An intelligent mold quality detection method, system, device and medium

Abstract

The application provides a kind of intelligent mold quality detection method and system, equipment, medium, belong to quality detection technical field, this method includes: control robot obtains the first mold image of the mold to be detected on mechanical turntable;Target mold image and target mold information corresponding to target mold image are obtained from mold database based on the first mold image;Target mold information includes mold scanning control parameter, mold type and mold label information;Based on mold scanning control parameter, the mold to be detected on mechanical turntable is scanned to obtain mold scanning data based on robot, and the volume of the mold is determined based on mold scanning data and mold three-dimensional profile data;Based on mold label information, the image data of the mold is obtained, and the mold quality detection result is determined based on image data, the volume of the mold and mold three-dimensional profile data.The application can realize the high-precision quality detection of diversified glass mold.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of quality detection, and more particularly relates to an intelligent mold quality detection method and system, equipment and medium. BACKGROUND

[0002] In the field of glass product production, a glass mold is a core tool for determining the shape, volume and surface precision of a product, and the quality of the glass mold directly affects the functionality and appearance compliance of glass bottles and other products, so professional mold quality detection equipment is needed to accurately detect the mold.

[0003] With the increasing demand for personalization in the glass product market, mold detection requirements are higher. In the prior art, mold detection relies on tool measurement of key dimensions of the mold and image recognition of contour lines to generate a detection reference. However, this approach not only has low efficiency, but also has high size measurement deviation and large image recognition error, resulting in insufficient reliability of the detection results and prone to misjudgment of unqualified molds, increasing production costs and quality risks.

[0004] Therefore, an intelligent mold quality detection method is needed to further improve the mold quality detection accuracy while ensuring detection efficiency. SUMMARY

[0005] The application aims to provide an intelligent mold quality detection method and system, equipment and medium to further improve mold quality detection accuracy while ensuring detection efficiency.

[0006] The first aspect of the embodiment of the application provides an intelligent mold quality detection method, which is executed by a control terminal device, and the control terminal device is in communication connection with a robot and a mechanical turntable; the mechanical turntable is used to place a mold to be detected, and the robot is used to collect data of the mold to be detected placed on the mechanical turntable and send the collected data to the control terminal device; the control terminal device is used to control the working state of the robot and the mechanical turntable, and obtain a mold quality detection result based on the data sent by the robot;

[0007] The method comprises:

[0008] In response to receiving a mold detection instruction, the robot acquires a first mold image of the mold to be detected on the mechanical turntable; in response to receiving the first mold image sent by the robot, target mold image and target mold information corresponding to the target mold image are obtained from the mold database based on the first mold image; the target mold image is an image matched with the first mold image, and the target mold information includes mold scanning control parameters, mold type and mold label information;

[0009] The robot scans the mold to be detected on the mechanical turntable based on the mold scanning control parameter to obtain mold scanning data, and determines a mold volume value and mold three-dimensional profile data based on the mold scanning data.

[0010] If the mold type is a glass mold containing an image, image data of the mold is obtained based on the mold label information, and a mold quality detection result is determined based on the image data, the mold volume value and the mold three-dimensional profile data; if the mold type is a glass mold not containing an image, the mold quality detection result is determined based on the mold volume value and the mold three-dimensional profile data.

[0011] In a second aspect of the embodiment, an intelligent mold quality detection system is provided, which is applied to a control terminal device, and the control terminal device is in communication connection with a robot and a mechanical turntable; the mechanical turntable is used for placing a mold to be detected, the robot is used for collecting data of the mold to be detected placed on the mechanical turntable, and the collected data is sent to the control terminal device; the control terminal device is used for controlling working states of the robot and the mechanical turntable, and obtaining a mold quality detection result based on the data sent by the robot.

[0012] The system comprises:

[0013] A mold image recognition module is configured to, in response to receiving a mold detection instruction, control the robot to obtain a first mold image of a mold to be detected on the mechanical turntable, and in response to receiving the first mold image sent by the robot, obtain a target mold image and target mold information corresponding to the target mold image from a mold database based on the first mold image; the target mold image is an image matched with the first mold image, and the target mold information comprises mold scanning control parameters, a mold type and mold label information.

[0014] A quality detection module is configured to control the robot to scan the mold to be detected on the mechanical turntable based on the mold scanning control parameter to obtain mold scanning data, and determine a mold volume value and mold three-dimensional profile data based on the mold scanning data.

[0015] A quality analysis module is configured to, if the mold type is a glass mold containing an image, obtain image data of the mold based on the mold label information, and determine a mold quality detection result based on the image data, the mold volume value and the mold three-dimensional profile data; if the mold type is a glass mold not containing an image, determine the mold quality detection result based on the mold volume value and the mold three-dimensional profile data.

[0016] In a third aspect of the embodiment, a control terminal device is provided, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the steps of the intelligent mold quality detection method when executing the computer program.

[0017] In a fourth aspect, the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the intelligent mold quality detection method.

[0018] The intelligent mold quality detection method and system, device and medium provided by the embodiment of the present application have the following beneficial effects:

[0019] The embodiment of the present application controls the terminal device to automatically control the robot to obtain the first mold image, and controls the terminal device to automatically identify the mold to be detected and related information based on the image and the matching of the mold scanning control parameter, the mold type and the mold label information from the database. Meanwhile, the robot and the mechanical turntable cooperatively scan according to the preset parameter, automatically complete data acquisition and analysis, shorten the time consumption of single mold detection, quickly respond to the mold detection demand, and improve the automation level and the mold quality detection efficiency.

[0020] The embodiment of the present application accurately controls the scanning process based on the mold scanning control parameter, and determines the mold volume value and the three-dimensional profile data based on the mold scanning data. Through this standardized step, the error can be reduced. Meanwhile, considering that part of the mold has local images, the traditional detection method cannot effectively detect the images. Therefore, the embodiment of the present application divides the mold into two types including images and no images, carries out differential quality detection processes for the two types of molds, automatically switches the detection logic according to the mold type, improves the comprehensive quality detection precision of the mold with images, and makes the present application widely applicable to the quality detection of various glass molds and meet diversified needs.

[0021] To sum up, the embodiment of the present application realizes the automation and intelligence of the mold quality detection through the communication connection of the terminal device, the robot and the mechanical turntable. The robot can accurately obtain the image data of the mold to be detected, and perform high-precision scanning based on the mold scanning control parameter to obtain accurate mold volume value and mold three-dimensional profile data, thereby improving the detection efficiency and precision. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0023] Figure 1 A flowchart of an intelligent mold quality detection method provided by an embodiment of the present application is shown in the figure.

[0024] Figure 2A structural schematic diagram of a mold quality detection device provided by an embodiment of the present application is provided.

[0025] Figure 3 A second cross-sectional profile data schematic diagram of a mold provided by an embodiment of the present application is provided.

[0026] Figure 4 A structural block diagram of an intelligent mold quality detection system provided by an embodiment of the present application is provided.

[0027] Figure 5 A schematic block diagram of a control terminal device provided by an embodiment of the present application is provided. DETAILED DESCRIPTION

[0028] In the following description, specific details are set forth in order to provide a thorough understanding of embodiments of the present application. However, persons having ordinary skill in the art will readily understand that embodiments of the present application can be practiced without these specific details. In other instances, well-known structures, devices, circuits, and methods have not been described in detail in order to avoid obscuring the description of the present application.

[0029] In order to make the purpose, technical scheme and advantages of the present application clearer, specific embodiments will be described below with reference to the accompanying drawings.

[0030] Reference is made to Figure 1 , Figure 1 A flowchart of an intelligent mold quality detection method provided by an embodiment of the present application is provided. The method can be executed by a control terminal device, which is in communication connection with a robot and a mechanical turntable respectively. The mechanical turntable is used to place a mold to be detected, and the robot is used to collect data of the mold to be detected placed on the mechanical turntable and send the collected data to the control terminal device. The control terminal device is used to control the working states of the robot and the mechanical turntable, and obtain a mold quality detection result based on the data sent by the robot. Specifically, the method can include S101-S103.

[0031] S101: In response to receiving a mold detection instruction, the robot is controlled to obtain a first mold image of a mold to be detected on the mechanical turntable. In response to receiving the first mold image sent by the robot, a target mold image and target mold information corresponding to the target mold image are obtained from a mold database based on the first mold image. The target mold image is an image matched with the first mold image, and the target mold information includes mold scanning control parameters, a mold type and mold label information.

[0032] In the embodiment, the target mold image is obtained from the mold database based on the first mold image, specifically including: performing initial feature extraction on the first mold image to obtain contour type features, size features and pattern features; obtaining a candidate mold image set from the mold database based on the contour type features, size features and pattern features; if the candidate mold image set includes one mold image, the mold image is taken as the target mold image; if the candidate mold image set includes at least two mold images, the similarity of the first mold image and each mold image in the candidate mold image set is calculated respectively, and the mold image with the highest similarity to the first mold image is taken as the target mold image.

[0033] In the embodiment, the robot can be a movable mechanical arm, and the robot can include a camera and a laser scanning device. The mold detection instruction is a control signal triggering the mold quality detection process, used to start the detection device to perform a preset detection action. The mold detection instruction can be an instruction issued by the staff operating on the control terminal device, and the control terminal device can receive the instruction. The first mold image is an initial visual image collected by the robot using its own camera on the mechanical turntable for detecting the mold, serving as the basic data source for mold matching. The mold database is a structured data set for storing various standard mold images and supporting information, used to provide a matching reference. The target mold information is the full-quantity mold data associated with the target mold image, which can include mold scanning control parameters, mold type identification and mold label information, etc.

[0034] The mold scanning control parameters refer to specific parameters for controlling the robot and the mechanical turntable to perform scanning actions, which can include robot movement paths (such as Y-axis linear paths along the mold height direction, etc.), robot movement speeds (such as 0 mm / s~2 mm / s), laser scanning frequencies (such as 300 Hz~800 Hz), mechanical turntable rotation speeds (such as 0° / s~15° / s), etc. The mold type is a category based on whether the mold contains an image, used to distinguish the detection logic. Each mold type corresponds to a type identification, such as "1" representing a glass mold containing an image, and "0" representing a glass mold not containing an image. The mold label information is information for uniquely identifying the mold and associated basic attributes, which can include mold unique number, mold design version and mold applicable product model (such as 500 ml glass bottle), etc.

[0035] Contour type features reflect the overall shape and contour of the mold, used to initially distinguish mold forms. These features may include contour shape, contour symmetry, and the number of key inflection points. Dimensional features reflect the mold's geometric dimensions, used to quantitatively distinguish mold specifications. These may include the mold's maximum diameter, mold height, key cross-sectional dimensions (such as bottle mouth diameter and bottle body mid-section diameter), and allowable dimensional deviations. Pattern features reflect the surface pattern attributes of the mold. These may include the presence or absence of pattern markings (e.g., "0" represents no pattern, "1" represents pattern), pattern type, pattern location, and coordinates of key pattern feature points. The candidate mold image set is a collection of mold images obtained from the mold database based on the initial features, providing a range for further precise matching.

[0036] The underlying consideration of this embodiment is to automate the retrieval of mold detection parameters through image feature matching, replacing manual selection of detection programs and improving detection efficiency and accuracy. Specifically, this embodiment selects three initial features: contour type, size, and pattern, as these three features are the most core distinguishing dimensions for glass molds. The contour determines the basic shape of the mold, the size determines the mold specifications, and the pattern determines the unique appearance of the mold. Combining these three features can quickly narrow down the matching range and reduce database retrieval and computation costs. This embodiment divides the candidate set and processes them accordingly to obtain matching results. A single image directly identifies the target, while multiple images are sorted by similarity to select the optimal one. This avoids the risk of mismatch in single feature matching and ensures matching accuracy through similarity quantification. This embodiment establishes a correspondence between target mold information and mold images, enabling automated processes for image matching, parameter retrieval, and scanning detection. This meets the needs of automated, unattended detection and avoids errors and efficiency losses from manual parameter input.

[0037] like Figure 2 As shown, Figure 2 This is a schematic diagram of a mold quality inspection device provided in this embodiment. Exemplarily, the mold quality inspection device may include a control terminal device 110, a robot 120, and a mechanical turntable 130. The mold to be inspected can be placed on the mechanical turntable 130. The robot 120 can carry a laser scanning head 121. Additionally, a camera can be installed on the robot 120. Taking the inspection of a 500ml glass mold containing an image as an example, the control terminal device 110 can receive mold inspection commands input by the operator. After parsing the mold inspection commands, the control terminal device 110 sends a standby preparation signal to the robot 120 and the mechanical turntable 130, confirming that the device is in an executable state.

[0038] The control terminal device 110 sends an image acquisition instruction to the robot 120, controls the robot 120 to carry the industrial camera (camera) to a preset shooting position, shoots an image of the mold (i.e., a first mold image), and sends the image to the control terminal device 110. The control terminal device 110 first performs grayscale processing on the first mold image, then extracts contour type features through a Canny edge detection algorithm and a Harris corner detection algorithm, identifies that the mold as a whole is a cylindrical contour, has a key corner point a, and has a high edge smoothness level; then the control terminal device 110 extracts size features through pixel-to-actual-size conversion (preset 10 pixels = 1 mm): the maximum diameter of the mold is 65 mm, the height is 220 mm, and the bottle opening diameter is 32 mm; finally, the control terminal device 110 extracts pattern features in the image: the mold bottom contains a line and text combination pattern, the pattern position is located in the center area of the mold bottom, and the pattern has 12 key feature points.

[0039] According to the above extracted initial features, the control terminal device 110 can generate a label for the mold: contour type = cylindrical contour; size range = diameter 60-70 mm, height 210-230 mm; pattern = yes. The control terminal device 110 calls a mold database index module, takes “contour type = cylindrical contour; size range = diameter 60-70 mm, height 210-230 mm; pattern = yes” as a retrieval condition, and triggers database fuzzy matching. Through feature comparison with images in the database, three groups of mold images that meet the conditions are filtered out, forming a candidate mold image set, and the IDs of the three images are assumed to be STD-500ML-001, STD-500ML-003, and STD-500ML-007.

[0040] The control terminal device 110 identifies that the candidate mold image set contains three images, and enables similarity calculation. Using a feature point matching method, the similarity between the first mold image and each candidate image is calculated according to contour features 40% + size features 35% + pattern features 25%: the similarity with STD-500ML-001 is 96.2%, the similarity with STD-500ML-003 is 89.5%, and the similarity with STD-500ML-007 is 87.8%. The STD-500ML-001 image with the highest similarity is selected as the target mold image, and the target image ID and matching similarity are recorded.

[0041] The control terminal device 110 retrieves the corresponding target mold information from the mold database based on the image ID (such as STD-500ML-001) of the target mold image. Among them, the mold scanning control parameters are: robot movement path = linear path along the mold height Y axis (0mm-220mm or 220mm-0mm), movement speed 0.5mm / s, laser scanning frequency 500Hz and turntable rotation speed 5° / s; the mold type is 01 (glass mold containing image); the mold label information is "mold number STD-500ML-001, design version V2.1, applicable product 500ml glass bottle".

[0042] S102: Based on the mold scanning control parameters, the robot 120 scans the mold to be detected on the mechanical turntable 130 to obtain mold scanning data, and determines the mold volume value and the mold three-dimensional contour data based on the mold scanning data.

[0043] In this embodiment, the mold scanning data refers to the original data obtained by the robot 120 scanning the mold, which is used for subsequent calculation of volume and contour, for example, can include three-dimensional point cloud data, cross-sectional contour data, etc. The mold volume value refers to the volume quantization result of the mold cavity type, which is used to represent the capacity compliance, for example, can include positive volume, negative volume and total volume, etc. The mold three-dimensional contour data refers to the three-dimensional structure data reflecting the shape of the mold, which is used to represent the shape compliance, for example, can include key cross-sectional contour, surface feature point coordinates, etc.

[0044] The consideration behind this embodiment is that the mold scanning control parameters can standardize the scanning process, avoid path deviation or speed fluctuation caused by manual operation, ensure the consistency and accuracy of the scanning data, and realize the detection requirements of automation and high precision; the mold volume value is directly related to the capacity function of the mold forming product, and the mold three-dimensional contour data is directly related to the product shape adaptability, and the combination of the two can comprehensively cover the core quality detection dimension of the mold, and improve the detection efficiency and reliability.

[0045] For example, taking the detection of a glass bottle mold forming (mold type LM-C202410) as an example, the specific detection process can include:

[0046] The control terminal device 110 can retrieve the scanning control parameters corresponding to the mold from the database, including the robot movement path as a linear path along the mold height direction Y axis (starting point Y=0mm, end point Y=200mm), robot movement speed 0.5mm / s, laser scanning frequency 500Hz and mechanical turntable 130 rotation speed 5° / s.

[0047] The control terminal device 110 sends a rotation instruction to the mechanical turntable 130, the mechanical turntable 130 rotates at a constant speed of 5° / s, driving the mold to rotate smoothly; at the same time, the control terminal device 110 sends a scanning instruction to the robot 120, the robot 120 carries the laser scanning head 121 to move along the preset Y-axis path, the laser scanning head 121 continuously emits laser at a frequency of 500 Hz, collects three-dimensional coordinate data of the mold surface, generates mold scanning data containing three-dimensional point cloud data and bottle opening horizontal cross section and bottle body longitudinal cross section contour data, and the data is transmitted to the control terminal device 110 in real time and stored in 1.stl format.

[0048] The control terminal device 110 can call the measurement analysis software, and the mold scanning data in 1.stl format is input, the effective cavity type range of the mold is first circled by the cavity type area segmentation algorithm, the positive volume 506609.736, the negative volume 0.000 and the total volume 506609.736 are automatically calculated to determine the mold volume value; wherein the positive volume refers to the volume of the effective cavity type area of the mold to be detected, that is, the core space volume inside the mold which can be used to contain the glass molten liquid and finally determines the capacity of the formed glass product. The negative volume refers to the invalid volume loss in the effective cavity type area of the mold, that is, the space volume that cannot normally contain the glass molten liquid due to defects such as recesses, scratches and impurity residues on the cavity wall of the mold. The total volume refers to the actual available volume of the effective cavity type of the mold, and the calculation logic is total volume = positive volume - negative volume, and the total volume is the final index comprehensively reflecting the actual capacity of the mold cavity type.

[0049] The control terminal device 110 further extracts the three-dimensional contour features corresponding to the key dimensions of the mold bottle opening diameter and bottle body height by the edge extraction and contour fitting algorithm, and generates mold three-dimensional contour data containing key cross-sectional contour lines and surface feature point coordinates.

[0050] S103: If the mold type is a glass mold containing an image, image data of the mold is obtained based on the mold label information, and a mold quality detection result is determined based on the image data, the mold volume value and the mold three-dimensional contour data; if the mold type is a glass mold not containing an image, a mold quality detection result is determined based on the mold volume value and the mold three-dimensional contour data.

[0051] In this embodiment, the image data is data containing image mold surface pattern features, which is used to verify the pattern compliance, for example, can include pattern feature point coordinates and line contour data. The mold quality detection result is a mold quality judgment conclusion obtained based on the detection data, which is used to represent whether the mold is qualified.

[0052] The embodiment adopts differential detection logic according to the difference of mold types. The consideration behind the embodiment is that the pattern of the image-containing mold directly affects the appearance of the molded product. For example, if the bottom of the 500ml mold is engraved with a pattern, missing image detection can easily lead to false detection. The mold without image only needs to ensure that the volume and contour meet the capacity and assembly requirements. This logic avoids missing appearance defects for image-containing molds and avoids redundant detection for molds without images, balancing the comprehensiveness and efficiency of detection.

[0053] For example, to detect a glass mold containing an image and a glass mold without an image, the specific implementation process can include:

[0054] The control terminal device 110 extracts the mold type from the target mold information. The detection method for the glass mold containing the image can include:

[0055] The control terminal device 110 retrieves the corresponding image data from the database based on the mold unique number in the mold label information, including the bottom pattern feature point coordinates and line contour data. At the same time, the specific volume value 506609.736 and the three-dimensional contour data of the mold are called, and the image data and the pattern area in the three-dimensional contour data are compared to verify whether the feature point deviation is ≤0.02mm, whether the volume value is within the standard range, and whether all three are qualified. If so, the mold quality detection result is determined to be qualified.

[0056] The detection method for the glass mold without an image can include:

[0057] The control terminal device 110 directly calls the specific volume value and three-dimensional contour data of the mold to verify whether the volume value matches the nominal capacity of the glass bottle and whether the critical dimensions such as the mouth diameter and body height in the three-dimensional contour data are within the tolerance range. There is no need to process image data, and if both indicators are qualified, the mold quality detection result is determined to be qualified.

[0058] The control terminal device 110 generates a mold quality detection result report according to the detection result, including the determination conclusion, detection data and deviation details, for quality traceability.

[0059] From the above, it can be concluded that the embodiment automatically controls the robot 120 to obtain the first mold image through the control terminal device 110, and matches the mold scanning control parameters, mold type and mold label information from the database based on the image. In this way, the control terminal device 110 can automatically identify the mold to be detected and the related information. At the same time, the robot 120 and the mechanical turntable 130 cooperate with each other according to the preset parameters to scan, automatically complete data acquisition and analysis, shorten the time consumption of single mold detection, quickly respond to mold detection requirements, and improve the automation level and mold quality detection efficiency.

[0060] The embodiment is based on mold scanning control parameters to accurately control the scanning process, and combines the mold volume value determined by the mold scanning data and the three-dimensional profile data. Through this standardized step, the error can be reduced. At the same time, considering that part of the mold has a local image, the traditional detection method cannot effectively detect the image. Therefore, the embodiment divides the mold into two types: image-containing and image-free, and carries out differentiated quality detection processes for the two types of molds. According to the type of the mold, the detection logic is automatically switched, the overall quality detection precision of the mold containing the image is improved, and the application can be widely used for quality detection of various glass molds to meet diversified needs.

[0061] In summary, the embodiment realizes the automation and intelligentization of mold quality detection by controlling the communication connection between the terminal device 110 and the robot 120 and the mechanical turntable 130. The robot 120 can accurately obtain image data of the mold to be detected, and perform high-precision scanning based on mold scanning control parameters to obtain accurate mold volume values and mold three-dimensional profile data, thereby improving the detection efficiency and precision.

[0062] In an embodiment of the application, before responding to the mold detection instruction, an intelligent mold quality detection method further comprises:

[0063] In response to receiving the robot calibration instruction, the robot 120 is controlled to obtain the measured coordinates of the first calibration position, the measured coordinates of the second calibration position and the measured coordinates of the third calibration position based on the calibration path parameters; the first calibration position, the second calibration position and the third calibration position are mark points preset on the mechanical turntable 130 and not on the same straight line; the measured coordinates correspond to the robot coordinate system;

[0064] The absolute coordinates of the first calibration position, the absolute coordinates of the second calibration position and the absolute coordinates of the third calibration position are obtained; the absolute coordinates correspond to the world coordinate system; the first calibration position is on the Z-axis in the world coordinate system, the second calibration position has a different X-axis coordinate value in the world coordinate system from the first calibration position, and the third calibration position has a different Y-axis coordinate value in the world coordinate system from the first calibration position;

[0065] The first coordinate offset between the measured coordinates of the first calibration position and the absolute coordinates of the first calibration position is calculated, and the translation deviation of the robot 120 is determined based on the first coordinate offset;

[0066] A first measurement vector is determined based on the measured coordinates of the first calibration position and the measured coordinates of the second calibration position, and a second measurement vector is determined based on the measured coordinates of the first calibration position and the measured coordinates of the third calibration position;

[0067] A first absolute vector is determined based on the absolute coordinates of the first calibration position and the absolute coordinates of the second calibration position, and a second absolute vector is determined based on the absolute coordinates of the first calibration position and the absolute coordinates of the third calibration position.

[0068] calculating a first plane rotation deviation based on the projection of the first measurement vector on the first plane and the first absolute vector;

[0069] calculating a second plane rotation deviation based on the projection of the first measurement vector on the second plane and the first absolute vector;

[0070] calculating a third plane rotation deviation based on the projection of the second measurement vector on the third plane and the second absolute vector;

[0071] calibrating the position of the robot 120 based on the translation deviation, the first plane rotation deviation, the second plane rotation deviation and the third plane rotation deviation; the robot 120 after completing the position calibration is used to collect data of the mold placed on the mechanical turntable 130 for detection, and send the collected data to the control terminal device 110.

[0072] In the embodiment, the robot calibration instruction is a control signal for triggering the robot position calibration process, used to start the calibration action. The calibration path parameter is a path configuration data preset to control the robot 120 to move to a specified position, and the robot 120 can collect the coordinates of each calibration position at the specified position. The first calibration position, the second calibration position and the third calibration position are non-collinear marker points preset on the mechanical turntable 130, used to provide a calibration reference, for example, the marker points can be set as metal material (such as stainless steel), convex shape (such as spherical / cylindrical), size (such as diameter 5mm), and the position coordinates of each marker point are recorded. The measurement coordinates are the actual collected coordinates of the calibration positions in the robot coordinate system. The robot coordinate system is a three-dimensional coordinate system established with the robot 120 base as the reference, used to position the end position of the robot 120. The absolute coordinates are the standard coordinates of the calibration positions in the world coordinate system, used to provide a calibration reference value. The world coordinate system can be a global three-dimensional coordinate system of the mold quality detection device, used to unify the calibration positions and the robot 120 position reference.

[0073] The first coordinate offset is the difference between the first calibration position measurement coordinates and the absolute coordinates, used to calculate the translation deviation, for example, which can include X-axis offset value, Y-axis offset value and Z-axis offset value. The translation deviation is the offset of the overall position of the robot 120 relative to the standard position, used to represent the robot translation error, for example, which can include X-axis translation value, Y-axis translation value and Z-axis translation value. The first measurement vector is used to calculate the rotation deviation. The first absolute vector is used to provide a rotation deviation calculation reference. The first plane, the second plane and the third plane are X-Y plane, X-Z plane and Y-Z plane respectively, used to split the rotation deviation dimension. The first plane rotation deviation, the second plane rotation deviation and the third plane rotation deviation are the angle differences of the vector projections on the corresponding planes respectively, used to represent the robot rotation error.

[0074] The consideration behind this embodiment is that the robot 120 and the mechanical turntable 130 are active during work, and positional deviation will occur after continuous and long-term work. Calibrating the robot 120 before data acquisition can establish a unified and accurate spatial coordinate reference for the robot 120 and the mechanical turntable 130, eliminate positional deviation caused by equipment installation deviation, mechanical wear and tear, and environmental interference (such as temperature deformation), and ensure the accuracy of subsequent mold data acquisition (image acquisition, scanning detection).

[0075] For example, the robot 120 is prone to translational deviation (such as 0.05 mm overall deviation of the X-axis) or rotational deviation (such as 0.02° tilt around the Y-axis) due to base loosening and joint wear during long-term use. If not calibrated, the laser scanning head 121 will deviate from the preset scanning area, resulting in misplacement of mold scanning data. Calibration can ensure that the volume value and three-dimensional contour data of the mold collected by the robot 120 are consistent with the actual mold shape, avoiding false positives or false negatives of qualified molds. This embodiment has a robot coordinate system and a world coordinate system, and calibration can establish a precise mapping relationship between the two. For example, the three non-collinear calibration positions preset by the mechanical turntable 130 have fixed absolute coordinates. By calibrating the measurement coordinates collected by the robot 120 with the absolute coordinates, it is ensured that when the robot 120 and the mechanical turntable 130 cooperate in scanning, their motion trajectories strictly match, with no timing or positional deviation.

[0076] Considering that two points can only determine a straight line and cannot cover three-dimensional rotational error, this embodiment selects three non-collinear calibration positions. Three points can construct a complete three-dimensional coordinate system, ensuring that the deviation calculation covers all dimensions of translation and rotation, meeting the requirement of full-space precision guarantee in the purpose. Considering that the robot error is essentially a combination of overall positional deviation and attitude tilt, this embodiment splits the translational deviation and the three-plane rotational deviation. This classification calculation can accurately locate the error source and avoid the lack of precision caused by general correction.

[0077] For example, taking the robot calibration before detection of a 500 ml glass mold containing images as an example, the specific implementation process of calibration is as follows:

[0078] After the control terminal device 110 receives the robot calibration instruction, it triggers the calibration process. The control terminal device 110 sends a static instruction to the mechanical turntable 130. The control terminal device 110 retrieves the preset calibration path parameters from the system database and simultaneously sends a calibration instruction to the robot 120. The robot 120 moves along the preset calibration path with the laser scanning head 121. When it reaches the first calibration position, it stops for 1 s and collects its measurement coordinates in the robot coordinate system. The second calibration position and the third calibration position are sequentially completed, and the measurement coordinates are sent to the control terminal device 110.

[0079] The control terminal device 110 calls the absolute coordinates of the three calibration positions from the world coordinate system parameter library, and converts the absolute coordinates and the measurement coordinates into the same coordinate system. The first coordinate offset of the first calibration position is calculated after the conversion: the X-axis offset is 0.2 mm, the Y-axis offset is 0.1 mm, and the Z-axis offset is 0 mm. The control terminal device 110 determines the offset as the robot translation deviation. The translation deviation can reflect the position deviation caused by the loosening of the robot base and provide data support for subsequent translation correction.

[0080] The control terminal device 110 calculates the first measurement vector and the second measurement vector based on the measurement coordinates, and calculates the first absolute vector and the second absolute vector based on the absolute coordinates. Accordingly, the direction and distance relationship between the calibration positions can be characterized by vectors, providing direction dimension data for rotation deviation calculation and avoiding the defect that single coordinate comparison cannot reflect the rotation error.

[0081] The control terminal device 110 projects the first measurement vector onto the X-Y plane (Z component is 0), compares it with the first absolute vector, and calculates the first plane rotation deviation (such as 0.1°); projects the first measurement vector onto the X-Z plane (Y component is 0), compares it with the first absolute vector, and calculates the second plane rotation deviation -0.5°; projects the second measurement vector onto the Y-Z plane (X component is 0), compares it with the second absolute vector, and calculates the third plane rotation deviation 0.8°. This step can split the rotation error in three-dimensional space into three orthogonal planes, avoiding the cross interference of different dimension errors and ensuring the accuracy of rotation deviation calculation.

[0082] The control terminal device 110 sends the translation deviation and the three plane rotation deviations to the robot 120, and the robot 120 automatically adjusts the robot joint parameters through the translation deviation and the three plane rotation deviations. After adjustment, the robot 120 collects the measurement coordinates of the three calibration positions again until the deviation from the absolute coordinates is ≤0.01 mm (satisfying the threshold requirement), and then determines that the calibration is completed.

[0083] This embodiment eliminates the translation deviation and the rotation deviation of the robot 120 caused by long-term use, ensures the accurate alignment of the laser scanning head 121 to the mold scanning area, avoids the misplacement of the scanning data, makes the subsequent mold volume values and three-dimensional contour data consistent with the actual mold shape, and improves the accuracy of data acquisition. The calibration process compares the measurement coordinates and the absolute coordinates of the calibration positions, constructs the mapping relationship between the robot coordinate system and the world coordinate system, ensures that the motion trajectory strictly matches when the robot 120 and the mechanical turntable 130 cooperatively scan, and provides a stable cooperative reference for subsequent mold full-surface scanning and image acquisition.

[0084] In an embodiment of the present application, the mold scanning control parameters include a robot movement path, a robot movement speed, a laser scanning frequency, and a rotary table rotation speed; the robot 120 is controlled based on the mold scanning control parameters to scan the mold to be detected on the mechanical rotary table 130 to obtain mold scanning data, including:

[0085] The mechanical rotary table 130 is controlled based on the rotary table rotation speed to rotate;

[0086] The robot 120 is controlled based on the robot scanning path, the robot movement speed, and the laser scanning frequency to scan the mold to be detected on the mechanical rotary table 130 to obtain three-dimensional point cloud data, first cross-sectional profile data, and second cross-sectional profile data in sequence;

[0087] The three-dimensional point cloud data is a full-surface discrete three-dimensional coordinate set of the mold, the first cross-sectional profile data is a circular profile of a bottle mouth horizontal cross section of the mold, and the second cross-sectional profile data is a continuous curve of a bottle body longitudinal symmetric cross section of the mold; the three-dimensional point cloud data, the first cross-sectional profile data, and the second cross-sectional profile data are taken as the mold scanning data.

[0088] In the embodiment, the mold volume value and the mold three-dimensional profile data are determined based on the mold scanning data, specifically including: a target coordinate system is constructed based on the first cross-sectional profile data and the second cross-sectional profile data; a mold three-dimensional model is generated based on the target coordinate system, the three-dimensional point cloud data, the first cross-sectional profile data, and the second cross-sectional profile data; and the mold volume value and the mold three-dimensional profile data are determined based on the mold three-dimensional model.

[0089] In the embodiment, the target coordinate system is constructed based on the first cross-sectional profile data and the second cross-sectional profile data, specifically including: a bottle mouth center position is determined based on the first cross-sectional profile data; the bottle mouth center position is taken as an origin of the target coordinate system; a bottle body center line is extracted based on the second cross-sectional profile data, an upper end point of the bottle body center line is a bottle mouth, a lower end point of the bottle body center line is a bottle bottom, and a direction from the upper end point to the lower end point of the bottle body center line is taken as a Y-axis positive direction of the target coordinate system; and a right direction perpendicular to the Y-axis is taken as an X-axis positive direction of the target coordinate system.

[0090] In the embodiment, the target coordinate system is a local three-dimensional coordinate system constructed based on the mold key section profile, which is used to unify the coordinate reference of the mold scanning data and ensure the coordinate consistency of subsequent modeling and data calculation. The origin and axis system direction of the target coordinate system are determined based on the characteristics of the mold itself rather than relying on the equipment reference. The bottle body centerline is the centerline of the longitudinally symmetrical section extracted from the second section profile data, which runs through the bottle body and represents the longitudinal symmetry reference of the mold. The upper end point corresponds to the bottle mouth, and the lower end point corresponds to the bottle bottom. The mold three-dimensional model is a digital model of the mold generated based on the target coordinate system, which integrates three-dimensional point cloud data and first and second section profile data. The model intuitively reflects the geometric shape of the mold surface and is the basis for calculating the mold volume and extracting three-dimensional profile data. The circular profile is a cluster of discrete points that constitute the circular profile of the bottle mouth transverse section, which collectively represents the transverse geometric shape of the bottle mouth and provides data support for determining the position of the bottle mouth center. The continuous curve is a cluster of points that constitute the continuous curve of the longitudinally symmetrical section of the bottle body. The points are arranged continuously and accurately reflect the longitudinal bending shape of the bottle body, providing data basis for extracting the bottle body centerline.

[0091] In the embodiment, the first section profile (bottle mouth transverse) and the second section profile (bottle body longitudinal) are selected to construct the target coordinate system. The bottle mouth center and the bottle body centerline are inherent symmetrical features of the glass mold, which are stable in position and easy to extract from the scanning data. This can avoid the cumulative error when using the equipment coordinate system as the reference, ensure the strong correlation between the coordinate system and the mold shape, and provide a reliable reference for subsequent modeling. In the embodiment, the turntable is controlled to rotate first and then the robot 120 scans. This is because the rotation of the turntable can expose the circumferential surface of the mold in turn, combined with the movement of the robot 120 along the height direction of the bottle body, which can realize the full-surface coverage of the mold without blind area. In the embodiment, the three-dimensional point cloud and the section profile data are combined to generate a three-dimensional model. The three-dimensional point cloud can provide full-surface details, and the section profile data can ensure the accuracy of the key shape. The two complement each other to balance the modeling accuracy and efficiency, avoiding redundant calculation when only using point cloud modeling or missing details when only using profile modeling. In the embodiment, the three types of data are divided into scanning data because different data correspond to different functions: two types of section profiles are used to define the reference, and the point cloud is used to complete the surface. Classified processing can improve data processing efficiency.

[0092] In the embodiment, the robot scanning path includes a first scanning path, a second scanning path, a third scanning path, and a fourth scanning path; the laser scanning frequency includes a first laser scanning frequency and a second laser scanning frequency, and the first laser scanning frequency is greater than the second laser scanning frequency.

[0093] Based on the robot scanning path, the robot movement speed, and the laser scanning frequency, the robot 120 scans the mold to be detected on the mechanical turntable 130 to obtain three-dimensional point cloud data, first section profile data, and second section profile data in turn.

[0094] controlling the robot 120 to scan the mold to be detected on the mechanical turntable 130 based on the first scanning path, the robot moving speed and the first laser scanning frequency, to obtain three-dimensional point cloud data; the starting point of the first scanning path corresponds to a bottle bottom data collection point of the mold to be detected, and the ending point of the first scanning path corresponds to a first bottle mouth data collection point of the mold to be detected; the first bottle mouth data collection point is at the same horizontal plane as the bottle mouth of the mold to be detected;

[0095] controlling the robot 120 to move to a second bottle mouth data collection point of the mold to be detected based on the second scanning path and the robot moving speed; and controlling the robot 120 to scan the bottle mouth of the mold to be detected on the mechanical turntable 130 based on the second laser scanning frequency, to obtain first cross-sectional profile data; the starting point of the second scanning path is the first bottle mouth data collection point, and the ending point of the second scanning path is the second bottle mouth data collection point, which is located above the bottle mouth of the mold to be detected;

[0096] controlling the mechanical turntable 130 to stop rotating;

[0097] controlling the robot 120 to move to the first bottle mouth data collection point of the mold to be detected based on the third scanning path and the robot moving speed; the starting point of the third scanning path is the second bottle mouth data collection point, and the ending point of the third scanning path is the first bottle mouth data collection point;

[0098] controlling the robot 120 to scan the mold to be detected on the mechanical turntable 130 based on the fourth scanning path, the robot moving speed and the second laser scanning frequency, to obtain second cross-sectional profile data; the starting point of the fourth scanning path is the first bottle mouth data collection point, and the ending point of the fourth scanning path is the bottle bottom data collection point.

[0099] Exemplary, the control terminal device 110 receives the mold detection instruction, and retrieves the corresponding mold scanning control parameters of the model from the mold database: the rotating speed of the rotating table is set to 5° / s; the moving speed of the robot is set to 0.5 mm / s; the first laser scanning frequency is set to 500 Hz (for high-frequency sampling of three-dimensional point cloud), and the second laser scanning frequency is set to 200 Hz (for high-frequency sampling of cross section profile). At the same time, the control terminal device 110 retrieves the scanning path parameters of the robot 120: the starting point of the first scanning path is the bottle bottom data collection point, corresponding to the center of the mold bottle bottom, and the preset Y=210 mm, and the endpoint is the first bottle mouth data collection point, which is the same horizontal plane as the bottle mouth, and Y=0 mm; the starting point of the second scanning path is the first bottle mouth data collection point, Y=0 mm, and the endpoint is the second bottle mouth data collection point, which is 10 cm above the bottle mouth, Y=-10 cm; the starting point of the third scanning path is the second bottle mouth data collection point Y=-10 cm, and the endpoint is the first bottle mouth data collection point, Y=0 mm; the starting point of the fourth scanning path is the first bottle mouth data collection point, and the endpoint is the bottle bottom data collection point.

[0100] The control terminal device 110 sends a rotating instruction to the mechanical rotating table 130, and the rotating table drives the mold to rotate at a uniform speed of 5° / s. The servo motion mechanism of the mechanical platform ensures that the shaking amplitude of the rotating table is less than or equal to 0.001 mm, avoids the offset of the mold, and makes the circumferential surface of the mold enter the scanning range of the laser scanning head 121 in turn.

[0101] Three-dimensional point cloud data collection: the control terminal device 110 sends a first scanning instruction to the robot 120, and the robot 120 carries the laser scanning head 121 to move along the first scanning path: from the bottle bottom data collection point to the first bottle mouth data collection point at a speed of 0.5 mm / s, and the laser scanning head 121 continuously samples at a first laser scanning frequency (500 Hz), collects a three-dimensional coordinate point every 0.002 seconds, covers the whole surface of the bottle bottom, bottle body and bottle mouth, accumulatively collects about 1.1 million discrete points, forms three-dimensional point cloud data, and stores it in the.stl format, which ensures the restoration of the details such as the corner arc of the mold and the bottle bottom groove.

[0102] After the three-dimensional point cloud data collection is completed, the control terminal device 110 sends a second scanning instruction, and the robot 120 moves along the second scanning path: from the first bottle mouth data collection point to the second bottle mouth data collection point at a speed of 0.5 mm / s, and in this process, the laser scanning head 121 stops sampling, and only through the servo positioning of the robot joint, the center of the scanning head is aligned with the center of the horizontal cross section of the bottle mouth.

[0103] First cross-section profile data collection: After the robot 120 reaches the second bottle mouth data collection point, the control terminal device 110 instructs the laser scanning head 121 to start scanning at a second laser scanning frequency (200 Hz), and the turntable still rotates at 5° / s, which drives the bottle mouth transverse cross-section to be completely exposed. A coordinate point is collected every 0.005 seconds, and a total of 300 points are collected to form the first cross-section profile data of the bottle mouth transverse circular profile, which is stored in.csv format.

[0104] Mechanical turntable 130 stops rotating: After the first cross-section profile collection is completed, the control terminal device 110 sends a stop instruction, and the turntable is quickly stationary through a servo brake mechanism, with an angle deviation of ≤0.01° after being stationary, so as to avoid mold deviation during subsequent longitudinal cross-section collection.

[0105] Robot 120 moves back to first bottle mouth data collection point: The control terminal device 110 sends a third scanning instruction, and the robot 120 moves along a third scanning path: moves back to the first bottle mouth data collection point from the second bottle mouth data collection point at a speed of 0.5 mm / s, and ensures that the scanning direction coincides with the longitudinal symmetrical cross-section of the bottle body through encoder feedback.

[0106] Second cross-section profile data collection: After the robot 120 reaches the first bottle mouth data collection point, the control terminal device 110 sends a fourth scanning instruction, and the robot 120 moves along a fourth scanning path: moves from the first bottle mouth data collection point to the bottle bottom data collection point at a speed of 0.5 mm / s, and the laser scanning head 121 samples at the second laser scanning frequency to form the second cross-section profile data of the longitudinal continuous curve of the bottle body, which is stored in.csv format. As shown in Figure 3 Figure 3 The second cross-section profile is obtained for a plurality of molds during detection.

[0107] The robot 120 transmits the three types of data obtained in real time to the control terminal device 110. The control terminal device 110 performs circular fitting on the first cross-section profile data, calculates the bottle mouth center position (set as X=0, Y=0, Z=0) by analyzing the coordinate distribution of the circular profile point set, and sets this position as the origin of the target coordinate system; the control terminal device 110 performs symmetry analysis on the second cross-section profile data, extracts the bottle body center line that penetrates the bottle body, determines the upper end point (bottle mouth center) and lower end point (bottle bottom center) on the center line, and sets the direction of the center line from the upper end point to the lower end point as the positive direction of the Y axis; the control terminal device 110 sets the direction perpendicular to the Y axis and pointing to the right side of the bottle body as the positive direction of the X axis, and determines the Z axis direction according to the right-hand rule, to complete the construction of the target coordinate system.

[0108] ​The terminal device 110 is controlled to map the three-dimensional point cloud data to the target coordinate system based on the target coordinate system, while integrating the first cross-sectional profile data and the second cross-sectional profile data, to finally generate a three-dimensional mold model. The terminal device 110 circumscribes a mold effective cavity region based on the three-dimensional mold model, automatically calculates a cavity positive volume (set as 499750 The terminal device 110 extracts key profile features of the mold full surface from the three-dimensional mold model to form three-dimensional profile data of the mold, and completes data determination.

[0109] The embodiment constructs a target coordinate system based on key cross sections of the mold itself, avoids cumulative deviation of the device coordinate system, and avoids quality misjudgment caused by reference deviation. The embodiment cooperates with the turntable rotation and the robot 120 movement, combines laser high-frequency sampling, can cover the mold circumferential surface and the longitudinal surface, avoids blind areas such as bottle body rotation angle and bottle bottom inner side, and improves measurement efficiency and accuracy.

[0110] In an embodiment of the present application, the mold quality detection result is determined based on the image data, the mold volume value and the three-dimensional profile data of the mold, including: obtaining mold plate image data corresponding to a target mold image, a standard volume value and standard three-dimensional profile data; calculating image similarity of the image data and the mold plate image data; calculating volume difference of the mold volume value and the standard volume value; calculating three-dimensional profile similarity of the three-dimensional profile data of the mold and the standard three-dimensional profile data; and determining the mold quality detection result based on the image similarity, the volume difference and the three-dimensional profile similarity.

[0111] In the embodiment, the mold plate image data is standard image data corresponding to the target mold image, is a reference for image data comparison, and is used to verify mold appearance feature compliance. The standard volume value is a standard volume determined in the target mold design stage, is a reference for mold volume value comparison, reflects a preset capacity of a mold formed product, and is used to verify mold function compliance. The standard three-dimensional profile data is standard shape data of the target mold, is a reference for three-dimensional profile data comparison of the mold, and is used to verify mold shape compliance. The image similarity is a quantitative index of the similarity degree of the image data and the mold plate image data, for example, can be calculated by comparing pattern feature points, lines and the like, and represents the compliance degree of mold surface patterns and the like. The volume difference is a difference between the mold volume value and the standard volume value, directly reflects the deviation of the mold capacity and the standard value, and represents the compliance degree of the mold capacity function. The three-dimensional profile similarity is a quantitative index of the similarity degree of the three-dimensional profile data of the mold and the standard three-dimensional profile data, for example, can be calculated by comparing key profile features, dimensions and the like, and represents the compliance degree of the mold shape.

[0112] ​Exemplarily, the control terminal device 110 retrieves the standard data corresponding to the mold of the model from the mold database: the template image data (including the standard pattern of the bottle bottom logo), the standard volume value (500ml, corresponding to the volume 500000 ), and the standard three-dimensional contour data (including the standard diameter of the bottle mouth 32.0mm, the standard height of the bottle body 220.0mm, and the maximum diameter of the bottle body 65.0mm). The control terminal device 110 calls the image comparison module to align the actual pattern (image data) of the bottle bottom of the mold to be detected with the template image data, calculates the image similarity through image recognition, and obtains the result as 98.6%. The control terminal device 110 retrieves the volume value of the mold to be detected (499750 ) determined in the foregoing, and directly calculates the difference value with the standard volume value (500000 ), and obtains the volume difference value as -250 , and the absolute value as 250 .

[0113] The control terminal device 110 maps the three-dimensional contour data of the mold to be detected to the standard three-dimensional contour data to the same target coordinate system, compares the key size deviations such as the diameter of the bottle mouth (actual 32.03mm, standard 32.0mm) and the height of the bottle body (actual 219.95mm, standard 220.0mm), and calculates the three-dimensional contour similarity in combination with the contour edge smoothness, and obtains the result as 99.3%.

[0114] The control terminal device 110 presets the quality judgment threshold: the image similarity is greater than or equal to 95%, the absolute value of the volume difference is less than or equal to 500 , and the three-dimensional contour similarity is greater than or equal to 98%. The three indexes calculated are compared with the threshold: the image similarity 98.6% is greater than or equal to 95%, the absolute value of the volume difference 250 is less than or equal to 500 , and the three-dimensional contour similarity 99.3% is greater than or equal to 98%. All indexes meet the requirements, and it is judged that the mold quality detection result is qualified, and a detection report containing the values of the indexes is generated.

[0115] The embodiment covers the mold quality requirements from the three dimensions of image, volume, and three-dimensional contour, and avoids the missed judgment caused by detecting only a single dimension. The embodiment quantifies the detection results through the image similarity, the volume difference, and the three-dimensional contour similarity, and improves the credibility of the detection results.

[0116] Corresponding to the foregoing embodiment, an intelligent mold quality detection method, Figure 4 is a structural block diagram of an intelligent mold quality detection system provided by an embodiment of the present application. Only parts related to the embodiments of the present application are shown for ease of description. For reference Figure 4The intelligent mold quality detection system 20 is applied to a control terminal device, and the control terminal device is in communication connection with a robot and a mechanical turntable; the mechanical turntable is used for placing a mold to be detected; the robot is used for collecting data of the mold to be detected placed on the mechanical turntable and sending the collected data to the control terminal device; the control terminal device is used for controlling working states of the robot and the mechanical turntable and obtaining a mold quality detection result based on the data sent by the robot; and the intelligent mold quality detection system 20 comprises a mold image recognition module 21, a quality detection module 22 and a quality analysis module 23.

[0117] The mold image recognition module 21 is configured to, in response to receiving a mold detection instruction, control the robot to acquire a first mold image of the mold to be detected on the mechanical turntable; in response to receiving the first mold image sent by the robot, obtain a target mold image and target mold information corresponding to the target mold image from a mold database based on the first mold image; the target mold image is an image matched with the first mold image, and the target mold information comprises mold scanning control parameters, a mold type and mold label information.

[0118] The quality detection module 22 is configured to control the robot to scan the mold to be detected on the mechanical turntable based on the mold scanning control parameters to obtain mold scanning data, and determine a mold volume value and mold three-dimensional profile data based on the mold scanning data.

[0119] The quality analysis module 23 is configured to, if the mold type is a glass mold containing an image, acquire image data of the mold based on the mold label information, and determine a mold quality detection result based on the image data, the mold volume value and the mold three-dimensional profile data; if the mold type is a glass mold not containing an image, determine the mold quality detection result based on the mold volume value and the mold three-dimensional profile data.

[0120] In an embodiment of the present application, the intelligent mold quality detection system 20 further comprises:

[0121] The robot calibration module is configured to:

[0122] In response to receiving a robot calibration instruction, control the robot to acquire a measured coordinate of a first calibration position, a measured coordinate of a second calibration position and a measured coordinate of a third calibration position based on calibration path parameters; the first calibration position, the second calibration position and the third calibration position are marking points preset on the mechanical turntable and not on the same straight line; and the measured coordinates all correspond to a robot coordinate system.

[0123] obtaining absolute coordinates of the first calibration position, absolute coordinates of the second calibration position and absolute coordinates of the third calibration position; the absolute coordinates correspond to a world coordinate system; the first calibration position is on a Z axis in the world coordinate system, the second calibration position is different from the first calibration position in an X axis coordinate value in the world coordinate system, and the third calibration position is different from the first calibration position in a Y axis coordinate value in the world coordinate system;

[0124] calculating a first coordinate offset between the measured coordinates of the first calibration position and the absolute coordinates of the first calibration position, and determining a translation deviation of the robot based on the first coordinate offset;

[0125] determining a first measurement vector based on the measured coordinates of the first calibration position and the measured coordinates of the second calibration position, and determining a second measurement vector based on the measured coordinates of the first calibration position and the measured coordinates of the third calibration position;

[0126] determining a first absolute vector based on the absolute coordinates of the first calibration position and the absolute coordinates of the second calibration position, and determining a second absolute vector based on the absolute coordinates of the first calibration position and the absolute coordinates of the third calibration position;

[0127] calculating a first plane rotation deviation based on a projection of the first measurement vector on a first plane and the first absolute vector;

[0128] calculating a second plane rotation deviation based on a projection of the first measurement vector on a second plane and the first absolute vector;

[0129] calculating a third plane rotation deviation based on a projection of the second measurement vector on a third plane and the second absolute vector;

[0130] calibrating the position of the robot based on the translation deviation, the first plane rotation deviation, the second plane rotation deviation and the third plane rotation deviation; the robot after completing the position calibration is used for data acquisition of a mold to be detected placed on a mechanical turntable, and the acquired data is sent to a control terminal device.

[0131] In an embodiment of the present application, the mold image recognition module 21 is specifically configured to: perform initial feature extraction on the first mold image to obtain contour type features, size features and pattern features; match a candidate mold image set from a mold database based on the contour type features, the size features and the pattern features; if the candidate mold image set includes one mold image, the mold image is taken as a target mold image; if the candidate mold image set includes at least two mold images, the similarity of the first mold image and each mold image in the candidate mold image set is calculated respectively, and the mold image with the highest similarity to the first mold image is taken as the target mold image.

[0132] In an embodiment of the present application, the mold scanning control parameters include a robot movement path, a robot movement speed, a laser scanning frequency and a rotary table rotation speed; the quality detection module 22 is specifically configured to: control the rotation of the mechanical rotary table based on the rotary table rotation speed; control the robot to scan the mold to be detected on the mechanical rotary table based on the robot scanning path, the robot movement speed and the laser scanning frequency, and sequentially obtain three-dimensional point cloud data, first cross-sectional profile data and second cross-sectional profile data; the three-dimensional point cloud data is a discrete three-dimensional coordinate set of the full surface of the mold, the first cross-sectional profile data is a circular profile of the bottle mouth horizontal cross section of the mold, and the second cross-sectional profile data is a continuous curve of the bottle body vertical symmetric cross section of the mold; and the three-dimensional point cloud data, the first cross-sectional profile data and the second cross-sectional profile data are taken as mold scanning data.

[0133] In an embodiment of the present application, the quality detection module 22 is specifically further configured to: construct a target coordinate system based on the first cross-sectional profile data and the second cross-sectional profile data; generate a mold three-dimensional model based on the target coordinate system, the three-dimensional point cloud data, the first cross-sectional profile data and the second cross-sectional profile data; and determine a mold volume value and mold three-dimensional profile data based on the mold three-dimensional model.

[0134] In an embodiment of the present application, the quality detection module 22 is specifically further configured to: determine a bottle mouth center position based on the first cross-sectional profile data; take the bottle mouth center position as the origin of the target coordinate system; extract a bottle body center line based on the second cross-sectional profile data, the upper end point of the bottle body center line being the bottle mouth and the lower end point being the bottle bottom, and taking the direction from the upper end point to the lower end point of the bottle body center line as the positive direction of the Y axis of the target coordinate system; and taking the right direction perpendicular to the Y axis as the positive direction of the X axis of the target coordinate system.

[0135] In an embodiment of the present application, the quality analysis module 23 is specifically configured to: obtain template image data corresponding to the target mold image, a standard volume value and standard three-dimensional profile data; calculate an image similarity between the image data and the template image data; calculate a volume difference value between the mold volume value and the standard volume value; calculate a three-dimensional profile similarity between the mold three-dimensional profile data and the standard three-dimensional profile data; and determine a mold quality detection result based on the image similarity, the volume difference value and the three-dimensional profile similarity.

[0136] Referring to Figure 5 , Figure 5 The control terminal device provided in an embodiment of the present application is shown in a schematic block diagram. As shown in Figure 5The control terminal device 300 in the embodiment shown can include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The above-mentioned processors 301, input devices 302, output devices 303, and memories 304 complete communication with each other through a communication bus 305. The memory 304 is configured to store a computer program, and the computer program includes program instructions. The processor 301 is configured to execute the program instructions stored in the memory 304. The processor 301 is configured to invoke the program instructions to execute the functions of the modules in the above-mentioned system embodiments, for example Figure 4 The functions of the mold image recognition module 21, the quality detection module 22, and the quality analysis module 23 shown are described.

[0137] It should be understood that, in the embodiments of the present application, the processor 301 can be a central processing unit (CPU), and the processor can also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0138] The input device 302 can include a touchpad, a fingerprint collection sensor (for collecting fingerprint information and direction information of the fingerprint of a user), a microphone, etc., and the output device 303 can include a display (LCD, etc.), a speaker, etc.

[0139] The memory 304 can include read-only memory and random access memory, and provide instructions and data for the processor 301. A portion of the memory 304 can also include non-volatile random access memory. For example, the memory 304 can also store information of a mold type.

[0140] In specific implementations, the processor 301, the input device 302, and the output device 303 described in the embodiments of the present application can execute the implementation manners described in the embodiments of the intelligent mold quality detection method provided by the embodiments of the present application, and can also execute the implementation manners of the control terminal device 300 described in the embodiments of the present application, which will not be described here.

[0141] In another embodiment of the present application, a computer readable storage medium is provided, which stores a computer program. The computer program includes program instructions, which, when executed by a processor, implement all or part of the processes of the above-mentioned embodiment methods. The computer program can also instruct related hardware to complete the implementation. The computer program can be stored in a computer readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0142] The computer readable storage medium can be an internal storage unit of the control terminal device, such as a hard disk or a memory of the control terminal device. The computer readable storage medium can also be an external storage device of the control terminal device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the computer readable storage medium can include both the internal storage unit and the external storage device of the control terminal device. The computer readable storage medium is used to store the computer program and other programs and data required by the control terminal device. The computer readable storage medium can also be used to temporarily store data that has been output or will be output.

[0143] Those skilled in the art can appreciate that the modules / units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the above description. Whether the functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0144] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the control terminal device and the units described above can refer to the corresponding processes in the above-mentioned method embodiments, which will not be described here.

[0145] In several embodiments provided in the present application, it should be understood that the disclosed control terminal device and method can be implemented in other manners. For example, the division of the system embodiments described above is merely an example, and the division of the modules / units can be different, for example, a plurality of modules / units can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the modules / units shown or discussed can be indirect coupling or communication connection through some interfaces or modules / units, and can be electrical, mechanical or other forms of connection.

[0146] The modules / units described as separate components can or can not be physically separate, and the components shown as modules / units can or can not be physical modules / units, i.e., can be located in one place or distributed on a plurality of network modules / units. Part or all of the modules / units can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0147] In addition, the functional modules / units in each embodiment of the present application can be integrated into a processing module / unit, or each module / unit can exist physically, or two or more modules / units can be integrated into one module / unit. The integrated module / unit can be realized in the form of hardware or in the form of a software functional module / unit.

[0148] The above is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any skilled person in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An intelligent mold quality inspection method, characterized in that, The method is executed by a control terminal device, which is communicatively connected to both a robot and a mechanical turntable. The mechanical turntable is used to place the mold to be inspected, and the robot is used to collect data from the mold placed on the mechanical turntable and send the collected data to the control terminal device. The control terminal device is used to control the working status of the robot and the mechanical turntable, and to obtain the mold quality inspection result based on the data sent by the robot; The method includes: In response to receiving a mold inspection command, the robot is controlled to acquire a first mold image of the mold to be inspected on the mechanical turntable; In response to receiving a first mold image sent by the robot, a target mold image and target mold information corresponding to the target mold image are obtained from the mold database based on the first mold image; the target mold image is an image that matches the first mold image, and the target mold information includes mold scanning control parameters, mold type, and mold label information. Based on the mold scanning control parameters, the robot is controlled to scan the mold to be inspected on the mechanical turntable to obtain mold scanning data. Based on the mold scanning data, the mold volume value and the mold three-dimensional contour data are determined. If the mold type is a glass mold containing an image, then the image data of the mold is obtained based on the mold label information, and the mold quality inspection result is determined based on the image data, the mold volume value, and the mold three-dimensional contour data; if the mold type is a glass mold without an image, then the mold quality inspection result is determined based on the mold volume value and the mold three-dimensional contour data.

2. The intelligent mold quality inspection method as described in claim 1, characterized in that, Prior to receiving the mold inspection command, the method further includes: In response to receiving a robot calibration command, the robot is controlled to acquire the measurement coordinates of the first, second, and third calibration points based on calibration path parameters; the first, second, and third calibration points are pre-set markers on the mechanical turntable that are not on the same straight line; the measurement coordinates all correspond to the robot coordinate system; Obtain the absolute coordinates of the first, second, and third coordinates; all absolute coordinates correspond to the world coordinate system; the first coordinate is located on the Z-axis in the world coordinate system, the second coordinate is different from the first coordinate on the X-axis in the world coordinate system, and the third coordinate is different from the first coordinate on the Y-axis in the world coordinate system. Calculate the first coordinate offset between the measured coordinates of the first target positioning and the absolute coordinates of the first target positioning, and determine the translational deviation of the robot based on the first coordinate offset; A first measurement vector is determined based on the measurement coordinates of the first and second locators, and a second measurement vector is determined based on the measurement coordinates of the first and third locators. A first absolute vector is determined based on the absolute coordinates of the first and second locators, and a second absolute vector is determined based on the absolute coordinates of the first and third locators. The rotational deviation of the first plane is calculated based on the projection of the first measurement vector onto the first plane and the first absolute vector. The rotational deviation of the second plane is calculated based on the projection of the first measurement vector onto the second plane and the first absolute vector. The rotational deviation of the third plane is calculated based on the projection of the second measurement vector onto the third plane and the second absolute vector. The position of the robot is calibrated based on the translational deviation, the first plane rotation deviation, the second plane rotation deviation, and the third plane rotation deviation; after the position calibration is completed, the robot is used to collect data on the mold to be tested placed on the mechanical turntable and send the collected data to the control terminal device.

3. The intelligent mold quality inspection method as described in claim 1, characterized in that, The step of obtaining the target mold image from the mold database based on the first mold image includes: Initial feature extraction is performed on the first mold image to obtain contour type features, size features, and pattern features; Based on the contour type features, size features, and pattern features, a set of candidate mold images is obtained by matching from the mold database; If the candidate mold image set includes a mold image, then that mold image is taken as the target mold image; If the candidate mold image set includes at least two mold images, then the similarity between the first mold image and each mold image in the candidate mold image set is calculated respectively, and the mold image with the highest similarity to the first mold image is taken as the target mold image.

4. The intelligent mold quality inspection method as described in claim 1, characterized in that, The mold scanning control parameters include robot movement path, robot movement speed, laser scanning frequency, and turntable rotation speed; The process involves controlling the robot to scan the mold to be inspected on the mechanical turntable based on the mold scanning control parameters, thereby obtaining mold scanning data, including: The rotation of the mechanical turntable is controlled based on the turntable's rotation speed; Based on the robot scanning path, robot moving speed and laser scanning frequency, the robot is controlled to scan the mold to be inspected on the mechanical turntable, and three-dimensional point cloud data, first cross-sectional contour data and second cross-sectional contour data are obtained in sequence. The three-dimensional point cloud data is a set of discrete three-dimensional coordinates of the entire surface of the mold. The first cross-sectional contour data is the circular contour of the transverse cross-section of the bottle mouth of the mold. The second cross-sectional contour data is the continuous curve of the longitudinal symmetrical cross-section of the bottle body of the mold. The three-dimensional point cloud data, the first cross-sectional contour data and the second cross-sectional contour data are used as the mold scanning data.

5. The intelligent mold quality inspection method as described in claim 4, characterized in that, The process of determining the mold volume value and the mold three-dimensional contour data based on the mold scanning data includes: A target coordinate system is constructed based on the first cross-sectional contour data and the second cross-sectional contour data; A 3D model of the mold is generated based on the target coordinate system, the 3D point cloud data, the first cross-sectional contour data, and the second cross-sectional contour data. The mold volume value and mold three-dimensional contour data are determined based on the mold three-dimensional model.

6. The intelligent mold quality inspection method as described in claim 5, characterized in that, The construction of the target coordinate system based on the first cross-sectional contour data and the second cross-sectional contour data includes: The center position of the bottle mouth is determined based on the first cross-sectional contour data; the center position of the bottle mouth is taken as the origin of the target coordinate system; The bottle body centerline is extracted based on the second cross-sectional contour data. The upper end of the bottle body centerline is the bottle mouth, and the lower end is the bottle bottom. The direction from the upper end to the lower end of the bottle body centerline is taken as the positive Y-axis direction of the target coordinate system; the direction to the right perpendicular to the Y-axis is taken as the positive X-axis direction of the target coordinate system.

7. The intelligent mold quality inspection method as described in claim 1, characterized in that, The process of determining the mold quality inspection result based on the image data, the mold volume value, and the mold three-dimensional contour data includes: Acquire the template image data, standard volume value, and standard 3D contour data corresponding to the target mold image; Calculate the image similarity between the image data and the template image data; Calculate the volume difference between the mold volume value and the standard volume value; Calculate the three-dimensional contour similarity between the mold's three-dimensional contour data and the standard three-dimensional contour data; The mold quality inspection result is determined based on the image similarity, the volume difference, and the three-dimensional contour similarity.

8. An intelligent mold quality inspection system, characterized in that, The system is applied in a control terminal device, which is communicatively connected to a robot and a mechanical turntable. The mechanical turntable is used to place the mold to be inspected, and the robot is used to collect data from the mold placed on the mechanical turntable and send the collected data to the control terminal device. The control terminal device is used to control the working status of the robot and the mechanical turntable, and to obtain the mold quality inspection result based on the data sent by the robot; The system includes: A mold image recognition module is used to, in response to receiving a mold detection command, control the robot to acquire a first mold image of the mold to be inspected on the mechanical turntable; and in response to receiving the first mold image sent by the robot, obtain a target mold image and target mold information corresponding to the target mold image from a mold database based on the first mold image; the target mold image is an image that matches the first mold image, and the target mold information includes mold scanning control parameters, mold type, and mold label information; The quality inspection module is used to control the robot to scan the mold to be inspected on the mechanical turntable based on the mold scanning control parameters to obtain mold scanning data, and to determine the mold volume value and mold three-dimensional contour data based on the mold scanning data. The quality analysis module is used to: if the mold type is a glass mold containing an image, obtain image data of the mold based on the mold label information, and determine the mold quality inspection result based on the image data, the mold volume value, and the mold three-dimensional contour data; if the mold type is a glass mold without an image, determine the mold quality inspection result based on the mold volume value and the mold three-dimensional contour data.

9. A control terminal device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images