Intelligent mold management and control system based on equipment data visualization

Abstract

The invention discloses an intelligent mold management and control system based on equipment data visualization, and relates to the technical field of mold production, the intelligent mold management and control system comprises a product identification module, a detection analysis module, a model establishment module and an output feedback module, the product identification module photographs a product needing to be produced through a machine vision device, obtains a product image and collects and identifies product parameters; the product parameters comprise a cavity, a core, a production mode, size, shape, core pulling times and depth; the detection and analysis module detects and analyzes the product through the product parameters, obtains 3D point cloud information and constructs product information; the problems that in an existing scheme, data standards cannot be unified, the process parameters of the mold cannot be adjusted in time, and the field process cannot be adjusted in time can be solved, meanwhile, the data standards of the process parameters are unified, and data sharing of the process parameters is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of mold analysis, and in particular to an intelligent mold management and control system based on equipment data visualization. Background Art

[0002] At present, the adjustment and simulation of mold process parameters are mainly based on moldflow software. After the software simulates the process parameters, the simulation results of the parts are checked for defects before production. Most of the process parameters on site are in the hands of experienced workers. When new employees join the company, they cannot quickly get started. The accumulated process parameters of actual mold production cannot be accumulated in real time, which increases the company's employment costs. When old employees leave, new employees cannot be quickly trained. There is currently no relevant solution. Therefore, there is an urgent need for a mold intelligent management and control system based on equipment data visualization. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a mold intelligent management and control system based on equipment data visualization, which is used to solve the problems in the existing solutions that data standards cannot be unified, the process parameters of the mold cannot be adjusted in time, and the on-site process cannot be adjusted in time. At the same time, the data standards of the process parameters are unified to realize data sharing of the process parameters.

[0004] To achieve the above objectives, the first aspect of the present invention provides a mold intelligent management and control system based on equipment data visualization, including a product identification module, a detection and analysis module, a model building module, and an output feedback module.

[0005] Product recognition module: Use machine vision devices to take photos of products to be produced, obtain product images, and collect and identify product parameters, including cavity, core, production method, size, shape, core pulling times and depth;

[0006] Detection and analysis module: detect and analyze the product through the product parameters, obtain 3D point cloud information, and construct product information;

[0007] Model building module: constructing an image processing model, and training and processing the product image through the product image processing model;

[0008] Output feedback module: Establish a process mechanism model, unify data standards, collect product process parameters in real time through sensors, and feed them back to the image processing model for adjusting the production process.

[0009] Preferably, the detection and analysis module is in communication and electrical connection with a plurality of the product identification modules;

[0010] The product identification module communicates and is electrically connected with the machine vision device and various types of sensors; wherein the machine vision device includes a 3D vision camera and its carrier device;

[0011] The detection and analysis module is in communication and electrical connection with the model building module;

[0012] The model building module is in communication and electrical connection with the output feedback module.

[0013] Preferably, the process of photographing the product to be produced by a machine vision device, acquiring the product image, and collecting and identifying the product parameters includes:

[0014] Taking photos of the products to be produced by using the three groups of 3D vision cameras;

[0015] The first group of 3D vision cameras are installed at an angle of less than 30° and greater than zero relative to the opening and closing surface of the product to be produced. The second group of 3D vision cameras are installed at an angle perpendicular to the opening and closing surface of the product to be produced. The third group of 3D vision cameras are installed at an angle of 120°-150° relative to the opening and closing surface of the product to be produced for taking pictures.

[0016] The product image is acquired through the three groups of 3D vision cameras to collect and identify the cavity, core, production method, size, shape, number of core pulling times and depth of the product to be produced.

[0017] Preferably, the detecting and analyzing the product through the product parameters, acquiring 3D point cloud information, and constructing product information includes:

[0018] Performing image processing on the photos taken by the three groups of 3D vision cameras at different angles;

[0019] The positions of the three groups of 3D vision cameras taking photos at different angles are concentrated in the same three-dimensional coordinate system;

[0020] The data point clouds of the photos at different angles and the model point clouds of the products to be produced are respectively denoted as X={x i}(x=1, 2,...,M) and Y={y i}(y=1, 2, ..., M);

[0021] The minimum objective functions of the point-to-point and point-to-surface data point clouds of the photos at different angles are calculated, and the expressions are:

[0022]

[0023] Among them, R is the rotation matrix of the data point cloud, t is the translation vector of the data point cloud, M is the Mth data point cloud, For point y j The plane normal vector of the model point cloud is j, and point x is the point in the model point cloud after rotation and translation transformation. i The nearest point number;

[0024] The features of the data point cloud are matched with the features of the model point cloud, the docking features are identified and extracted to obtain the 3D point cloud information, and the product information is obtained by the distance from the 3D point cloud information to the straight line where the docking features are located. The expression is:

[0025]

[0026] Among them, q(t) represents the distance from the 3D point cloud information to the straight line where the docking feature is located, p represents the point on the straight line where the docking feature is located, OA represents the straight line vector where the docking feature is located, and t is the translation vector of the data point cloud.

[0027] Preferably, the image processing of the three groups of photos taken by the 3D vision cameras at different angles includes:

[0028] Randomly select a 7x7 pixel grid of the photos at different angles, calculate the pixel at the center of the grid, regard the pixel at the center of the grid as 1, and obtain the pixel points of the remaining grids;

[0029] Move the 7x7 pixel grid upward horizontally or vertically to obtain the image processing result, expressed as:

[0030]

[0031] Among them, g(i,j) is the image processing result;

[0032] Image processing is completed by calculating 3x3 pixel values ​​multiple times.

[0033] Preferably, the step of constructing an image processing model and training the product image through the product image processing model includes:

[0034] Pre-constructing an image processing model, wherein the image processing model includes a deep convolutional neural network model or an RBF neural network model;

[0035] Acquiring output feedback module data, wherein the output feedback module data is acquired when producing the product, including production cycle, casting pressure and injection time;

[0036] The image processing model is trained using the output feedback module data, and the trained image processing model is marked as the image processing model.

[0037] Preferably, the process mechanism model is established and the data standard is unified, including:

[0038] Collect the hierarchical structure, data classification, data nature, data characteristics, and data content of process parameter data according to the pre-built process mechanism model;

[0039] Determine network node information in the process mechanism, and determine common characteristics of the connection between process parameter data sharing data and network nodes through the network node information;

[0040] A unified data standard is established based on the common characteristics, and the process parameter data is shared through mapping between the process parameter data.

[0041] Preferably, the real-time acquisition of process parameters of the product by sensors and the feedback to the image processing model for adjusting the production process include:

[0042] The sensor is used to collect process parameters, and the products of each group of process parameters are compared. The process parameters of products with fewer defects are recorded, and the process parameters of products with fewer defects are fed back to the image processing model to optimize the production process.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. The present invention uses a 3D visual camera to take photos of products, generates 3D point clouds from the photos, and processes the photos of the produced products;

[0045] 2. The present invention can timely adjust the process parameters by feeding back the process parameter data, and can timely adjust the on-site process and optimize the production process;

[0046] 3. The present invention establishes a process mechanism model and unifies the data standards of process parameters, thereby generating standard rules for mold process parameter data and realizing data sharing of process parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0049] Figure 2 A pixel grid table of photos at different angles of the present invention;

[0050] Figure 3A data standard table of process parameters of an embodiment of the present invention;

[0051] Figure 4 A parameter table is produced for an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] See also Figure 1 The first embodiment of the present application provides a mold intelligent management and control system based on equipment data visualization, including a product identification module, a detection and analysis module, a model building module, and an output feedback module.

[0054] Product recognition module: Use machine vision devices to take photos of products to be produced, obtain product images, and collect and identify product parameters, including cavity, core, production method, size, shape, core pulling times and depth;

[0055] Detection and analysis module: detect and analyze the product through the product parameters, obtain 3D point cloud information, and construct product information;

[0056] Model building module: constructing an image processing model, and training and processing the product image through the product image processing model;

[0057] Output feedback module: Establish a process mechanism model, unify data standards, collect product process parameters in real time through sensors, and feed them back to the image processing model for adjusting the production process.

[0058] Further, the detection and analysis module is in communication and electrical connection with a plurality of the product identification modules;

[0059] The product identification module communicates and is electrically connected with the machine vision device and various types of sensors; wherein the machine vision device includes a 3D vision camera and its carrier device;

[0060] The detection and analysis module is in communication and electrical connection with the model building module;

[0061] The model building module is in communication and electrical connection with the output feedback module.

[0062] Furthermore, the process of photographing the product to be produced by a machine vision device, acquiring the product image, and collecting and identifying the product parameters includes:

[0063] Taking photos of the products to be produced by using the three groups of 3D vision cameras;

[0064] The first group of 3D vision cameras are installed at an angle of less than 30° and greater than zero relative to the opening and closing surface of the product to be produced. The second group of 3D vision cameras are installed at an angle perpendicular to the opening and closing surface of the product to be produced. The third group of 3D vision cameras are installed at an angle of 120°-150° relative to the opening and closing surface of the product to be produced for taking pictures.

[0065] The product image is acquired through the three groups of 3D vision cameras to collect and identify the cavity, core, production method, size, shape, number of core pulling times and depth of the product to be produced.

[0066] Furthermore, the detecting and analyzing the product through the product parameters, acquiring 3D point cloud information, and constructing product information includes:

[0067] Performing image processing on the photos taken by the three groups of 3D vision cameras at different angles;

[0068] The positions of the three groups of 3D vision cameras taking photos at different angles are concentrated in the same three-dimensional coordinate system;

[0069] The data point clouds of the photos at different angles and the model point clouds of the products to be produced are respectively denoted as X={x i}(x=1, 2,...,M) and Y={y i}(y=1, 2, ..., M);

[0070] The minimum objective functions of the point-to-point and point-to-surface data point clouds of the photos at different angles are calculated, and the expressions are:

[0071]

[0072] Among them, R is the rotation matrix of the data point cloud, t is the translation vector of the data point cloud, M is the Mth data point cloud, n yj For point y j The plane normal vector of the model point cloud is j, and point x is the point in the model point cloud after rotation and translation transformation. i The nearest point number;

[0073] The features of the data point cloud are matched with the features of the model point cloud, the docking features are identified and extracted to obtain the 3D point cloud information, and the product information is obtained by the distance from the 3D point cloud information to the straight line where the docking features are located. The expression is:

[0074]

[0075] Among them, q(t) represents the distance from the 3D point cloud information to the straight line where the docking feature is located, p represents the point on the straight line where the docking feature is located, OA represents the straight line vector where the docking feature is located, and t is the translation vector of the data point cloud.

[0076] For example, taking the mold as an example, the mold opening area is 32X22 square centimeters; the number of standard holes is 26, the hole diameter is 5mm, the depth is 5mm, and the height of the step hole is 10mm; the number of side cores is 6, and the diameter of the deep hole is 50mm; the number of bottom cores is 1, with a size of 200mmx70mmx5mm; the number of cavities is 1, with a size of 320mmx220mmx5mm, a depth of 750mm, and one cavity per mold.

[0077] Furthermore, the image processing of the three groups of photos taken by the 3D vision cameras at different angles includes:

[0078] Randomly select a 7x7 pixel grid of the photos at different angles, calculate the pixel at the center of the grid, regard the pixel at the center of the grid as 1, and obtain the pixel points of the remaining grids;

[0079] Move the 7x7 pixel grid upward horizontally or vertically to obtain the image processing result, expressed as:

[0080]

[0081] Among them, g(i,j) is the image processing result;

[0082] For example, Figure 2 As shown, the pixel value of any point is the average of the surrounding N*N pixels. The pixel value of the red point in the figure = the sum of the surrounding blue pixel values / 7*7. The calculation result is regarded as unit 1, and the surrounding blue is 1 / 49 of the total pixel value. Then the matrix is ​​used to calculate the surrounding pixel values.

[0083] Image processing is completed by calculating 3x3 pixel values ​​multiple times.

[0084] For example, select the 7*7 pixel values ​​marked in blue above and use 3*3 of them for calculation

[0085]

[0086] M5'=M1*C1+M2*C2+M3*C3+M4*C4+M5*C5+M6*C6+M7*C7+M8*C8+M9*C9

[0087] Gx and Gy are convolution factors, selected respectively

[0088]

[0089]

[0090]

[0091]

[0092] A is the 7*7 pixel value marked in blue above, and 3*3 is used for calculation

[0093] Select

[0094] To improve efficiency, select When the calculation exceeds 225, the pixel value is taken as 225. Then multiple 3*3 pixel value calculations are performed to complete the image analysis.

[0095] It should be explained that, in the embodiment of the present invention, the image acquisition of mechanical parts will inevitably be mixed with noise in the process of converting the shape structure and internal features of the measured object into data information that can be recognized by the computer. The reduction of the signal-to-noise ratio leads to poor quality of the target image, which interferes with the inspection and testing of the parts. Noise will reduce the image quality, blur the image and submerge the features, making subsequent image recognition difficult. Therefore, before feature segmentation and extraction, the image must be filtered and denoised. These denoising operations have an important impact on subsequent feature detection and extraction. Image filtering mainly removes noise by filtering, thereby improving image quality and maintaining image edge information. Mean filtering and median filtering in the spatial domain are mainly used. A 7X7 grid is used after median filtering analysis.

[0096] Furthermore, the constructing of the image processing model and training the product image through the product image processing model include:

[0097] Pre-constructing an image processing model, wherein the image processing model includes a deep convolutional neural network model or an RBF neural network model;

[0098] Acquiring output feedback module data, wherein the output feedback module data is acquired when producing the product, including production cycle, casting pressure and injection time;

[0099] The image processing model is trained using the output feedback module data, and the trained image processing model is marked as the image processing model.

[0100] Furthermore, the process mechanism model is established and the data standard is unified, including:

[0101] Collect the hierarchical structure, data classification, data nature, data characteristics, and data content of process parameter data according to the pre-built process mechanism model;

[0102] Determine network node information in the process mechanism, and determine common characteristics of the connection between process parameter data sharing data and network nodes through the network node information;

[0103] Establishing a unified data standard according to the common characteristics, and sharing the process parameter data through mapping between the process parameter data;

[0104] It should be explained that in the embodiment of the present invention, the product process parameters are stored in the data report, and the process parameter data is sent to the process mechanism model for integration through the network node. The corresponding data fields have defined standard fields to improve the accuracy of the mechanism model. It breaks through the data islands in the production process, speeds up the data circulation, and improves the value of the system data assets. Data standards such as Figure 3 .

[0105] Furthermore, the real-time acquisition of process parameters of the product by the sensor and the feedback to the image processing model for adjusting the production process include:

[0106] The sensor is used to collect process parameters, and the products of each group of process parameters are compared. The process parameters of products with fewer defects are recorded, and the process parameters of products with fewer defects are fed back to the image processing model to optimize the production process.

[0107] For example, Figure 3 As shown in the figure, when the sensor actually reads the production cycle of 137S, the casting pressure of 825bar, the injection time of 130S, and the cooling time of 140S, the on-site process personnel find that the product production has the least defects, and update the real-time monitoring data of the sensor to the system through the negative feedback mechanism to guide the next production. The production parameters in the current system are production cycle of 135S, casting pressure of 827bar, injection time of 131S, and cooling time of 139S.

[0108] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A mold intelligent management and control system based on equipment data visualization, including a product identification module, a detection and analysis module, a model building module, and an output feedback module, characterized in that: Product recognition module: Use machine vision devices to take photos of products to be produced, obtain product images, and collect and identify product parameters, including cavity, core, production method, size, shape, core pulling times and depth; Detection and analysis module: detect and analyze the product through the product parameters, obtain 3D point cloud information, and construct product information; Model building module: constructing an image processing model, and training and processing the product image through the product image processing model; Output feedback module: Establish a process mechanism model, unify data standards, collect product process parameters in real time through sensors, and feed them back to the image processing model for adjusting the production process.

2. According to claim 1, a mold intelligent management and control system based on equipment data visualization is characterized in that: The detection and analysis module is in communication and electrical connection with a plurality of the product identification modules; The product identification module communicates and is electrically connected with the machine vision device and various types of sensors; wherein the machine vision device includes a 3D vision camera and its carrier device; The detection and analysis module is in communication and electrical connection with the model building module; The model building module is in communication and electrical connection with the output feedback module.

3. According to claim 1, a mold intelligent management and control system based on equipment data visualization is characterized in that: The method of photographing the product to be produced by a machine vision device, obtaining the product image, and collecting and identifying the product parameters includes: Taking photos of the products to be produced by using the three groups of 3D vision cameras; The first group of 3D vision cameras are installed at an angle of less than 30° and greater than zero relative to the opening and closing surface of the product to be produced. The second group of 3D vision cameras are installed at an angle perpendicular to the opening and closing surface of the product to be produced. The third group of 3D vision cameras are installed at an angle of 120°-150° relative to the opening and closing surface of the product to be produced for taking pictures. The product image is acquired through the three groups of 3D vision cameras to collect and identify the cavity, core, production method, size, shape, number of core pulling times and depth of the product to be produced.

4. According to claim 1, a mold intelligent management and control system based on equipment data visualization is characterized in that: The detecting and analyzing the product through the product parameters, obtaining 3D point cloud information, and constructing product information includes: Performing image processing on the photos taken by the three groups of 3D vision cameras at different angles; The positions of the three groups of 3D vision cameras taking photos at different angles are concentrated in the same three-dimensional coordinate system; The data point clouds of the photos at different angles and the model point clouds of the products to be produced are respectively denoted as X={x i }(x=1,2,...,M) and Y={y i }(y=1, 2, ..., M); The minimum objective functions of the point-to-point and point-to-surface data point clouds of the photos at different angles are calculated, and the expressions are: Among them, R is the rotation matrix of the data point cloud, t is the translation vector of the data point cloud, M is the Mth data point cloud, n yj For point y j The plane normal vector of the model point cloud is j, and point x is the point in the model point cloud after rotation and translation transformation. i The nearest point number; The features of the data point cloud are matched with the features of the model point cloud, the docking features are identified and extracted to obtain the 3D point cloud information, and the product information is obtained by the distance from the 3D point cloud information to the straight line where the docking features are located. The expression is: Among them, q(t) represents the distance from the 3D point cloud information to the straight line where the docking feature is located, p represents the point on the straight line where the docking feature is located, OA represents the straight line vector where the docking feature is located, and t is the translation vector of the data point cloud.

5. According to claim 1, a mold intelligent management and control system based on equipment data visualization is characterized in that: The image processing of the three groups of photos taken by the 3D vision cameras at different angles includes: Randomly select a 7x7 pixel grid of the photos at different angles, calculate the pixel at the center of the grid, regard the pixel at the center of the grid as 1, and obtain the pixel points of the remaining grids; Move the 7x7 pixel grid upward horizontally or vertically to obtain the image processing result, expressed as: Among them, g(i,j) is the image processing result; Image processing is completed by calculating 3x3 pixel values ​​multiple times.

6. The mold intelligent management and control system based on equipment data visualization according to claim 1 is characterized in that: The constructing of the image processing model and performing training processing on the product image by the product image processing model include: Pre-constructing an image processing model, wherein the image processing model includes a deep convolutional neural network model or an RBF neural network model; Acquiring output feedback module data, wherein the output feedback module data is acquired when producing the product, including production cycle, casting pressure and injection time; The image processing model is trained using the output feedback module data, and the trained image processing model is marked as the image processing model.

7. The mold intelligent management and control system based on equipment data visualization according to claim 1 is characterized in that: The process mechanism model is established and the data standard is unified, including: Collect the hierarchical structure, data classification, data nature, data characteristics, and data content of process parameter data according to the pre-built process mechanism model; Determine network node information in the process mechanism, and determine common characteristics of the connection between process parameter data sharing data and network nodes through the network node information; A unified data standard is established based on the common characteristics, and the process parameter data is shared through mapping between the process parameter data.

8. The mold intelligent management and control system based on equipment data visualization according to claim 1 is characterized in that: The process parameters of the product are collected in real time by sensors and fed back to the image processing model for adjusting the production process, including: The sensor is used to collect process parameters, and the products of each group of process parameters are compared. The process parameters of products with fewer defects are recorded, and the process parameters of products with fewer defects are fed back to the image processing model to optimize the production process.

Images