Implementation method of automatic recognition of electrode shape based on UGNX model analysis

Abstract

The present invention discloses a method for realizing automatic recognition of electrode shapes based on UGNX model analysis, and relates to the technical field of electrode recognition. The present invention comprises the following steps: analyzing the electrode model, searching for the electrode base and the electrode head; setting the electrode rotation coordinate system and automatically rotating the electrode model; analyzing the electrode head, and using color to mark the analysis results. The present invention improves the rate of automatic rotation of the model and the accuracy of recording the positions of the electrode base and the electrode head by setting the electrode rotation coordinate system, and by using color to mark the analysis results, it is convenient for designers to compare and observe the correctness of the processing, and also convenient for the subsequent recognition work of the model processing. By analyzing the electrode model, the recognition rate of the electrode model is improved, the workload of the designer's naked eye recognition is reduced, and the impact on the subsequent electrode programming and detection is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of electrode identification, and in particular relates to a method for realizing automatic electrode shape identification based on UGNX model analysis. Background Art

[0002] In electrode automatic programming (CNC) and automatic detection (CMM), it is necessary to accurately analyze the electrode shape features, find the electrode base and electrode head, and then adopt different strategies to generate CAM programs and CMM detection points. Among them, CAM (Computer Aided Manufacturing) mainly refers to: by directly or indirectly connecting computers with manufacturing processes and production equipment, using computer systems to plan and manage manufacturing processes, control and operate production equipment, process data required in product manufacturing, control and process the flow of materials (blanks and parts, etc.), test and inspect products, etc. The CMM (the Capability Maturity Model for software) process is the mechanism for producing products. Whether it is process improvement or capability determination, process evaluation is required, and process evaluation is usually based on some proposed evaluation models.

[0003] The current electrode shape analysis completely relies on the designer's experience. The designer observes with the naked eye based on past experience. However, this method is inefficient and will have a great impact on subsequent electrode programming and testing. Summary of the invention

[0004] The purpose of the present invention is to provide a method for realizing automatic recognition of electrode shape based on UGNX model analysis, which solves the technical problem that the current electrode shape analysis completely relies on the experience of the designer. The designer observes with the naked eye based on past experience, but this method is inefficient and will also have a great impact on subsequent electrode programming and detection.

[0005] To achieve the above object, the present invention is achieved through the following technical solutions:

[0006] A method for realizing automatic recognition of electrode shape based on UGNX model analysis comprises the following steps:

[0007] The computer analyzes the electrode model and finds the electrode base and electrode head;

[0008] The computer sets the electrode rotation coordinate system and automatically rotates the electrode model;

[0009] The computer analyzes the electrode tip and uses color to code the results.

[0010] Optionally, during the computer analysis of the electrode model, a unique barcode number is generated for each electrode, and the corresponding relationship between the two is uploaded to the database of the local area network.

[0011] Optionally, when analyzing the electrode model, the computer collects images of the electrode surface at various angles to obtain the original image, uses histogram equalization to increase the contrast of the original image to improve the image quality, and displays the processed original image. The processed image is subjected to an image recognition process, and edge detection, image segmentation and morphological operations are performed in sequence to complete the preparation for recognition.

[0012] Optionally, the electrode surface is divided into several parts based on the original image, a part is selected and edge detection is performed based on all associated images of the part, the detected edges are marked to obtain an edge marking image, the electrode base and the electrode head are segmented along the marked edges, and the segmented electrode base area, electrode head area and other areas are represented by a binary image. Image binarization is to set the grayscale value of the pixel points on the image to 0 or 255, that is, the process of presenting an obvious black and white effect on the entire image. In digital image processing, binary images occupy a very important position. The binarization of the image greatly reduces the amount of data in the image, thereby highlighting the outline of the target. The grayscale image of 25 brightness levels is selected through an appropriate threshold to obtain a binary image that can still reflect the overall and local features of the image. In digital image processing, binary images occupy a very important position. First, the binarization of the image is conducive to the further processing of the image, making the image simple, and reducing the amount of data, which can highlight the outline of the target of interest. Secondly, to process and analyze the binary image, the grayscale image must first be binarized to obtain Binarize the image, repeat steps 3-5 times to identify and detect the electrode base area and the electrode head area, the computer compares and integrates the results, and completes the identification and marking of all electrode bases and electrode heads. After image segmentation, use the binary image to display the electrode base and electrode head area. The electrode base area can be displayed in a dark color, and the electrode head area in a light color. At the same time, the electrode base area can be displayed in a light color and the electrode head area in a dark color by inverting the image. Subsequently, a morphological operation is performed on the image to eliminate other areas, and then a dilation operation is performed (the dilation operation is also a type of morphological operation, so it can also be understood as performing two morphological operations) to make the electrode base and electrode head areas fully and completely cover the corresponding electrode base and electrode head, and finally obtain a complete electrode base and electrode head image.

[0013] Optionally, the computer determines the direction of the XY axis vector according to the long and short sides of the base, defines the electrode shape rotation coordinate system, rotates the electrode shape according to the rotation coordinate system, and the computer conducts a comprehensive analysis of the appearance shape, material properties, exposed area, and fracture resistance of the electrode head. The computer outputs the electrode head shape information and automatically uses different colors to distinguish and mark according to the characteristics. The identified structure can be used for automatic electrode programming and automatic electrode point output. Color differentiation can be performed according to appearance shape, color differentiation can also be performed according to material properties, etc. For example, this type of electrode head characteristic is used as a reference for color differentiation marking, or multiple characteristics of the electrode head are combined as a distinction standard.

[0014] The embodiments of the present invention have the following beneficial effects:

[0015] One embodiment of the present invention improves the rate of automatic rotation of the model and the accuracy of recording the positions of the electrode base and the electrode head by setting an electrode rotation coordinate system. By using color to mark the analysis results, it is convenient for designers to compare and observe the correctness of computer processing, and also facilitates the subsequent computer recognition of model processing. By analyzing the electrode model, the recognition rate of the electrode model is improved, the workload of the designer's naked eye recognition is reduced, and the impact on subsequent electrode programming and detection is reduced.

[0016] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings constituting a part of the present application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 It is a structural schematic diagram of an automatic electrode shape recognition step according to an embodiment of the present invention;

[0019] Figure 2 Schematic diagram of the structure of the automatic electrode shape recognition process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present invention and its application or use.

[0021] In order to keep the following description of the embodiments of the present invention clear and concise, detailed descriptions of well-known functions and well-known components are omitted.

[0022] See also Figure 1-2 As shown, in this embodiment, a method for realizing automatic recognition of electrode shape based on UGNX model analysis is provided, comprising the following steps:

[0023] S1. Computer analyzes the electrode model and finds the electrode base and electrode head;

[0024] S2. The computer sets the electrode rotation coordinate system and automatically rotates the electrode model;

[0025] S3. The computer analyzes the electrode tip and uses color to mark the analysis results.

[0026] The application of one aspect of this embodiment is: the computer first analyzes the electrode model, searches for the base shape, determines the XY axis vector direction according to the long and short sides of the base, defines the electrode shape rotation coordinate system, and then rotates the electrode shape according to the rotation coordinate system, analyzes the electrode head, analyzes the electrode head shape characteristics, outputs the electrode head shape information, automatically colors it, and finally uses the identified structure for automatic electrode programming and automatic electrode output. It should be noted that the electrical equipment involved in this application can be powered by a battery or an external power supply.

[0027] By setting the electrode rotation coordinate system, the rate of automatic rotation of the model and the accuracy of recording the positions of the electrode base and the electrode head are improved. By using color to mark the analysis results, it is convenient for designers to compare and observe the correctness of computer processing, and it is also convenient for the computer to recognize the model processing later. By analyzing the electrode model, the recognition rate of the electrode model is improved, the workload of the designer's naked eye recognition is reduced, and the impact on subsequent electrode programming and detection is reduced.

[0028] like Figure 1 As shown, in the process of analyzing the electrode model by the computer in this embodiment, a unique barcode number is generated for each electrode, and the corresponding relationship between the two is uploaded to the database of the local area network.

[0029] like Figure 1 As shown, when analyzing the electrode model of this embodiment, the computer collects images of the electrode surface at various angles to obtain the original image, pre-processes the image first, and then optimizes the image. Histogram equalization is used to increase the contrast of the original image to facilitate the identification of the edge of the image, and the processed original image is displayed. The processed image is subjected to the image recognition process, and the edge detection, image segmentation and morphological operation are performed in sequence to complete the preparation for recognition.

[0030] like Figure 1As shown, in this embodiment, the electrode surface is divided into several parts based on the original image, one part is selected and edge detection is performed based on all associated images of the part, and the detected edges are marked to obtain an edge marked image. The electrode base and the electrode head are segmented along the marked edges, and the segmented electrode base area, the electrode head area and other areas are represented by a binary image. Image binarization is to set the grayscale value of the pixel on the image to 0 or 255, that is, the process of presenting the entire image with an obvious black and white effect. In digital image processing, binary images occupy a very important position. The binarization of the image greatly reduces the amount of data in the image, thereby highlighting the outline of the target. The grayscale image of 25 brightness levels is selected through an appropriate threshold to obtain a binary image that can still reflect the overall and local features of the image. In digital image processing, binary images occupy a very important position. First, the binarization of the image is conducive to the further processing of the image, making the image simple, and reducing the amount of data, which can highlight the outline of the target of interest. Secondly, to process and analyze the binary image, the grayscale image must first be binarized to obtain a binary image.

[0031] like Figure 1 As shown, the present embodiment repeats steps 3-5 times to identify and detect the electrode base area and the electrode head area. The computer compares and integrates the results to complete the identification marking of all electrode bases and electrode heads. After image segmentation, the electrode base and electrode head areas are displayed in a binary image. The electrode base area can be displayed in a dark color, and the electrode head area in a light color. At the same time, the electrode base area can be displayed in a light color and the electrode head area in a dark color by inverting the image. Subsequently, a morphological operation is performed on the image to eliminate other areas, and then an expansion operation is performed (the expansion operation is also a type of morphological operation, so it can also be understood as performing two morphological operations) so that the electrode base and electrode head areas fully and completely cover the corresponding electrode base and electrode head, and finally a complete electrode base and electrode head image is obtained. An exclusive barcode number is generated for each electrode base and electrode head, and the correspondence between the two and the relationship with the electrode are uploaded to the database of the local area network. The electrode base barcode includes information such as the length, width, and height dimensions of the electrode base, chamfer dimensions, and position coordinates.

[0032] like Figure 2 As shown, the computer of this embodiment determines the direction of the XY axis vector according to the long and short sides of the base, defines the electrode shape rotation coordinate system, and rotates the electrode shape according to the rotation coordinate system.

[0033] like Figure 1 As shown, the computer of this embodiment performs a comprehensive analysis on the appearance shape, material properties, exposed area, and fracture resistance of the electrode head, and one of them can also be set for special analysis.

[0034] like Figure 2 As shown, the computer of this embodiment outputs the shape information of the electrode head, and automatically uses different colors to distinguish and mark according to the characteristics. The recognized structure can be used for automatic electrode programming and automatic electrode point output. Color distinction can be made according to the appearance shape, and color distinction can also be made according to the material characteristics. For example, this type of electrode head characteristic is used as a reference for color distinction marking, or multiple characteristics of the electrode head are combined as a distinction standard.

[0035] Embodiment 1:

[0036] The steps of preprocessing the original image in this embodiment include: the main control node obtains the average CPU utilization and average action response time of all edge computing devices in the database for image preprocessing in each cycle; the main control node calculates the state of each edge computing device for the next cycle through the load prediction model; the main control node calculates the state of each edge computing device for the next cycle through the load prediction model; the main control node schedules the image preprocessing action of each edge computing device for the next cycle; the edge computing device performs image preprocessing according to the instructions of the main control node, and feeds back the response time and CPU utilization of each action; the main control node integrates the images preprocessed by all edge computing devices.

[0037] The preprocessing actions are as follows: first, determine whether the image is out of focus: make differences in the horizontal and vertical directions of the image respectively, and accumulate the differences. When the total difference value exceeds the threshold, it is judged to be out of focus and the image is deleted; then, perform image quality evaluation: perform image quality evaluation on all images within a cycle, and retain images with better quality; finally, perform image recognition: calculate the directional gradient features of each regional surface through a sliding window, compare them with the features of the object for classification, and determine whether there is an object. If there is an object, further extract the edge and tip features of the object.

[0038] The image quality evaluation method may be based on Tenengrad gradient, Laplacian gradient, SMD (grayscale variance), SMD2 (grayscale variance product), and the like.

[0039] The load prediction model is to establish a first-order Markov prediction model corresponding to the average CPU utilization rate of each edge computing device in each cycle of image preprocessing; the predicted state is divided into overload O, normal N, and low load U according to the CPU utilization rate, and the following state transition probability matrix is ​​generated:

[0040]

[0041] PUU represents the probability of transferring from the low load state in the previous cycle to the low load state in the next cycle, PNU represents the probability of transferring from the normal state in the previous cycle to the low load state in the next cycle, POU represents the probability of transferring from the overload state in the previous cycle to the low load state in the next cycle, PUN represents the probability of transferring from the low load state in the previous cycle to the normal state in the next cycle, PNN represents the probability of transferring from the normal state in the previous cycle to the normal state in the next cycle, PON represents the probability of transferring from the overload state in the previous cycle to the normal state in the next cycle, PUO represents the probability of transferring from the low load state in the previous cycle to the overload state in the next cycle, PNO represents the probability of transferring from the normal state in the previous cycle to the overload state in the next cycle, and POO represents the probability of transferring from the overload state in the previous cycle to the overload state in the next cycle.

[0042] The response prediction model can establish a response time prediction model for the average action response time of each image preprocessing cycle of the entire edge image processing system, or it can establish a response time prediction model for the average action response time of each image preprocessing cycle of each edge computing device.

[0043] The response time prediction model may be a Kalman filter, a particle filter, an ARMAX model (Autoregressive moving-average model with exogenous inputs), or the like.

[0044] The steps for the master control node to schedule the image preprocessing actions of each edge computing device in the next cycle are as follows: the master control node migrates all preprocessing actions of the overloaded edge computing device in the next cycle to the unloaded edge computing device for processing according to the load prediction model results; the master control node migrates the preprocessing actions of the edge computing device with response timeout in the next cycle to the edge computing device with surplus response time according to the response prediction model; on the premise of ensuring that the edge device is not overloaded and timed out while migrating within the edge device, all preprocessing actions that are too late for the edge computing device to be performed are migrated to the master control node for processing.

[0045] The edge image processing system consists of a main control node and at least one edge computing device connected through one or more of the network, Bluetooth or USB, and also includes a conventional peripheral interface (USB interface), network module, power supply module, Bluetooth module, etc. The main control node can be a cloud server or an edge computing device node. The edge computing device is equipped with an image processor and is connected to an image input device.

[0046] The main control node includes: a resource management module: used to collect node status information uploaded by the edge computing device management module in distributed edge computing, including node memory status, CPU status, GPU status, running task information, response time, etc.; a task management module: used to send image preprocessing tasks to the edge computing device and complete scheduling; a data preprocessing module: used to preprocess and split execution task data; a result aggregation module: used to collect the analysis results of the operation modules of each edge computing device, and organize the result set according to the serial number; an output module: used to output the analysis result set to the application layer.

[0047] The edge computing device includes: a node management module, which is used to collect status information of the edge computing device, including memory status, CPU status, GPU status, and running task information, as well as receive computing tasks, start the computing module and pass specific computing parameters to the computing module, and the computing module is used to specifically execute the image preprocessing algorithm.

[0048] Build an edge image processing system, extract the entire simple task of image preprocessing from the entire image processing task to the edge computing device for processing, and by establishing a load prediction model and a load prediction model, realize the edge computing of image preprocessing and the optimal configuration of edge computing devices, improve the utilization of computing resources, save data transmission time and cost, and reduce data latency.

[0049] Embodiment 2:

[0050] The steps of optimizing the original image in this embodiment include: extracting the image to be processed; determining the quality characteristics of the image; inputting the image and the quality characteristics into an image processing model, and determining an optimized image of the image based on the output of the image processing model.

[0051] An image processing model is selected based on initial features of the image to be processed, wherein the initial features include at least one of features related to an object in the image, a scanning type for acquiring the image, and / or a reconstruction algorithm for acquiring the image.

[0052] The quality feature includes at least one of a noise feature, an artifact feature, a grayscale distribution, a global grayscale, a resolution, and a contrast of the image.

[0053] The noise feature includes at least one of a noise distribution, a noise intensity, and a noise rate; and / or the artifact feature includes at least one of an artifact distribution, an artifact intensity, and an artifact rate.

[0054] The training process of obtaining the image processing model includes: obtaining multiple groups of training samples, one of the multiple groups of training samples includes at least: sample images, quality features of sample images; the training process of obtaining the image processing model includes: obtaining multiple groups of training samples, one of the multiple groups of training samples includes at least: sample images, quality features of sample images.

[0055] The sample quality feature includes at least one of a sample noise feature, a sample artifact feature, a sample grayscale distribution, a sample global grayscale, a sample resolution, and a sample contrast of the sample image.

[0056] The sample noise feature includes at least one of a sample noise distribution, a sample noise intensity, and a sample noise rate; and / or the sample artifact feature includes at least one of a sample artifact distribution, a sample artifact intensity, and a sample artifact rate.

[0057] The quality weight is positively correlated with sample resolution, the quality weight is negatively correlated with sample noise intensity, the quality weight is negatively correlated with sample artifact intensity; and / or the quality weight is negatively correlated with sample contrast.

[0058] The training process of obtaining the image processing model includes: acquiring multiple qualified images related to the scanning device type; processing the multiple qualified images to generate multiple sample images, wherein the processing includes at least one of segmentation, adding noise and / or adding artifacts; and training the initial image processing model based on the multiple sample images and their corresponding multiple sample quality features to obtain the image processing model.

[0059] The training process of obtaining an image processing model includes: acquiring multiple qualified images related to the object type; processing the multiple qualified images to generate multiple sample images, wherein the processing includes at least one of segmentation, adding noise and / or adding artifacts; and training an initial image processing model based on the multiple sample images and their corresponding multiple sample quality features to obtain an image processing model.

[0060] The image processing model is a deep learning model, which includes a deep neural network (DNN) model, a multi-layer neural network (MLP) model, a convolutional neural network (CNN) model, a generative adversarial neural network (GAN) model, and / or a deep convolutional encoding and decoding (DCED) neural network model.

[0061] Embodiment 2:

[0062] The image segmentation method of the present embodiment is: acquiring multiple related reference images; registering the acquired multiple reference images; marking the edges of the multiple reference images according to the features of different edges on the multiple reference images after registration; determining the control points on the multiple reference images after registration according to the edges of the processed images; obtaining the control points of the processed image according to the multiple control points on the multiple reference images; and identifying the model area to be marked according to the control points of the processed image.

[0063] The model includes one or more second edges, and obtaining a model further includes generating a correlation factor of the control point of the model according to a relationship between the control point of the model and the one or more second edges in the model.

[0064] The image segmentation method further includes: acquiring and processing image data; reconstructing an image based on the image data, wherein the image includes one or more first edges; matching the model with the reconstructed image based on the one or more first edges and the one or more second edges; and adjusting the one or more second edges of the model based on the one or more first edges.

[0065] Adjusting one or more second edges of the model includes: performing a similarity transformation on the second edges; performing an affine transformation on the second edges according to a correlation factor; or fine-tuning the second edges according to an energy function.

[0066] The above-mentioned embodiments may be combined with each other.

[0067] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0068] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "lateral, vertical, perpendicular, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the devices or elements referred to must have a specific direction or be constructed and operated in a specific direction. Therefore, they cannot be understood as limiting the scope of protection of the present invention. The directional words "inside and outside" refer to the inside and outside relative to the contours of each component itself.

Claims

1. A method for realizing automatic recognition of electrode shape based on UGNX model analysis, characterized in that: The method comprises the following steps: analyzing the electrode model, searching for the electrode base and the electrode head, collecting images of the electrode surface at various angles to obtain the original image when analyzing the electrode model, using histogram equalization to increase the contrast of the original image, and displaying the processed original image, dividing the electrode surface into several parts based on the original image, selecting a part and performing edge detection based on all associated images of the part, marking the detected edges to obtain an edge marking image, segmenting the electrode base and the electrode head along the marked edges, representing the segmented electrode base area, electrode head area and other areas with a binary image, determining the XY axis vector direction according to the long and short sides of the base, defining a normalized coordinate system for the electrode shape, and normalizing the electrode shape according to the normalized coordinate system; Set the electrode rotation coordinate system and automatically rotate the electrode model; Analyze the electrode tip and use color coding to analyze the results.

2. The method for realizing automatic electrode shape recognition based on UGNX model analysis according to claim 1, characterized in that: During the analysis of the electrode model, a unique barcode number is generated for each electrode, and the corresponding relationship between the two is uploaded to the database of the local area network.

3. The method for realizing automatic electrode shape recognition based on UGNX model analysis according to claim 1, characterized in that: Repeat the identification and detection of the electrode base area and the electrode head area 3-5 times, compare and integrate the results, and complete the identification and marking of all electrode bases and electrode heads.

4. The method for realizing automatic electrode shape recognition based on UGNX model analysis according to claim 1, characterized in that: A comprehensive analysis is conducted on the appearance, shape, material properties, exposed area and fracture resistance of the electrode head.

5. The method for realizing automatic electrode shape recognition based on UGNX model analysis according to claim 1, characterized in that: Output the electrode head shape information, and automatically use different colors to distinguish and mark according to the characteristics. The identified structure is used for automatic electrode programming and automatic electrode output.

Images