Graphite column synthetic block image data acquisition system

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a graphite column synthesis block image data acquisition system, which comprises a microscope, the microscope comprises a main body, the main body is sequentially provided with an ocular lens, an objective lens and an objective table from top to bottom, and the objective table is rotatably arranged on the main body through a rotating mechanism; the rotating mechanism comprises a fixed rod and a movable rod, the first end of the fixed rod is connected with the main body, the second end of the fixed rod is provided with a rotating part, the first end of the movable rod is connected with the rotating part, and the second end of the movable rod is fixed with the objective table; the width of the fixing rod is gradually increased from the first end to the second end of the fixing rod, and a limiting piece is slidably arranged on the fixing rod, the objective table can move left and right through the rotating mechanism, so that images of graphite column synthesis blocks at different angles are acquired, use is convenient and fast, and the acquisition efficiency is greatly improved.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a graphite column synthesis block image data acquisition system. Background Art

[0002] During the diamond production stage, the newly produced graphite block is embedded with multiple diamond crystals, which are usually very small and difficult to observe directly with the naked eye. Therefore, it is necessary to place the graphite block synthesis block on a microscope and use the high magnification function of the microscope to observe the graphite column synthesis block to determine its quality, including crystal distribution and crystal form. In the traditional quality inspection process, technicians need to use microscopes to manually evaluate the quality of graphite column synthesis blocks. This process is not only time-consuming, but also requires high experience and professional ability of technicians.

[0003] To this end, a camera can be used to capture images of samples, which can greatly improve detection efficiency and reduce time and labor costs. For example, the Chinese invention patent with application number 202311766077.5 provides a mobile phone holder and microscope assembly installed on a microscope. For another example, the Chinese invention patent with application number 201480031984.8 provides an adapter system for installing an image capture device on an optical observation instrument. By setting a mobile phone holder or fixture on the microscope, the user can easily complete the image acquisition work by installing the mobile phone or other image acquisition device on the mobile phone holder. It has good stability and can be applied to different models of mobile phones, and has a wide applicability.

[0004] However, when collecting images of graphite column synthetic blocks, in order to ensure detection accuracy, it is often necessary to collect images of the same graphite column synthetic block at multiple angles. The current common adjustment method is to adjust the position of the eyepiece or the objective lens so that the microscope can collect images of the graphite column synthetic blocks at different positions. However, this method has many problems. For example, the adjustment process of the eyepiece and the objective lens is very cumbersome. For example, after the eyepiece position is moved, the position of the camera needs to be moved synchronously to ensure that the two can correspond. When batch production inspection is carried out in a factory or production line, the shooting angle needs to be adjusted frequently. Such a cumbersome operation method is difficult to ensure its work efficiency. Therefore, there is an urgent need for a graphite column synthetic block image data acquisition system that can easily adjust the shooting angle in a production line. Summary of the invention

[0005] The object of the present invention is to provide a graphite column synthesis block image data acquisition system that is convenient for adjusting the shooting angle, so as to improve the image acquisition efficiency.

[0006] In order to solve the above-mentioned technical problems, the present invention specifically adopts the following technical solutions: a graphite column synthesis block image data acquisition system, comprising a microscope and an image acquisition device, the microscope comprising a main body, the main body is provided with an eyepiece, an objective lens and a stage from top to bottom in sequence, the stage is provided with a groove for accommodating the graphite column synthesis block, the stage is rotatably arranged on the main body through a rotating mechanism; the main body is also provided with a knob for adjusting the height and angle of the objective lens and the eyepiece, and a mounting bracket for fixing the image acquisition device;

[0007] The rotating mechanism comprises a fixed rod and a movable rod, wherein the first end of the fixed rod is connected to the main body, the second end of the fixed rod is provided with a rotating member, the first end of the movable rod is connected to the rotating member, and the second end of the movable rod is fixed to the loading platform; the width of the fixed rod gradually increases from the first end to the second end thereof, and a limit member is slidably provided on the fixed rod;

[0008] When the loading platform is located at the initial position, the limiting member is located at the connection position between the fixed rod and the movable rod, so that the fixed rod and the movable rod are located on the same axis; when the loading platform is rotated, the loading platform drives the movable rod to rotate around the second end of the fixed rod, and the limiting member moves toward the first end of the fixed rod;

[0009] Wherein, the image acquisition device specifically includes:

[0010] An image data acquisition module is used to acquire images to be detected of the cross section of the graphite column synthesis block from the production line in real time; the images to be detected include complete images of the cross section;

[0011] The detection module is used to detect the image to be detected using a pre-trained quality detection model to obtain the crystal distribution grade and crystal type grade of the graphite column synthetic block.

[0012] As an improvement, the acquisition system further includes:

[0013] A training set construction module, which is used to collect the cross-sectional images of the graphite column synthesis block under various production parameters for constructing a training set; the cross-sectional images include complete cross-sectional images; the production parameters include equipment parameters; the equipment parameters further include power parameters and pressure parameters;

[0014] The model building module is used to input the manually annotated graphite column synthetic block cross-section image into a pre-built model for training, thereby obtaining a quality inspection model.

[0015] As an improvement, the acquisition system further comprises: a database for storing production parameters corresponding to each crystal distribution grade and crystal form grade;

[0016] The equipment parameter matching module is used to match the crystal distribution grade and crystal type grade distribution detected by the detection module from the database to the corresponding production parameters.

[0017] As an improvement, the partition re-inspection module is used to determine whether the crystal distribution level of the detection module is lower than the threshold level. If so, the image data acquisition module is triggered to perform a partition re-inspection on the crystal level.

[0018] As an improvement, the rotating mechanism further comprises an elastic member, and two ends of the elastic member are respectively connected to the limiting member and the movable rod;

[0019] When the movable rod rotates, the limiting member moves toward the first end of the fixed rod; when the force acting on the movable rod disappears, the elastic member is reset to drive the movable member and the limiting member to reset.

[0020] As an improvement, the second end of the fixed rod is recessed toward its first end to form a recessed portion, and correspondingly, the movable rod protrudes toward its first end to form a protruding portion that cooperates with the recessed portion, and the protruding portion can rotate around the center point of the recessed portion.

[0021] As an improvement, the concave portion is V-shaped, and correspondingly, the convex portion is also V-shaped, and the opening of the concave portion is larger than the opening of the convex portion; or, the concave portion is arc-shaped, and correspondingly, the convex portion is also arc-shaped, and the curvature of the concave portion is larger than the curvature of the convex portion;

[0022] When the movable rod is rotated to a maximum angle in the first direction, the first side of the protrusion abuts against the first side of the recess, and the acute angle between the central axis of the movable rod and the central axis of the fixed rod is 20°-45°.

[0023] As an improvement, the acute angle between the central axis of the movable rod and the central axis of the fixed rod is 30°.

[0024] As an improvement, the objective lens is connected to the eyepiece to form a lens assembly, and a lighting assembly is provided on the lens assembly. The lighting assembly includes a shell, a lighting lamp provided on the shell, and a switch for controlling the lighting lamp to be turned on and off.

[0025] As an improvement, a plurality of anti-slip protrusions are provided at the bottom of the groove.

[0026] The principles and beneficial effects of the present invention are:

[0027] In view of the situation where the collection volume is large, such as in production lines, and images of graphite column composite blocks at different angles need to be collected, the present application proposes an image data collection system that is easy to operate and can flexibly adjust the collection angle.

[0028] Specifically, the present application connects the main body with the stage through a rotating mechanism. When in use, the acquisition position can be adjusted by only gently pushing or pulling the stage, without the need to frequently adjust and focus the position of the lens and the eyepiece of the microscope in the image acquisition device; further, the fixed rod and the movable rod are matched with a V-shaped or arc-shaped structure, which can limit the rotation angle during the rotation process, thereby preventing the stage from deviating from the image acquisition range during the rotation process; further, the present application also arranges an elastic member between the movable member and the limit member. When no external force acts on the stage and the movable rod, the elastic member can drive the movable rod to automatically reset. When resetting, the elastic member touches the movable rod and makes a sound, prompting the user to restore the initial position, which is simple and intuitive. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without paying creative labor.

[0030] Figure 1 It is a schematic diagram of the overall structure of a graphite column synthesis block image data acquisition system according to an exemplary embodiment of the present invention;

[0031] Figure 2 is a structural schematic diagram of a lens assembly according to an exemplary embodiment of the present invention;

[0032] Figure 3 is a schematic structural diagram of a lighting assembly according to an exemplary embodiment of the present invention;

[0033] Figure 4 It is a schematic structural diagram of an assembled lens assembly and an illumination assembly according to an exemplary embodiment of the present invention;

[0034] Figure 5a A top view of a rotating mechanism of an exemplary embodiment of the present invention in an initial state;

[0035] Figure 5b A side view of a rotating mechanism of an exemplary embodiment of the present invention in an initial state;

[0036] Figure 5c A schematic diagram of the bendable angle of the rotating mechanism in the first direction according to an exemplary embodiment of the present invention;

[0037] Figure 5d A schematic diagram of the bendable angle of the rotating mechanism in the second direction according to an exemplary embodiment of the present invention;

[0038] Figure 5e It is a schematic diagram of a stopper of a rotating mechanism according to an exemplary embodiment of the present invention being located at different positions;

[0039] Figure 6 is a schematic diagram of the overall structure of a stage according to an exemplary embodiment of the present invention;

[0040] Figure 7 is a schematic diagram of an exemplary embodiment of the present invention with a stage located at different positions;

[0041] Figure 8 The present invention is a flowchart of a method for detecting the quality of a graphite column synthetic block based on image processing according to an exemplary embodiment of the present invention.

[0042] Markings in the figure: 1. Image acquisition device; 2. Mounting bracket; 3. Knob; 4. Illumination assembly; 41. Block; 42. Switch; 43. Illumination lamp; 5. Rotation mechanism; 51. Fixed rod; 52. Movable rod; 53. Limiting member; 54. Rotating member; 6. Stage; 61. Groove; 7. Lens assembly; 71. Eyepiece; 72. Objective lens; 73. Buckle. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0044] Herein, suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meanings by themselves. Therefore, "module", "component" or "unit" can be used mixedly.

[0045] In this document, the terms "upper", "lower", "inner", "outer", "front", "back", "one end", "the other end" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0046] In this document, unless otherwise clearly specified and limited, the terms "installed", "provided with", "connected", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0047] Herein "and / or" includes any and all combinations of one or more of the associated listed items.

[0048] Herein, "plurality" means two or more than two, ie, it includes two, three, four, five, etc.

[0049] Example 1

[0050] like Figure 1-7 The present invention provides a graphite column synthetic block image acquisition system, which includes a microscope and an image acquisition device mounted on the microscope through a mounting bracket, wherein the microscope includes a main body, and the main body is provided with an eyepiece, an objective lens and a stage 6 in sequence from top to bottom, and the stage 6 is rotatably arranged on the main body through a rotating mechanism 5;

[0051] like Figure 5a-5d As shown, the rotating mechanism includes a fixed rod 51 and a movable rod 52, the first end of the fixed rod 51 is connected to the main body, the second end of the fixed rod 51 is provided with a rotating member 54, the first end of the movable rod 52 is connected to the rotating member 54, and the second end of the movable rod 52 is fixed to the stage; the width of the fixed rod 51 gradually increases from the first end to the second end thereof, and a limiting member 53 is slidably provided on the fixed rod 51;

[0052] like Figure 5a and 5bWhen the loading platform is located at the initial position, the fixed rod 51 and the movable rod 52 are located on the same axis, and the limiting member 53 is located at the connection position of the fixed rod 51 and the movable rod 52 to fix the two; when the loading platform is rotated, the loading platform drives the movable rod 52 to rotate around the second end of the fixed rod 51 as the center, and the limiting member 53 moves toward the first end of the fixed rod 51;

[0053] like Figure 1 The main body is also provided with a knob 3 for adjusting the height and angle of the objective lens and the eyepiece, and a mounting bracket 2 for fixing the image acquisition device.

[0054] During specific use, in the initial state, the fixed rod and the movable rod are located on the same axis, and the rotating stage drives the movable rod to rotate. During the rotation, the first end of the movable rod generates a force on the limit piece in the direction of the central axis of the fixed rod, and then the limit piece automatically slides down and no longer contacts the movable rod. When the movable rod needs to be restored to its initial position, just push the slider back to the contact position between the fixed rod and the movable rod. After pushing it into place, the user can feel a sense of jamming and a "click" sound will be emitted to prompt the user.

[0055] like Figure 7 , showing a schematic diagram of the stage rotating from a first position to a second position and a third position in sequence, wherein the first position is when the stage is rotated to a maximum angle in the first direction, the second position is the initial position mentioned above, and the third position is when the stage is rotated to a maximum angle in the second direction.

[0056] In some embodiments, the rotating member may be a rotating pair. Of course, those skilled in the art may also select other rotating components.

[0057] In some embodiments, the positions of the mounting bracket and the lens assembly are relatively fixed, wherein the mounting bracket and the lens assembly can be installed on the same part of the microscope body, and when the knob is turned to move the lens assembly, the mounting bracket also moves accordingly, thereby avoiding the need to separately adjust the position of the image acquisition device after the lens assembly is moved, which causes cumbersome operations.

[0058] In some embodiments, the rotating mechanism further includes an elastic member (such as a spring, not shown in the figure), and two ends of the elastic member are respectively connected to the limit member and the movable rod;

[0059] When the movable rod rotates, the limiting member moves toward the first end of the fixed rod; when the force acting on the movable rod disappears, the elastic member returns to its original position and drives the movable member and the limiting member to return to their original position. Figure 5a Status shown.

[0060] On the one hand, when there is no external force acting on the loading platform and the movable rod, the elastic member can drive the movable rod to automatically reset. When resetting, the elastic member touches the movable rod and makes a sound, prompting the user to restore the initial position, which is simple and intuitive. On the other hand, the elastic member can also pull the movable rod to generate appropriate resistance, thereby preventing the operator from moving the loading platform too fast when pushing or pulling the loading platform.

[0061] In some embodiments, the fixed rod, the movable rod and the reset member are all made of metal. On the one hand, this can make the device more secure. On the other hand, the sound made by the metal elastic member touching the movable rod will be more obvious, which can serve as a good reminder.

[0062] In some embodiments, the second end of the fixed rod is recessed toward the first end thereof to form a recessed portion, and correspondingly, the movable rod is protruded toward the first end thereof to form the protruding portion that cooperates with the recessed portion, and the protruding portion can rotate around the center point of the recessed portion;

[0063] In some embodiments, the recessed portion is V-shaped, and correspondingly, the protruding portion is also V-shaped, and the opening of the recessed portion is larger than the opening of the protruding portion;

[0064] In some other embodiments, the recessed portion is arc-shaped, and correspondingly, the protruding portion is also arc-shaped, and the curvature of the recessed portion is greater than that of the protruding portion.

[0065] like Figure 5c and 5d When the movable rod 52 is rotated to the maximum angle in the first direction or the second direction, the first side of the protrusion abuts against the first side of the recessed portion. At this time, the acute angle between the central axis of the movable rod 52 and the central axis of the fixed rod 51 is α, wherein the angle α can be 20°-45° (preferably 30°).

[0066] In some embodiments, Figure 2 The objective lens 72 is connected to the eyepiece 71 to form a lens assembly, that is, the objective lens 72 and the eyepiece 71 are integrated, and the objective lens 72 is indirectly connected to the main body through the eyepiece 71.

[0067] In some embodiments, the eyepiece is provided with an illumination assembly, such as Figure 3 The lighting assembly includes a shell, a lighting lamp 43 arranged on the shell, and a switch 42 for controlling the lighting lamp to be turned on and off.

[0068] In other embodiments, since the eyepiece and the objective lens are designed as an integrated whole, the lighting assembly can also be installed on the objective lens in the lighting assembly.

[0069] Since diamonds mainly include single crystals, conjoined bodies and polycrystals, among which single crystals are the most common, and their specific forms include cubes, octahedrons and rhombic dodecahedrons, local reflection problems are prone to occur during the detection process. Once reflection occurs, the collected image is very likely to be inaccurate, which in turn affects the detection results. The present application effectively avoids the above problems by arranging an illumination component on the microscope so that the light source is evenly illuminated on the diamond to be detected. When the illumination component in the present application is in use, the illumination range of the illumination lamp can cover the active area of ​​the stage, so that the stage can always be illuminated within its adjustable range.

[0070] In some embodiments, Figure 2 and Figure 3 The eyepiece 72 is provided with a buckle 73 on the outside of the lens barrel, and the lighting assembly is provided with a block 41 that cooperates with the buckle. The buckle 73 and the block 41 cooperate with each other, so that the lighting assembly is detachably mounted on the lens assembly, and the user can install or remove it as needed.

[0071] In some embodiments, Figure 6 The stage is provided with a groove 61 for accommodating the graphite column synthesis block.

[0072] In some embodiments, the groove is a circular groove, and the depth of the groove is 0.5 cm.

[0073] In some embodiments, a plurality of anti-slip protrusions are disposed at the bottom of the groove. For example, the bottom of the groove may be made of a rough frosted material.

[0074] The circular groove in the present application can cooperate with the graphite block used in daily testing, and the rough structure design at the bottom of the groove can increase the friction between the graphite block and the stage (the groove) to ensure that the graphite block will not be displaced during the movement of the stage.

[0075] In some embodiments, the image acquisition device may be a mobile intelligent terminal, such as a smart phone. Specifically, the image acquisition device includes:

[0076] An image data acquisition module (e.g., a camera), used to acquire the image to be inspected of the cross section of the graphite column synthesis block from the production line in real time;

[0077] The detection module is used to use a pre-trained quality detection model to perform image processing on the image to be detected of the cross section of the graphite column synthesis block collected in real time from the production line by the image data acquisition module, so as to obtain the crystal distribution grade and crystal type grade of the graphite column synthesis block.

[0078] Furthermore, the image acquisition device also includes:

[0079] A database for storing equipment parameters corresponding to each crystal distribution grade and crystal form grade;

[0080] The equipment parameter matching module is used to match the crystal distribution grade and crystal type grade distribution detected by the detection module from the database to the corresponding production parameters, such as equipment parameters.

[0081] In some embodiments, the image acquisition device further comprises:

[0082] A training set construction module is used to collect cross-sectional images of graphite column synthesis blocks under various production parameters for constructing a training set; the cross-sectional images include complete cross-sectional images; the production parameters include equipment parameters; the equipment parameters further include power parameters and pressure parameters; preferably, the training set construction module obtains cross-sectional images of graphite column synthesis blocks under various production parameters from the above-mentioned image data acquisition module;

[0083] The model building module is used to input the manually annotated graphite column synthetic block cross-section image into the pre-built model for training, thereby obtaining a quality inspection model. Of course, in other embodiments, machine automatic annotation can also be used, and accordingly, the image acquisition device also includes a data annotation module.

[0084] In some embodiments, the image acquisition device further includes: a partition re-inspection module for determining whether the crystal distribution level of the detection module is lower than a threshold level, and if so, triggering the image data acquisition module to perform a partition re-inspection of the crystal level.

[0085] In summary, the present application proposes an image data acquisition system with a simple structure, easy operation, and flexible control of the acquisition angle, which effectively improves the image acquisition efficiency.

[0086] Example 2

[0001] Based on the above-mentioned graphite column synthesis block image acquisition system, the present invention also provides a graphite column synthesis block synthesis block quality detection method based on image processing, see Figure 8 , the method specifically comprises:

[0002] S1 collects cross-sectional images of graphite column synthetic blocks under various production parameters for constructing a training set; the cross-sectional images include complete cross-sectional images; the production parameters include equipment parameters, environmental parameters, and material parameters; the equipment parameters further include power parameters and pressure parameters.

[0003] In order to better collect the cross-sectional images of graphite column synthetic blocks, high-resolution cameras can be deployed in industrial environments to collect the cross-sectional images of graphite column synthetic blocks with high precision. Of course, in order to improve the robustness of the later model, images can be collected at multiple angles and distances.

[0004] In addition, when constructing the training set, graphite column synthetic blocks generated under various production parameters are covered as much as possible. In this embodiment, the so-called various production parameters include equipment parameters, environmental parameters and material parameters.

[0005] The most important equipment parameters are pressure and power parameters, so it is necessary to cover as many images of graphite column synthetic blocks produced under all pressures and temperatures that the equipment can reach as possible.

[0006] Environmental parameters mainly include the temperature and humidity of the production environment and other parameters that can affect the generation of graphite column synthetic blocks. Of course, under permitted conditions, images of graphite column synthetic blocks generated under various temperatures and humidities are collected as much as possible.

[0007] Material parameters mainly include the composition of the material. Raw materials may also fluctuate in content and other aspects, so it is necessary to collect images of graphite column synthetic blocks generated within the material fluctuation range.

[0008] In an ideal situation, all graphite column composite block images generated under various possible arrangements of equipment parameters, environmental parameters, and material parameters need to be collected. Of course, in actual production, we can only collect as many as possible.

[0009] S2 manually annotates the crystal distribution grade and crystal form grade in the cross-sectional image of the graphite column synthetic block.

[0010] In this embodiment, professionals mark the crystal distribution level and crystal type level in the cross-sectional image of the graphite column synthesis block to ensure the accuracy and consistency of the marking as much as possible. Of course, automatic marking by machine can also be used.

[0011] Because some backgrounds in the graphite block can be easily mistaken for crystals under light, in order to eliminate misjudgment, in this embodiment, when constructing the training set, background images of the graphite column synthetic block are collected and annotated as background as negative samples.

[0012] S3 inputs the annotated graphite column synthetic block cross-section image into the pre-built model for training, thereby obtaining a quality inspection model.

[0013] In this embodiment, before the image is input into the model, the image may be enhanced, specifically including:

[0014] 1. Resizing: Resize all images to a uniform resolution to ensure consistency of input data.

[0015] 2. Color conversion: Convert images to grayscale format to reduce the complexity of model training and speed up calculations.

[0016] 3. Sample expansion: Apply techniques such as random rotation, flipping, and scaling to enhance the model’s robustness to image deformations.

[0017] In this embodiment, the image classification model selected is based on fasterrcnn_resnet50_fpn, which combines ResNet-50 and Feature Graph Network (FPN) and is suitable for complex target detection tasks. ResNet-50 provides effective feature extraction capabilities, while FPN enhances the model's detection performance at multiple scales.

[0018] In order to cater to the quality inspection in the production process, that is, to perform grade quality inspection on both crystal distribution and crystal form, the classification head of the model is redefined for the above model in this embodiment. In fasterrcnn_resnet50_fpn, the original classifier is replaced with a new FastRCNNPredictor, whose output layer is configured to recognize 20 categories, namely 12 crystal form grades, 7 crystal distribution grades, and 1 background.

[0019] Specifically, the structure of the model is as follows:

[0020] Convolutional layer: A total of 50 convolutional networks, each layer contains multiple residual blocks to enhance learning ability and increase training depth.

[0021] Feature Pyramid Network (FPN): The structure contains a top-down path and lateral connections, which optimizes the fusion of multi-scale features.

[0022] ROI head: includes the region proposal network (RPN) and ROI alignment layer, followed by customized classification and regression layers.

[0023] Classifier: Based on the original Fast R-CNN predictor, it is expanded to 20 outputs (12 polymorph classes, 7 crystal distribution classes, and 1 background).

[0024] Activation function: Use the ReLU activation function for nonlinear transformation to ensure stability when the network depth increases.

[0025] The training process of the model in this implementation is as follows:

[0026] a. The model is deployed on appropriate hardware (GPU is preferred to accelerate computation). Images and objects (including bounding boxes and category labels) are loaded onto the same device for processing;

[0027] b. During the training process, batch processing is implemented. Each batch of images is forward propagated through the model, the loss between the predicted results and the actual labels is calculated, and then the weights are updated through backpropagation;

[0028] c. Use SGD optimizer for parameter optimization, set the initial learning rate, momentum and weight decay, and adjust the learning rate regularly to optimize the training process;

[0029] d. The loss function uses cross entropy loss, which is particularly suitable for multi-class classification problems.

[0030] e. Training parameters:

[0031] Training cycle: The model was optimized for 200 training cycles, and each cycle traversed the entire training set; Optimizer: The SGD optimizer was used, the initial learning rate was set to 0.005, and the learning rate was dynamically adjusted to optimize the convergence speed; Batch size: Each batch of data contained 16 images, which balanced the training speed and memory usage efficiency;

[0032] The deployment process of the model in this embodiment is as follows:

[0033] a. Performance evaluation:

[0034] During training, continuously monitor the model's loss and classification accuracy on the training set and validation set, and use indicators such as precision, recall, and F1 score to evaluate model performance;

[0035] b. Model optimization:

[0036] Adjust model parameters and training strategies based on performance feedback, such as increasing or decreasing the number of network layers, adjusting batch size or optimizer parameters when necessary;

[0037] Adjust model parameters and training strategies based on test results, such as modifying learning rate scheduling or adding data augmentation methods;

[0038] c. Actual deployment:

[0039] Deploy trained models in production environments to process images from production line cameras in real time.

[0040] Set up regular model maintenance and update cycles to accommodate changes that may occur during production.

[0041] The actual effects of this model are as follows:

[0042] a.Accuracy;

[0043] Overall Accuracy: The model achieved 93.5% accuracy on the independent test set.

[0044] Crystalline grade accuracy: The average accuracy was 95%, with the best category accuracy reaching 98% and the lowest at 92%.

[0045] Distribution level accuracy: The average accuracy is 90%, with the highest being 94% and the lowest being 85%.

[0046] b. Recall rate and F1 score;

[0047] Overall recall: 92%.

[0048] The overall F1 score is 92.5%, which shows that the model has a good balance between precision and recall.

[0049] c. Real-time performance;

[0050] In actual production, the model demonstrated efficient processing capabilities, processing and analyzing about 20 frames of images per second on average.

[0051] S4 collects the images to be inspected of the cross section of the graphite column synthetic block from the production line in real time and inputs them into the quality inspection model to obtain the crystal distribution grade and crystal type grade of the graphite column synthetic block and match the corresponding production parameters.

[0052] After predicting the crystal distribution grade and crystal form grade based on the image, the model will also match the production parameters of the graphite column synthesis block, so that the two can be combined to guide the subsequent parameter adjustment.

[0053] In the distribution level of crystals, the higher the level, the more uniform the crystal distribution. However, the distribution of crystals with lower levels will be locally clustered.

[0054] For images with densely distributed crystals, the effect of crystal grade detection is not very good. The reason is that when the crystals are too dense, they will interfere with each other during the image recognition and detection process, thus affecting the accuracy of the prediction.

[0055] Therefore, in order to improve the accuracy of crystal grade detection under low crystal distribution levels, in this implementation, if the crystal distribution level is lower than the threshold level, the crystal grade is re-inspected in different areas.

[0056] Generally speaking, in the detection above the crystal distribution threshold level (level 3 in this embodiment), the accuracy is acceptable, but once it is lower than level 3, the accuracy of the detection result will be unsatisfactory due to the above reasons.

[0057] In order to increase the speed of the entire system and reduce computing power requirements on the one hand, and to improve the reliability of detection on the other hand, in this embodiment, only images with crystal distribution levels below the threshold level are rechecked in partitions.

[0058] Specifically, the partition re-inspection method includes:

[0059] S401 partitions the acquired complete cross-sectional image to obtain partitioned images.

[0060] Specifically, the crystals in the complete cross-sectional image are segmented by instance to obtain the position of each crystal in the complete cross-sectional image; the image within the threshold range around each crystal is framed and cropped to obtain a partitioned image.

[0061] Since the detection of the crystal grade needs to avoid interference as much as possible, the complete crystal edge contour needs to be retained during the frame selection. In this embodiment, a threshold range is set for the crystal, and the threshold range includes the complete crystal image and a part of the background.

[0062] More specifically, the threshold range is a range enclosed by points that are at a threshold distance from the crystal edge. In simple terms, each pixel point at the crystal edge is extended outward by the threshold distance as the edge of the frame selection.

[0063] With this setting, the integrity of the crystal image can be ensured while the selection of adjacent crystals can be avoided as much as possible.

[0064] Of course, there may still be other crystals within the threshold range. In order to deal with this problem, in this embodiment, the distance between the closest points of the two crystals is calculated, and half of the distance is used as the threshold distance for frame selection. For example, the original threshold distance is 20 pixels. However, when framing, it is found that other crystals have invaded the image, so the distance between the closest points of the two crystals is calculated. If the distance is 10 pixels, then half of the 10 pixels, that is, 5 pixels, is used as the new threshold distance between the two crystals for re-frame selection and cropping, so that there is only one crystal in each partition image, which is convenient for detecting the crystal grade of each crystal.

[0065] S402 performs image enhancement processing on the partition image.

[0066] The image enhancement in this step is consistent with the image enhancement step in the main model and will not be repeated here.

[0067] S403 inputs the enhanced partition image into a pre-trained crystal quality detection sub-model to obtain the crystal grade after re-inspection.

[0068] In this step, the architecture of the crystal quality detection sub-model is different from the above quality detection model only in the output layer of the classifier, because it only detects the crystal grade, so the output layer is configured as 12 layers, corresponding to 12 crystal grades. The training set used during training is also the crystal partition image, rather than the complete image of the graphite synthesis block cross section.

[0069] This embodiment uses a pre-constructed quality detection model to perform real-time detection of the quality grade of the graphite column synthesis block, thereby avoiding the inconsistency of quality grade classification caused by different people performing manual quality grade classification, and the inconsistency caused by different quality grade classification standards of different manufacturers. At the same time, it also avoids reliance on manual experience and improves efficiency.

[0087] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0088] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A graphite column synthesis block image data acquisition system, characterized in that: The microscope comprises a main body, and the main body is provided with an eyepiece, an objective lens and a stage in order from top to bottom. The stage is provided with a groove for accommodating the graphite column synthesis block, and the stage is rotatably provided on the main body through a rotating mechanism; the main body is also provided with a knob for adjusting the height and angle of the objective lens and the eyepiece, and a mounting bracket for fixing the image acquisition device; The rotating mechanism comprises a fixed rod and a movable rod, wherein the first end of the fixed rod is connected to the main body, the second end of the fixed rod is provided with a rotating member, the first end of the movable rod is connected to the rotating member, and the second end of the movable rod is fixed to the loading platform; the width of the fixed rod gradually increases from the first end to the second end thereof, and a limit member is slidably provided on the fixed rod; When the loading platform is located at the initial position, the limiting member is located at the connection position between the fixed rod and the movable rod, so that the fixed rod and the movable rod are located on the same axis; when the loading platform is rotated, the loading platform drives the movable rod to rotate around the second end of the fixed rod, and the limiting member moves toward the first end of the fixed rod; Wherein, the image acquisition device specifically includes: An image data acquisition module is used to acquire images to be detected of the cross section of the graphite column synthesis block from the production line in real time; the images to be detected include complete images of the cross section; The detection module is used to detect the image to be detected using a pre-trained quality detection model to obtain the crystal distribution grade and crystal type grade of the graphite column synthetic block.

2. A graphite column synthesis block image data acquisition system according to claim 1, characterized in that Also includes: A training set construction module is used to construct a training set based on the collected cross-sectional images of graphite column synthesis blocks under various production parameters; the cross-sectional images include complete cross-sectional images; the production parameters include equipment parameters; the equipment parameters further include power parameters and pressure parameters; The model building module is used to input the annotated graphite column synthetic block cross-section image into a pre-built model for training, thereby obtaining a quality inspection model.

3. A graphite column synthesis block image data acquisition system according to claim 2, characterized in that Also included: a database for storing production parameters corresponding to each crystal distribution grade and crystal form grade; The parameter matching module is used to match the crystal distribution grade and crystal type grade distribution detected by the detection module from the database to the corresponding production parameters.

4. The graphite column synthesis block image data acquisition system according to claim 2, characterized in that Also includes: The partition re-inspection module is used to determine whether the crystal distribution level of the detection module is lower than the threshold level. If so, the image data acquisition module is triggered to perform a partition re-inspection on the crystal level.

5. The graphite column synthesis block image data acquisition system according to claim 1, characterized in that: The rotating mechanism further comprises an elastic member, and two ends of the elastic member are respectively connected to the limiting member and the movable rod; When the movable rod rotates, the limiting member moves toward the first end of the fixed rod; when the force acting on the movable rod disappears, the elastic member is reset to drive the movable member and the limiting member to reset.

6. The graphite column synthesis block image data acquisition system according to claim 1, characterized in that: The second end of the fixed rod is recessed toward the first end thereof to form a recessed portion, and correspondingly, the movable rod protrudes toward the first end thereof to form a protruding portion that cooperates with the recessed portion, and the protruding portion can rotate around the center point of the recessed portion.

7. The graphite column synthesis block image data acquisition system according to claim 6, characterized in that: The concave portion is V-shaped, and correspondingly, the convex portion is also V-shaped, and the opening of the concave portion is larger than the opening of the convex portion; or, the concave portion is arc-shaped, and correspondingly, the convex portion is also arc-shaped, and the curvature of the concave portion is larger than the curvature of the convex portion; When the movable rod is rotated to a maximum angle in the first direction, the first side of the protrusion abuts against the first side of the recess, and the acute angle between the central axis of the movable rod and the central axis of the fixed rod is 20°-45°.

8. The graphite column synthesis block image data acquisition system according to claim 7, characterized in that: The acute angle between the central axis of the movable rod and the central axis of the fixed rod is 30°.

9. The graphite column synthesis block image data acquisition system according to claim 1, characterized in that: The objective lens is connected with the eyepiece to form a lens assembly. The lens assembly is provided with a lighting assembly. The lighting assembly includes a shell, a lighting lamp arranged on the shell, and a switch for controlling the lighting lamp to be turned on and off.

10. The graphite column synthesis block image data acquisition system according to claim 1, characterized in that: A plurality of anti-slip protrusions are arranged at the bottom of the groove.

Images