the input unit 21. Specifically, it performs tasks such as receiving manually entered data, acquiring information from a database, and receiving measurement data from sensors.

[0033] Furthermore, the information acquisition unit 221 may also include input data validity verification functions (such as checking the range of input values ​​and logical consistency) and functions to supplement missing data (such as applying default values ​​and estimating from past data). In addition, the information acquisition unit 221 may also be capable of performing data fusion processing to integrate data from different input sources, or filtering processing to remove noise. This makes it possible to acquire highly reliable product information.

[0034] The gripping position setting unit 222 sets an appropriate gripping position for the item based on the acquired item information. Possible algorithms for setting the gripping position include methods based on the physical center of gravity, methods based on the geometric center of gravity, methods based on stress distribution analysis, and methods for predicting the gripping position using machine learning. Various methods can be used to set the gripping position.

[0035] For example, the gripping position setting unit 222 may calculate a gripping position by considering information such as the shape of the article, weight distribution, and arrangement of contents. The gripping position setting unit 222 may set a gripping position to a symmetrical point on a horizontal plane passing through the center of gravity of a rectangular parallelepiped, or it may prioritize gripping from a direction perpendicular to the long side of the article. Here, "symmetrical point on a horizontal plane passing through the center of gravity" means a plurality of points arranged symmetrically on a horizontal plane passing through the center of gravity (physical center of gravity or geometric center), which is the center of the article. This makes it possible to grip the article more securely.

[0036] If the object has a special shape or is unbalanced, the gripping position setting unit 222 may determine the gripping position by referring to the results of a mechanical analysis using the finite element method or the like. Furthermore, the gripping position setting unit 222 may also calculate gripping force adjustment parameters that take into account the material or surface condition (slipperiness, etc.) of the object.

[0037] The result of setting the gripping position may be displayed as a visual preview and may be editable by the operator. The method for setting the gripping position is not limited to the examples given herein.

[0038] The installation position setting unit 223 sets the installation position of the marker on the item based on the set gripping position. The algorithm for selecting the marker installation position includes selecting a position that does not interfere with the gripping operation, selecting a position that maximizes the visibility of the marker, and determining the arrangement pattern of multiple markers. Various methods can be used to set the installation position.

[0039] For example, when multiple markers are placed at diagonal positions on either side of the gripping center, the installation position setting unit 223 may calculate the coordinates of each marker and determine the arrangement that maximizes gripping accuracy. Alternatively, the installation position setting unit 223 may select a position that maximizes marker readability (e.g., a flat area with low light reflection) based on the shape or surface characteristics of the article. Furthermore, the installation position setting unit 223 may determine an arrangement that is highly adaptable to the environment, taking into account the article's usage environment (lighting conditions, viewing angle, etc.). The installation position may be defined by two-dimensional coordinates set on the surface of the article (e.g., the top surface) and converted into control parameters for the installation unit 24.

[0040] The marker generation unit 224 generates markers based on the item information and the set construction position. For example, the markers may store information about the gripping position, as well as additional information such as the weight of the item, the center of gravity, the gripping method, and the stacking method, in addition to the item information.

[0041] As a marker generation algorithm, two-dimensional code generation algorithms such as QR Code (registered trademark) and DataMatrix (registered trademark), and other one-dimensional code generation algorithms such as barcodes may be used. The method of generating markers is not limited to the examples given here.

[0042] In the marker generation unit 224, data may be organized in an efficient structured format (JSON, XML, etc.), and compression and encryption processing may be applied as needed. The physical size of the generated markers may also be adjusted considering the imaging distance and the resolution of the imaging device. In addition, the marker generation unit 224 may perform contrast adjustment according to environmental conditions (type of light source, surface reflection characteristics, etc.) and add redundancy to improve durability. Furthermore, the marker generation unit 224 may embed cross-reference information between multiple markers to generate markers that allow information recovery even if some markers cannot be read.

[0043] Furthermore, the marker generation unit 224 may encode the coordinate information of the four corners of the surface of the article as article information, and may also encode additional information such as the height of the article, the load mass or maximum load mass, information on a gripper suitable for gripping the article (gripper type, usage conditions, required suction force, etc.), and the recommended stacking method and number of layers when stacking articles into the marker. When using a two-dimensional code such as a QR code, it is possible to compactly encode this information and include it in the marker.

[0044] The output generation unit 225 outputs the marker information generated by the marker generation unit 224 and the construction position information set by the construction position setting unit 223 to the construction unit 24.

[0045] The output data format may be automatically selected to be the most suitable format according to the type of construction unit 24 (raster image format, vector image format, printer control commands, robot motion commands, etc.). Furthermore, the construction position may be converted to the equipment's coordinate system, and calibration correction may be applied as needed. The output data may also include information for quality verification (checksum, version information, etc.).

[0046] In addition, the output generation unit 225 may ensure traceability by automatically recording logs of the output processing (date and time, setting values, generation results, etc.). The output generation unit 225 may also monitor the operating status of the construction unit 24 and start outputting when ready, and may also be equipped with a feedback receiving function for verifying the construction results.

[0047] The application unit 24 is an application device that applies markers to specified positions on an item based on the markers and application positions output from the output generation unit 225. Various methods can be used for application, such as printing, label application, engraving, and laser marking.

[0048] Specifically, the application unit 24 may use printing equipment such as high-resolution inkjet printers, thermal transfer printers, or UV ink printers; labelers equipped with automatic label peeling and application mechanisms; laser marking equipment using YAG lasers or fiber lasers; or CNC-controlled mechanical marking equipment. These devices may be used in combination with precision positioning mechanisms such as XYZ stages or articulated robot arms.

[0049] Alternatively, the marker may be printed simultaneously during the manufacturing of the corrugated cardboard box. Specifically, the application unit 24 may be a box surface printing machine used in the corrugated cardboard box manufacturing process. The box surface printing machine may print the marker at a predetermined position at the same time as the manufacturing of the corrugated cardboard box. This allows gripping position information to be incorporated from the manufacturing stage of the corrugated cardboard box, which is the product, and eliminates the need for post-printing. By utilizing existing equipment of a corrugated cardboard box surface printing machine, it is possible to implement the technology disclosed herein while keeping additional capital investment to a minimum. Of course, it is also possible to print the marker on a pre-made corrugated cardboard box afterwards, or to apply the marker by attaching a label.

[0050] The construction unit 24 may be equipped with a conveyor belt, palletizer, pick-and-place robot, etc., for transporting and positioning the goods. The construction unit 24 may also be equipped with verification equipment such as a high-resolution camera and barcode reader for checking the quality of the construction. The construction unit 24 may apply a protective treatment to the markers, such as a transparent coating, to improve the durability of the markers after construction. The materials used for construction (ink, label material, etc.) may be selected according to the usage environment of the goods (temperature, humidity, abrasion, chemical substances, etc.).

[0051] [Manufacturing Process Procedure] Figure 2 is a flowchart showing the method for manufacturing a marker-attached article according to this disclosure.

[0052] In step S101, the information acquisition unit 221 acquires article information, including at least information regarding the dimensions of the article, via the input unit 21. In addition to the article information, additional information may also be acquired.

[0053] In step S103, the gripping position setting unit 222 sets the gripping position based on the acquired item information. Note that the gripping position may be pre-set independently of the item information.

[0054] In step S105, the installation position setting unit 223 sets the installation position of the marker based on the set gripping position.

[0055] In step S107, the marker generation unit 224 generates a marker based on the item information and the installation location. The marker encodes the information necessary for gripping the item and additional information.

[0056] In step S109, the output generation unit 225 outputs the generated marker and the set construction position to the construction unit 24.

[0057] In step S111, the installation unit 24 installs the outputted marker at the specified installation location.

[0058] Through the above process, markers generated based on the item information and installation location are applied to the item, and a marker-equipped item is manufactured in which the gripping position can be identified from the markers and installation location.

[0059] [Configuration of the recognition device] Figure 3 is a block diagram showing the configuration of a marker-marked article recognition device according to this disclosure. The recognition device 40 comprises an input unit 41, a controller 42, and an output unit 43.

[0060] The input unit 41 is an interface for acquiring images of an object on which a marker has been applied. For example, the input unit 41 is connected to an imaging device such as a camera that images the object on which the marker has been applied. The camera may be a 2D camera or a 3D camera, and may be used in combination with additional sensors such as a height sensor as needed.

[0061] Specifically, usable 2D cameras include industrial monochrome / color cameras, high frame rate cameras, cameras with wide-angle / telephoto lenses, cameras with telecentric lenses, low-light cameras, and infrared cameras. 3D cameras include TOF (Time of Flight) cameras, structured optical 3D scanners, stereo vision cameras, and LiDAR (Light Detection and Ranging).

[0062] In addition to these, auxiliary sensors for acquiring distance / height information, such as laser rangefinders, ultrasonic distance sensors, and optical height sensors, may be combined. Camera placement options include overhead placement for shooting from above an object, side placement for shooting from the side, and multi-angle placement for simultaneous shooting from multiple directions. Control of the shooting environment may include lighting devices such as LED arrays for uniform illumination, directional lighting, diffuse light sources, backlights, and strobe lighting, as well as optical filters such as light-shielding hoods and polarizing filters.

[0063] The controller 42 is a general-purpose computer equipped with a CPU (Central Processing Unit), memory, and an input / output unit. The controller 42 has a computer program (recognition program) installed on it that functions as the recognition device 40. By executing the computer program, the controller 42 functions as one of the multiple information processing circuits (421, 422, 423, 424, 425, 426) of the recognition device 40.

[0064] This disclosure provides an example of implementing multiple information processing circuits (421, 422, 423, 424, 425, 426) using software. However, it is also possible to configure the information processing circuits (421, 422, 423, 424, 425, 426) by preparing dedicated hardware for each of the information processing operations described below. Alternatively, the multiple information processing circuits (421, 422, 423, 424, 425, 426) may be configured using separate hardware.

[0065] As shown in Figure 3, the controller 42 includes multiple information processing circuits (421, 422, 423, 424, 425, 426), namely an image acquisition unit 421, a marker extraction unit 422, a construction position extraction unit 423, an information extraction unit 424, a gripping position calculation unit 425, and an output generation unit 426.

[0066] The image acquisition unit 421 acquires an image of the item on which the marker has been applied via the input unit 41. For example, it may acquire image data from a camera and perform preprocessing.

[0067] The image acquisition unit 421 may automatically apply preprocessing such as image distortion correction (lens distortion correction), noise reduction (median filter, Gaussian filter, etc.), contrast enhancement (histogram equalization, adaptive binarization, etc.), and resolution adjustment (up / downsampling). When multiple cameras are used, the image acquisition unit 421 may perform synchronized acquisition control (hardware trigger, software trigger) and calibration between cameras (stereo calibration).

[0068] In addition, the image acquisition unit 421 may be equipped with an automatic exposure adjustment function to respond to changes in lighting conditions, or a high dynamic range (HDR) synthesis function. Furthermore, if a 3D camera is used, it may perform depth map or point cloud data acquisition and conversion processing. The image acquisition unit 421 may temporarily store image data in an image buffer and perform time-series analysis of multiple frames (inter-frame difference, motion analysis, etc.).

[0069] The marker extraction unit 422 extracts markers from the acquired image. As a marker detection algorithm, the marker extraction unit 422 may perform a series of processes including adaptive binarization, connected component analysis, finder pattern detection, pattern verification, and code region extraction.

[0070] For example, image processing is performed with multiple binarization thresholds to accommodate variations in lighting conditions. Next, image moment analysis or machine learning-based object detection algorithms (such as YOLO, SSD, or Faster R-CNN) may be applied to detect characteristic shapes (such as the square structure of a QR code finder pattern).

[0071] The marker extraction unit 422 may execute a tracking algorithm that identifies and tracks individual markers when multiple markers are present in the image. Furthermore, the marker extraction unit 422 may execute an algorithm capable of detecting partially hidden or tilted markers. Thus, the marker extraction unit 422 may be designed to accommodate various shooting conditions (such as variations in lighting, reflections, and shadows).

[0072] Furthermore, the marker extraction unit 422 may implement statistical methods (such as Random Sample Consensus) to improve the accuracy of marker detection, as well as high-speed processing utilizing a GPU (Graphics Processing Unit) or DSP (Digital Signal Processor). The detection results are output along with a confidence score, and verification processing may be added for detection results with low confidence scores.

[0073] The construction location extraction unit 423 extracts the construction location of the extracted marker. Specifically, it implements a process to estimate the construction location in three-dimensional space from the position of the detected marker on the image.

[0074] For example, the construction position extraction unit 423 may use camera calibration technology to convert image coordinates within the acquired image to world coordinates in the three-dimensional space where the object is located. A projection transformation model using the camera's intrinsic parameters (focal length, image center, distortion coefficient, etc.) and extrinsic parameters (position and orientation) enables highly accurate coordinate transformation.

[0075] The construction position extraction unit 423 may perform position estimation using a PnP (Perspective-n-Point) algorithm that utilizes known object sizes or reference markers when using only a 2D camera. The construction position extraction unit 423 may perform 3D position calculation using depth information directly when using a 3D camera.

[0076] Furthermore, the construction location extraction unit 423 may perform highly accurate location estimation using the triangulation method when data from multiple viewpoints is available. It may also perform uncertainty evaluation of the location estimation based on image resolution and camera accuracy, and output the results along with a confidence interval. The construction location extraction unit 423 may also employ estimation accuracy improvement techniques such as a Kalman filter utilizing temporal continuity.

[0077] The information extraction unit 424 extracts item information and additional information from the marker. If the marker is a one-dimensional code or a two-dimensional code, the information extraction unit 424 may perform a series of decoding processes such as detecting a mask pattern, constructing a sampling grid, reading the data bit sequence, and error correction processing. The information extraction unit 424 may also restore information even if part of the marker is damaged using an advanced error correction algorithm (such as a Reed-Solomon code).

[0078] The data extracted by the information extraction unit 424 may be parsed according to a defined structured format (JSON, XML, etc.) and broken down into various types of information such as item dimensions, weight, gripping position information, and stacking conditions. Subsequently, data integrity verification (checksum verification, version check, etc.) may be performed automatically, and appropriate error handling may be executed if corrupted or incompatible data is detected.

[0079] In addition, if the data extracted by the information extraction unit 424 includes encrypted data, appropriate decryption processing may be performed. If the information extraction unit 424 obtains the same type of information from multiple markers, it may perform integration and consistency verification processing of multiple data related to multiple markers.

[0080] The information extraction unit 424 may, if a database (not shown) is available, refer to the database to obtain additional information associated with the ID contained in the marker.

[0081] The gripping position calculation unit 425 calculates the gripping position of the item based on the marker and the construction position. Specifically, it calculates the gripping position of the item in the coordinate system by combining the gripping position information extracted from the marker with the two-dimensional coordinates of the construction position.

[0082] For example, the gripping position calculation unit 425 may perform a transformation process from the two-dimensional coordinates on the surface where the marker is applied to the coordinate system of the article. The gripping position calculation unit 425 may define coordinate axes from the arrangement of the finder pattern or alignment pattern on the marker and transform them to the coordinate system of the article. An affine transformation matrix or a projection transformation matrix (homography matrix) may be used for this transformation.

[0083] The gripping position calculation unit 425 may calculate the gripping direction (approach vector) and gripping posture (gripper rotation angle) in addition to the three-dimensional coordinates (X, Y, Z) of the gripping position based on information indicating the shape of the article. Furthermore, dynamic adjustments to the gripping operation may be made according to the condition of the article (tilt, stacking state, etc.).

[0084] Furthermore, the possibility of grasping may be evaluated based on the range of motion of the robot arm or the location of the obstacle. Then, the optimal grasping position is selected from multiple candidates. The accuracy of the calculation of the grasping position may be quantified based on the uncertainty evaluation of the measurement system.

[0085] The output generation unit 426 outputs the calculated gripping position to the output unit 43. Specifically, it outputs the 3D coordinates of the gripping position, the gripping posture, and the gripping parameters (gripping force, approach speed, etc.) in a format that can be interpreted by the robot controller.

[0086] The output data may be generated in a format compatible with standard industrial robot control languages ​​(such as RAPID, KRL, and URScript) and common robot communication protocols (such as Ethernet / IP, PROFINET, and CC-Link).

[0087] To ensure real-time performance, a low-latency communication protocol may be employed, and the data transmission and reception status may be constantly monitored. Additionally, the system may allow the operator to pre-confirm the gripping operation. Processing results may be automatically recorded as logs and used for traceability and post-analysis.

[0088] The output unit 43 is an interface that outputs information on the calculated gripping position. For example, the output unit 43 may be connected to a gripping device for gripping an object, or to an industrial robot.

[0089] The output unit 43 may be equipped with a communication interface (industrial network such as Ethernet / IP, PROFINET, CC-Link, EtherCAT, DeviceNet, etc.) with an industrial robot controller to achieve high-speed and highly reliable data transfer.

[0090] The data format output from the output unit 43 may be a standardized JSON, XML, or a binary format specific to a particular robot manufacturer. Additionally, for applications requiring real-time performance, a low-latency protocol (such as UDP / IP) and priority control functions may be used.

[0091] Furthermore, for visual feedback to the operator, the output unit 43 may be connected to a display equipped with a graphical user interface (GUI). This may enable the display of the gripping position of the item and the item recognition status.

[0092] In addition, the output unit 43 may be equipped with a database linkage interface (SQL, NoSQL, etc.) and a log file output function for logging processing results. It may also support multiple protocols to enable simultaneous cooperation with multiple different robot systems.

[0093] [Processing steps for recognition method] Figure 4 is a flowchart showing the method for recognizing marked articles according to this disclosure.

[0094] In step S301, the image acquisition unit 421 acquires an image of the item on which the marker has been applied via the input unit 41.

[0095] In step S303, the marker extraction unit 422 extracts markers from the acquired image.

[0096] In step S305, the installation location extraction unit 423 extracts the installation location for each marker.

[0097] In step S307, the information extraction unit 424 extracts item information from the marker. In addition to item information, additional information may also be extracted.

[0098] In step S309, the gripping position calculation unit 425 calculates the gripping position for each marker.

[0099] In step S311, markers relating to the same item are identified. If multiple markers are detected in the image, it is determined which markers belong to the same item. This identifies multiple markers (a group of markers) belonging to the same item.

[0100] Step S313 determines whether there are multiple markers for the same item.

[0101] If there are multiple markers for the same item (if the answer is YES in step S313), in step S315, the multiple gripping positions calculated from the multiple markers are averaged to calculate a single gripping position. This allows for a more accurate gripping position. The process then proceeds to step S317.

[0102] If there are no multiple markers for the same item (the result is NO in step S313), proceed to step S317.

[0103] In step S317, the output generation unit 426 outputs the calculated gripping position via the output unit 43.

[0104] [Specific examples of markers] Figure 5 shows an example of a marker. In Figure 5, the marker is shown as a two-dimensional code CD1. The two-dimensional code CD1 includes finder patterns FP1, FP2, and FP3.

[0105] The finder patterns FP1, FP2, and FP3 define axes QX and QY. Specifically, a two-dimensional coordinate system is defined with the center of finder pattern FP1 as the origin, the direction from finder pattern FP1 to finder pattern FP2 as axis QX, and the direction from finder pattern FP1 to finder pattern FP3 as axis QY. This two-dimensional coordinate system becomes the two-dimensional coordinate system of the surface of the marked article.

[0106] The two-dimensional code CD1 encodes item information. The item information includes at least information regarding the dimensions of the item to which the two-dimensional code CD1 is applied. In addition, the item information may include additional information. The additional information may include, for example, at least one of the following: the gripping position of the item, the weight of the item, the position based on the geometric center of gravity of the item, the position based on the physical center of gravity of the item, the method of gripping the item, the conditions for gripping the item, the permitted stacking method for the item, the number of stacking layers, and information specified by the user for the item.

[0107] As shown in Figure 5, the two-dimensional code CD1 includes finder patterns FP1, FP2, and FP3, which are geometric figures that identify the two-dimensional coordinate axes fixed to the surface on which the two-dimensional code CD1 is applied. The two-dimensional coordinate axes QX and QY are identified by the arrangement of these finder patterns FP1, FP2, and FP3. Note that the number of finder patterns is not limited to three, and a configuration with four or more finder patterns is also possible.

[0108] Figure 6 is a top view showing a first example of marker placement. The two-dimensional code CD1 identifies the two-dimensional coordinate axes (axis QX and axis QY) on the surface where the two-dimensional code CD1 is applied.

[0109] In the example shown in Figure 6, the two-dimensional code CD1 is applied to the upper surface of the item OBJ at a predetermined position. When the vertices of the four corners of the upper surface of the item OBJ are designated as points P1 to P4, the coordinates of points P1 to P4 in the two-dimensional coordinate system defined by axes QX and QY are represented as (x1, y1), (x2, y2), (x3, y3), and (x4, y4), respectively. Furthermore, the gripping position for gripping the item OBJ is indicated by point P5, and the coordinates of point P5 are shown as (x5, y5).

[0110] The following information is encoded in the two-dimensional code CD1, as information regarding the dimensions of the item OBJ and the gripping position of the item OBJ. "Box vertices P1(x1, y1), P2(x2, y2), P3(x3, y3), P4(x4, y4) Grasp position P5(x5, y5)"

[0111] Furthermore, the two-dimensional code CD1 may also encode information such as the dimensions of the item OBJ and the placement of the two-dimensional code on the surface of the item OBJ, as well as the height of the item OBJ, the load mass or the set maximum load mass, information on a gripper suitable for gripping the item OBJ (gripper type, usage conditions, required suction force, etc.), and the recommended stacking method and number of layers when stacking the item OBJ. This makes it possible to provide not only the identification of the gripping position and gripping method, but also the information necessary for the proper handling of the item OBJ in a unified manner.

[0112] Therefore, by obtaining the installation location of the two-dimensional code CD1 and the information encoded by the two-dimensional code CD1, the positions of points P1 to P4, which are the vertices of the four corners on the top surface of the item OBJ, can be specifically obtained based on the installation location of the two-dimensional code CD1. Furthermore, since the positions of points P1 to P4 are known, the positions of the edges on the top surface of the item OBJ on which the two-dimensional code CD1 is installed can be specifically obtained.

[0113] Similarly, by obtaining the installation location of the two-dimensional code CD1 and the information encoded by the two-dimensional code CD1, it becomes possible to identify point P5, which is the gripping position for gripping the item OBJ, based on the installation location of the two-dimensional code CD1.

[0114] Furthermore, the two-dimensional code CD1 may not specify the coordinate information of point P5. The gripping position for gripping the item OBJ may be determined by calculating, for example, the geometric center of the item OBJ, the centroid of the item OBJ, or the optimal position for handling the item OBJ. In other words, the information of point P5 may be directly encoded in the two-dimensional code CD1, or it may be calculated based on the information obtained from the two-dimensional code CD1.

[0115] By detecting the two-dimensional code, it is possible to identify the appropriate gripping position of the item OBJ, along with information on its dimensions and shape. This enables efficient gripping of the item OBJ in an automated item OBJ handling system.

[0116] Figure 7 is a top view showing a second example of marker placement. Two-dimensional code CD1 identifies the two-dimensional coordinate axes (axes QX and QY) on the surface where two-dimensional code CD1 is applied. Similarly, two-dimensional code CD2 identifies the two-dimensional coordinate axes (axes QA and QB) on the surface where two-dimensional code CD2 is applied.

[0117] In the example shown in Figure 7, multiple two-dimensional codes CD1 and CD2 are applied to the upper surface of the item OBJ at predetermined positions. When the vertices of the four corners of the upper surface of the item OBJ are designated as points P1 to P4, the coordinates of points P1 to P4 in the two-dimensional coordinate system defined by axes QA and QB are represented as (a1, b1), (a2, b2), (a3, b3), and (a4, b4), respectively. Furthermore, the gripping position for gripping the item OBJ is indicated by point P5, and the coordinates of point P5 are shown as (a5, b5).

[0118] The following information is encoded in the two-dimensional code CD2, as information regarding the dimensions of the item OBJ and the gripping position of the item OBJ. "Box vertices P1(a1, b1), P2(a2, b2), P3(a3, b3), P4(a4, b4) Grasp position P5(a5, b5)"

[0119] Here, the first two-dimensional code CD1 is applied to one position on one side of the article OBJ, and the second two-dimensional code CD2 is applied to the opposite side of the gripping center on the same side. Specifically, the first two-dimensional code CD1 is located in the upper left portion of the article OBJ, and the second two-dimensional code CD2 is located in the lower right portion.

[0120] Each two-dimensional code encodes information regarding the dimensions of the object (OBJ) and the placement of that code on the surface of the object. This arrangement allows information to be obtained from one of the two-dimensional codes even if the other is not visible.

[0121] Furthermore, if both two-dimensional codes are detected, the gripping position information obtained from each two-dimensional code can be averaged to determine the gripping position with greater accuracy. For example, the average of the gripping position obtained from two-dimensional code CD1 and the gripping position obtained from two-dimensional code CD2 can be set as the final gripping position.

[0122] In this way, by placing multiple two-dimensional codes on one surface of the object OBJ, information redundancy is ensured and the accuracy of gripping position detection is improved. Furthermore, even if some two-dimensional codes are not visible due to the orientation of the object OBJ, it is possible to obtain the necessary information from the remaining two-dimensional codes, which enhances the robustness of the system. There may be two or more markers placed on one surface of the object OBJ. In this way, multiple distinct markers may be placed on one surface of the object OBJ. Note that "distinct" means that at least one of the placement position and pattern of the markers is different. Multiple markers that differ only in placement position but have the same pattern are also included in the concept of "distinct markers".

[0123] Figure 8 is a top view showing a third example of marker placement. Two-dimensional code CD1 identifies the two-dimensional coordinate axes (axes QX and QY) on the surface where two-dimensional code CD1 is applied. Two-dimensional code CD2 identifies the two-dimensional coordinate axes (axes QA and QB) on the surface where two-dimensional code CD2 is applied. Furthermore, two-dimensional code CD3 identifies the two-dimensional coordinate axes (axes QS and QT) on the surface where two-dimensional code CD3 is applied.

[0124] In the example shown in Figure 8, multiple two-dimensional codes CD1, CD2, and CD3 are applied to the upper surface of the item OBJ at predetermined positions. When the vertices of the four corners of the upper surface of the item OBJ are designated as points P1 to P4, the coordinates of points P1 to P4 in the two-dimensional coordinate system defined by axes QS and QT are represented as (s1, t1), (s2, t2), (s3, t3), and (s4, t4), respectively. Furthermore, the gripping position for gripping the item OBJ is indicated by point P5, and the coordinates of point P5 are shown as (s5, t5).

[0125] The following information is encoded in the two-dimensional code CD3, as information regarding the dimensions of the item OBJ and the gripping position of the item OBJ. "Box vertices P1(s1, t1), P2(s2, t2), P3(s3, t3), P4(s4, t4) Grasp position P5(s5, t5)"

[0126] As shown in Figure 8, three or more markers may be placed on one surface of the item OBJ. This allows for more accurate spatial position identification. In the example shown in Figure 8, two-dimensional codes CD1, CD2, and CD3 are applied to the top surface of the item OBJ at different locations on the item OBJ.

[0127] For example, multiple two-dimensional codes may be applied to different corners of the upper surface of an item OBJ. Specifically, two-dimensional codes CD1, CD2, and CD3 may be applied to the three corners of the upper surface of the item OBJ, respectively. Alternatively, two-dimensional codes may be applied to all four corners of the upper surface of the item OBJ.

[0128] Generally, three coordinate points are required to uniquely identify a plane in space. While it is possible to identify a plane using three finder patterns FP1, FP2, and FP3 contained within a single two-dimensional code, it is also possible to identify a plane using the finder patterns of each two-dimensional code together. This allows for more accurate plane identification compared to identifying a plane based on finder patterns contained within a single two-dimensional code. By using three finder patterns placed sufficiently far apart, the impact of small detection errors on the overall accuracy of plane identification can be suppressed.

[0129] Specifically, by using three points—finger pattern FP1 of 2D code CD1, finder pattern FP1 of 2D code CD2, and finder pattern FP1 of 2D code CD3—to identify the plane of the top surface of the object OBJ, it becomes possible to correct image distortion caused by changes in the camera's viewpoint and the effects of the shooting angle with high precision. This allows for the accurate calculation of the object OBJ's position and orientation in 3D space, significantly improving the gripping accuracy of the robot.

[0130] Furthermore, there are other advantages to applying a 2D code to the corners of an object object (OBJ). Corners are characteristic points that define the external shape of an object object, and if the position of the corners can be identified, the overall shape of the object object can be estimated. Also, since corners are in positions where interference is less likely to occur during gripping operations, there is less risk of the 2D code being obscured when gripping devices such as robot hands or grippers grasp the object object object. Moreover, even when multiple object objects are placed adjacent to each other, the 2D code applied to the corners is highly visible, making it possible to clearly identify each individual object object.

[0131] Alternatively, the gripping position may be calculated from the positional relationship of multiple two-dimensional codes, without directly using the gripping position information (point P5) encoded in the two-dimensional code. For example, the plane (gripping surface) on which the three two-dimensional codes exist can be identified from the three-dimensional positions of the finder patterns FP1 of the three two-dimensional codes applied to the upper surface of the object OBJ. Then, the midpoint of the line connecting the finder patterns FP1 of two two-dimensional codes located diagonally apart may be set as the center of the gripping position, and the direction parallel to the line connecting the finder patterns FP1 of two two-dimensional codes located on the long side of the object OBJ may be set as the gripping direction.

[0132] This eliminates the need to encode gripping position information within the two-dimensional code itself, allowing for a simpler code structure. In particular, simply applying a standardized two-dimensional code pattern to different locations makes it possible to identify the gripping position.

[0133] Furthermore, when applying four two-dimensional codes to the upper surface of an object OBJ, the intersection of the two diagonals connecting the finder patterns FP1 of two opposing sets of two-dimensional codes may be set as the gripping center. This makes it possible to pinpoint the geometric center of the object OBJ with high precision.

[0134] Figures 6-8 above show an example of using a two-dimensional code as a marker applied to an object OBJ, but it is also possible to use a one-dimensional code instead of a two-dimensional code. For example, one-dimensional codes may be placed along axes QX and QY, respectively, as shown in Figures 6 and 7.

[0135] In a one-dimensional code, information such as numbers or letters is represented by arranging bars of different widths or spacings from a starting position to an ending position. Here, a two-dimensional coordinate system can be defined by defining axes QX or QY based on the direction from the starting position to the ending position of the one-dimensional code.

[0136] Figure 9 is a top view showing a fourth example of marker placement. Unlike the examples shown in Figures 6-8, markers MK containing geometric shapes may be used instead of one-dimensional or two-dimensional codes.

[0137] A marker MK is applied to the top surface of the object OBJ to facilitate "edge detection," and its geometric center is positioned to coincide with the center of the gripping position of the object OBJ. The marker MK is formed as a frame that is sufficiently thick, has high contrast with the color of the object OBJ, and is less affected by ambient lighting conditions.

[0138] The marker MK can be detected using image processing technology, and its center position can be calculated. Alternatively, the gripping position can be more explicitly detected by applying a clear geometric shape to the gripping position. While such geometric shapes cannot directly encode information about the object OBJ, their simplicity offers advantages in terms of detection reliability and processing speed.

[0139] For example, the marker MK functions even when there is an opening (not shown) on the top surface of the item OBJ, or when the item is sealed and packaged with opaque tape TP. Even if the marker MK is divided by the tape TP, it can still be recognized using detection techniques such as image moment analysis to detect characteristic shapes or machine learning-based object detection algorithms. Alternatively, by targeting "two facing U-shaped figures," the possibility of detection can be increased even if part of the marker MK is hidden.

[0140] Furthermore, considering cases where the edges of the item OBJ (the opening edge EG) are sealed during packaging, the edges of the marker MK frame, which run parallel to the opening edge EG of the item OBJ, may be pre-installed by shifting them sufficiently toward the center of the item OBJ. In this case, by also installing a note stating, "Pack in a way that does not conceal this frame," it is possible to ensure the visibility of the marker MK.

[0141] The example shown in Figure 9, which uses markers MK containing geometric shapes, can be used when the shape and maximum load mass of the item OBJ are known and detailed additional information is not required. Even with simple markers, the gripping position of the item OBJ can be easily estimated. By applying easily detectable geometric shapes to the top surface of the item OBJ and detecting them with a camera, the objective of determining the gripping position of the item OBJ can be achieved.

[0142] Figure 10 is a top view showing an example of multiple items being stacked adjacent to each other. In Figure 10, nine item objects (OBJs) are arranged in a 3x3 grid, each adjacent to the others, and each item object object is marked. The method of arranging the item objects adjacent to each other and the number of item objects arranged adjacent to each other are not limited to the example given here.

[0143] When items (OBJs) are stacked adjacent to each other, it is difficult to detect the boundaries between them using cameras or laser radar. For example, when items with similar colors or shapes are placed close together, it is difficult to identify each item individually using image processing technology alone, which can hinder the accurate operation of automated picking systems.

[0144] In contrast, according to this disclosure, by placing markers at predetermined positions on each article OBJ, it becomes possible to clearly identify each article OBJ even when they are adjacent to each other. As shown in Figure 10, markers are placed at diagonal positions (upper left and lower right) of each article OBJ, making it easy to identify the contour and gripping position of each article OBJ. This is because markers provide high-contrast and clear visual features. In particular, when article OBJs of the same color and shape are densely packed together, it can be difficult to distinguish the boundaries of the article OBJs with normal image processing, but the presence of markers allows for clear identification of each article OBJ. By using markers that are easy to detect even when lighting conditions and shooting angles change, the contour and gripping position of article OBJs can be consistently identified in various environments.

[0145] Each marker contains encoded information specific to the item's object-oriented jig (OBJ), such as the item's dimensions or gripping position, which is necessary for palletizing or depalletizing. Therefore, simply by detecting the markers with a camera or similar device, it becomes possible to identify the precise location and orientation of each item's OBJ, as well as the appropriate gripping position, even in a densely packed arrangement.

[0146] Furthermore, by placing multiple markers at diagonal positions of the object object (OBJ), the gripping position can be calculated with greater accuracy. For example, by adopting the average value of the gripping position calculated from multiple markers detected for each object object, the accuracy of position detection is improved. In addition, even if some markers are hidden by other object objects or obstacles, the necessary information can be obtained from the remaining markers, thus improving the robustness of the system.

[0147] Thus, even in environments where multiple item objects (OBJs) are placed adjacent to each other, this disclosure makes it possible to accurately identify each item object and determine the appropriate gripping position. This significantly improves the efficiency and reliability of automated picking systems in warehouses and factories.

[0148] [Effects of the Embodiment] As described in detail above, the manufacturing apparatus, manufacturing method, and manufacturing program for marker-attached articles relating to this disclosure use a controller connected to an input unit and an installation unit. The controller acquires article information, including at least information regarding the dimensions of the article, via the input unit, generates a marker based on the article information and the installation position of the marker on the surface of the article, outputs the marker and installation position to the installation unit, and installs the marker at the installation position via the installation unit. Based on the installation position and the marker, a gripping position for gripping the article can be identified.

[0149] This allows for obtaining information about the packaging of items in environments where databases cannot be easily accessed, thereby improving the efficiency of item handling and transportation. In particular, it enables more accurate identification of the item's gripping position. Even in situations where item boundary detection or individual identification is difficult, attaching markers to each item facilitates detection by cameras or laser radar, improving the efficiency and accuracy of the automated picking system.

[0150] The marker may include a geometric figure that identifies a two-dimensional coordinate axis fixed to the surface on which the marker is applied. This makes it possible to establish a clear coordinate system on the surface of the object. Not only can the position of any point on the object be accurately determined based on the position of the marker, but the orientation and orientation of the object can also be accurately detected. Defining the two-dimensional coordinate axis allows for accurate coordinate transformation even if there is spatial distortion or a change in viewpoint when the object is imaged by a camera or sensor, enabling high-precision mapping of the object's gripping position to the robot's motion coordinate system.

[0151] The marker may include a one-dimensional or two-dimensional code. This makes it possible to simultaneously encode and store not only the location information of the item but also a wealth of additional information. These codes have a high information density and can store diverse data such as the dimensions, weight, center of gravity, allowable gripping force, and stacking conditions of the item in a limited space, allowing for immediate determination of the optimal handling method for each item. Furthermore, these codes are based on globally standardized technology and have high compatibility with existing image recognition systems and reading devices, enabling high recognition accuracy while keeping implementation costs down.

[0152] The marker may identify the coordinates of the surface on which the marker is installed, based on the installation location. This makes it possible to directly and accurately establish the correspondence between the actual dimensions of the object and the image coordinate system. Since the three-dimensional position or orientation of the object in the real world can be calculated with high accuracy from the two-dimensional image acquired by the camera or sensor, the accuracy of the robot's gripping operation is improved. Furthermore, even when multiple objects are stacked at various angles or arrangements, the relative positional relationship of each object can be accurately grasped based on the coordinate information obtained from each marker.

[0153] Furthermore, in logistics centers or manufacturing sites where the dimensions and shapes of goods vary, this system can significantly reduce manual measurement and complex calibration work by workers while simultaneously improving the flexibility and accuracy of automated systems.

[0154] The marker may have three or more finder patterns, and the arrangement of the finder patterns may determine the two-dimensional coordinate axes. This allows for the establishment of a robust and accurate coordinate system on the surface of the object. By using multiple finder patterns, accurate projection transformation is possible even with distortion due to changes in the camera's viewpoint or shooting angle, enabling stable coordinate detection regardless of the orientation and tilt of the object.

[0155] The installation location may be on the same plane as the surface where the gripping position appears. This makes it possible to identify the gripping position by having the camera or sensor capture only a single surface. This simplifies the placement of the imaging device and the sensing process, resulting in a simpler overall system configuration. In particular, when multiple items are stacked on a pallet, the gripping position of each item can be identified by taking a picture from above, eliminating the need for photography from multiple directions, leading to faster recognition processing and reduced processing load. Furthermore, because the positional relationship on the same plane can be measured directly, errors associated with coordinate transformation are reduced, improving the accuracy of gripping position detection.

[0156] Multiple different markers may be applied to one surface of an object. This allows for the acquisition of multiple pieces of information from a single surface, improving identification accuracy, enhancing error tolerance through redundancy, and enabling compatibility with various reading devices. Furthermore, because each marker has a complementary function, overall functionality is maintained even if some markers are damaged, improving reliability.

[0157] Multiple distinct markers may include a first marker and a second marker, and on one side of the article, the second marker may be placed on the opposite side of the gripping center of the article from the first marker. This allows for the integration of information from multiple markers to improve the accuracy of the gripping position. Specifically, by calculating the gripping position independently from each marker and adopting their average value, accidental errors in position detection can be offset, enabling more accurate identification of the gripping position. Furthermore, depending on the arrangement of the articles, one marker may be hidden by other articles or obstacles, but with this configuration, the gripping position can be calculated from the remaining markers, thus improving the robustness of the system.

[0158] The marker may also identify additional information about the item. This makes it possible to provide not only information indicating the gripping position, but also a variety of information necessary for efficient and safe handling of the item.

[0159] The additional information may include at least one of the following: the gripping position of the item, the weight of the item, the position based on the geometric center of gravity of the item, the position based on the physical center of gravity of the item, the gripping method of the item, the conditions for gripping the item, the acceptable stacking method for the item, the number of stacking layers, and user-specified information for the item. This allows for the adoption of a gripping strategy optimized for each item, enabling appropriate handling of items with various characteristics. As a result, the success rate of picking and palletizing operations improves, and the overall efficiency and safety of the logistics process are enhanced. Furthermore, this additional information contributes to the traceability of items and can also contribute to improving the accuracy of management throughout the entire supply chain.

[0160] In the marked article relating to this disclosure, the marker may hold article information, which includes at least information relating to the dimensions of the article, and encoded information based on the placement of the marker on the surface of the article, and may be able to identify a gripping position for gripping the article based on the placement and encoded information.

[0161] This allows for obtaining information about the packaging of goods in environments where databases cannot be easily accessed, thereby improving the efficiency of handling and transporting goods. In particular, it enables the precise identification of the gripping position of goods. Specifically, the goods themselves can "carry" their own dimensional information and gripping position information. As a result, even when goods are moved between different factories or logistics centers, the local automated system can instantly acquire item-specific information simply by reading a marker, identify the appropriate gripping position, and achieve efficient handling. In other words, since item information can be obtained directly from the goods themselves without relying on databases or external systems, compatibility between different systems is ensured, and delays and errors in information transmission can be prevented.

[0162] The recognition device, recognition method, and recognition program for marked articles relating to this disclosure use a controller connected to an input unit and an output unit. The controller acquires an image of an article marked with a marker via the input unit, extracts the marker and the placement of the marker on the surface of the article from the image, calculates a gripping position for gripping the article based on the marker and placement, and outputs the gripping position via the output unit.

[0163] This allows for obtaining information about the packaging of goods in environments where databases cannot be easily accessed, thereby improving the efficiency of handling and transporting goods. In particular, it enables the precise identification of the gripping position of goods. Furthermore, even if the shape or appearance of goods is diverse, a unified recognition algorithm can be applied by using a common identification means called a marker, improving the versatility and scalability of the system. Since the relationship from the marker to the gripping position is pre-encoded, the gripping position can be calculated directly without using complex shape recognition processing or machine learning, resulting in reduced processing load and improved recognition speed, which is expected to significantly improve work efficiency in logistics.

[0164] The controller may identify a two-dimensional coordinate axis fixed to the surface on which the marker is applied, based on the shape of the marker. This makes it possible to accurately determine the coordinate system on the surface of the object, regardless of the orientation or orientation of the object. Coordinate transformations can be performed appropriately even when the camera's shooting angle or shooting distance changes, enabling high-precision calculation of the position in three-dimensional space of any point on the surface of the object (especially the gripping position). Furthermore, since the coordinate axes can be identified from the shape of the marker itself, additional reference points or calibration jigs are unnecessary, simplifying the system configuration and reducing implementation costs.

[0165] The controller may calculate the coordinates of the surface on which the marker is installed, based on the marker and the installation location, with the installation location as the reference point. This allows for the establishment of an item-specific coordinate system, making it possible to identify the gripping position in a unified manner regardless of the dimensions or shape of the item. By adopting this coordinate system based on the installation location, the relative positional relationship of markers can be consistently interpreted even between different items, making it possible to construct a general-purpose recognition algorithm that can handle a wide variety of items.

[0166] The controller may, when multiple markers are extracted from an image, calculate a gripping position for each marker, identify a group of markers relating to the same item based on the multiple gripping positions, and then average the multiple gripping positions relating to the group of markers to calculate a single gripping position. This significantly improves the accuracy of gripping position detection. The gripping positions calculated individually from each marker may contain errors due to factors such as the camera's shooting angle, lighting conditions, and noise in image processing, but by averaging multiple measurements, these errors can be statistically offset, allowing for the identification of a more reliable gripping position.

[0167] Each of the functions described in the embodiments above may be implemented by one or more processing circuits. These processing circuits may include programmed processors, electrical circuits, and other devices such as application-specific integrated circuits (ASICs), or circuit components arranged to perform the described functions.

[0168] Furthermore, computer programs may be provided on computer-readable, non-temporary recording media such as CD-ROMs (Compact Disc-Read Only Memory). Computer programs may also be provided via telecommunication lines such as the Internet.

[0169] According to this disclosure, in environments where databases cannot be easily accessed, information on the packaging of goods can be obtained, thereby improving the efficiency of handling and transporting goods, and increasing the labor productivity of users involved in handling and transporting goods. As a result, for example, it can contribute to Goal 8 of the United Nations Sustainable Development Goals (SDGs), "Promote inclusive and sustainable economic growth and full and productive employment and decent work for all."

[0170] Although several embodiments have been described, it is possible to modify or transform the embodiments based on the above disclosure. All components of the above embodiments, and all features described in the claims, may be taken individually and combined, provided that they do not conflict with each other. [Explanation of Symbols]

[0171] 20 Manufacturing equipment 21 Input section 22 controllers 24 Construction Department 221 Information Acquisition Department 222 Grip position setting section 223 Construction position setting section 224 Marker generation unit 225 Output generation unit 40 Recognition device 41 Input section 42 controllers 43 Output section 421 Image acquisition unit 422 Marker extraction section 423 Construction position extraction part 424 Information extraction part 425 Grip position calculation unit 426 Output generation unit

Claims

1. A manufacturing apparatus comprising an input unit, a construction unit, and a controller, The aforementioned controller, Through the input unit, article information including at least information regarding the dimensions of the article is acquired. Based on the aforementioned article information and the placement position of the marker on the surface of the article, the marker is generated. The marker and the installation position are output to the installation unit. The marker is installed at the installation location via the installation section. Based on the aforementioned construction position and the aforementioned marker, the gripping position for gripping the article can be identified. Manufacturing equipment.

2. The manufacturing apparatus according to claim 1, wherein the marker includes a geometric figure that identifies a two-dimensional coordinate axis fixed to the surface on which the marker is applied.

3. The manufacturing apparatus according to claim 2, wherein the marker includes a one-dimensional code or a two-dimensional code.

4. The manufacturing apparatus according to claim 1, wherein the marker identifies the coordinates of the surface on which the marker is installed, with respect to the installation position.

5. The aforementioned marker has three or more finder patterns, The arrangement of the finder pattern determines the two-dimensional coordinate axes. The manufacturing apparatus according to claim 2.

6. The manufacturing apparatus according to claim 1, wherein the work position is arranged on the same plane as the surface in which the gripping position appears.

7. The manufacturing apparatus according to claim 1, wherein a plurality of different markers are applied to one surface of the article.

8. The manufacturing apparatus according to claim 7, wherein the plurality of distinct markers include a first marker and a second marker, and the second marker is applied on one surface of the article on the opposite side of the gripping center of the article from the first marker.

9. The manufacturing apparatus according to claim 1, wherein the marker identifies additional information relating to the article.

10. The manufacturing apparatus according to claim 9, wherein the additional information includes at least one of the following: the gripping position of the article, the weight of the article, the position of the article based on its geometric center of gravity, the position of the article based on its physical center of gravity, the method of gripping the article, the conditions for gripping the article, the permitted stacking method for the article, the number of stacking layers, and information specified by the user for the article.

11. A manufacturing method for controlling an input unit and a controller connected to an installation unit, The aforementioned controller, Through the input unit, article information including at least information regarding the dimensions of the article is acquired. Based on the aforementioned article information and the placement position of the marker on the surface of the article, the marker is generated. The marker and the installation position are output to the installation unit. The marker is installed at the installation location via the installation section. Based on the aforementioned construction position and the aforementioned marker, the gripping position for gripping the article can be identified. Manufacturing method.

12. A manufacturing program executed in a controller connected to an input unit and a construction unit, The aforementioned controller The steps include obtaining article information, which includes at least information regarding the dimensions of the article, via the input unit, A step of generating the marker based on the article information and the placement position of the marker on the surface of the article, The steps include outputting the marker and the installation position to the installation unit, The steps include: installing the marker at the installation location via the installation section; Equipped with, Based on the aforementioned construction position and the aforementioned marker, the gripping position for gripping the article can be identified. Manufacturing program.

13. Articles on which markers have been applied, The marker holds article information, which includes at least information relating to the dimensions of the article, and encoded information based on the installation position of the marker on the surface of the article. Based on the aforementioned construction location and the aforementioned encoded information, the gripping position for gripping the article can be identified. Goods.

14. A recognition program executed in a controller connected to an input unit and an output unit, The aforementioned controller, The steps include: acquiring an image of an item on which a marker has been applied via the input unit; The steps include extracting the marker and the placement position of the marker on the surface of the article from the aforementioned image, A step of calculating a gripping position for gripping the article based on the marker and the installation position, The steps include outputting the gripping position via the output unit, A recognition program equipped with the following features.

15. The aforementioned controller, The steps include: identifying a two-dimensional coordinate axis fixed to the surface on which the marker is to be installed, based on the shape of the marker; The recognition program according to claim 14, further comprising the following:

16. The aforementioned controller, Based on the marker and the installation location, the steps include calculating the coordinates of the surface on which the marker is installed, with respect to the installation location. The recognition program according to claim 14, further comprising the following:

17. The aforementioned controller, When multiple markers are extracted from the aforementioned image, the step of calculating the gripping position for each marker, A step of identifying a group of markers relating to the same article based on multiple gripping positions, A step of averaging the multiple gripping positions relating to the group of markers to calculate a single gripping position, The recognition program according to claim 14, further comprising the following:

Citation Information

Patent Citations

  • Packing recognition device

    JP2012192490A