Method for Establishing Adaptive Model Library, Material Identification Method and Material Detection Equipment

Abstract

The present invention provides a method for establishing an adaptive model library, a method for material identification, and a material detection device. The method includes obtaining material update data of a blanking system, where the material update data includes a material line drawing and material attribute information; converting the material line drawing into a primary target picture based on the shooting attributes of an image acquisition device and the material attribute information; performing area filling on the target area of the primary target picture, and performing binary image processing on the filled primary target picture to obtain a horizontal minimum circumscribed rectangle area of effective pixels; expanding and contracting the horizontal minimum circumscribed rectangle area based on a preset rule to obtain a training area of the primary target picture; performing model training on the training area of the primary target picture to obtain a material model and adding it to the adaptive model library, and automatically updating the adaptive model library by automatically synchronizing the data of the blanking system, effectively simplifying the update steps of the adaptive model library and improving the automation level.

Description

Technical Field

[0001] The present invention relates to the technical field of material identification, and particularly to a method for establishing an adaptive model library, a material identification method, and a material detection device. Background Art

[0002] For enterprises in the whole machine manufacturing industry, there are usually tens of thousands or hundreds of thousands of materials in the blanking workshop. With the increase and change of models, the number of materials will continue to increase. Since the materials are made of metal and generally accompanied by a certain degree of rust, the identification of materials is generally manually marked after cutting. After blanking, it is necessary to manually count and check the material information and then enter it into the Warehouse Management System (WMS), which requires a large amount of labor. At the same time, most of the materials in the blanking workshop are stacked together after one-time cutting of multiple pieces, but generally only one piece at the top of the stack is marked (if each material is manually marked, the workload is huge). The marked piece is prone to loss, resulting in the stagnation of unmarked materials. Over time, the accumulation of unidentifiable stagnant materials will become more and more, and the unidentifiable materials can only be treated as waste, resulting in huge waste. The method of manual identification is prone to missing marks or marking errors. If no secondary inventory is carried out to determine, it will also cause the stagnation and waste of mis-marked materials. Currently, for unmarked or mis-marked materials, the length, width, thickness and other information are usually measured with a caliper, and then manually searched from the Digital Integration Platform (DIP blanking system), which is time-consuming and laborious.

[0003] The method of identifying materials generally uses a charge coupled device (CCD camera) to take pictures of the physical object, and then matches and searches with an established material image model library. The establishment of the model library requires a physical picture of the material for training. However, the blanking system can generally only provide a line drawing converted from the CAD drawing of the material, and the size of the drawing is also quite different from the actual shooting size of the material, and the standard model cannot be directly established through this drawing. The common method is to use a CCD to take a real picture of each new material for standard model training and then store it in the library, and then wait to call the model for matching and searching when the same material is encountered next time; at the same time, if the materials with very small distinguishability are encountered, it is also impossible to quickly identify whether they are un-stored materials or stored materials. This method is suitable for the situation where the types of materials are not many.

[0004] However, in fact, for the huge number of materials in the whole machine blanking workshop, it is not always a process from scratch. Taking pictures of existing materials and storing them in the model library consumes a huge amount of labor costs and is inefficient. At the same time, when new materials are added, it is necessary to manually take pictures and store them, and the operation process is cumbersome and the degree of automation is low. Summary of the Invention

[0005] The present invention provides a method for establishing an adaptive model library, a method for material identification, and a material detection device, which are used to solve the defect of cumbersome identification when materials are updated in the prior art, realize automatic synchronization of data in the blanking system, and improve the automation level.

[0006] The present invention provides a method for establishing an adaptive model library, including:

[0007] Obtaining material update data of the blanking system, where the material update data includes a material line drawing and material attribute information;

[0008] Based on the shooting attributes of the image acquisition device and the material attribute information, converting the material line drawing into a primary target picture;

[0009] Performing area filling on the target area of the primary target picture, performing binary image processing on the filled primary target picture, and obtaining a horizontal minimum circumscribed rectangle area of effective pixels;

[0010] Based on a preset rule, expanding and shrinking the horizontal minimum circumscribed rectangle area to obtain a training area of the primary target picture;

[0011] Performing model training on the training area of the primary target picture, and adding the obtained material model to the adaptive model library.

[0012] According to the method for establishing an adaptive model library provided by the present invention, after converting the material line drawing into a primary target picture, it further includes:

[0013] Increasing the blank edge area of the primary target picture to obtain a secondary target picture as the primary target picture.

[0014] According to the method for establishing an adaptive model library provided by the present invention, the material attribute information includes: length information, width information, and thickness information;

[0015] The converting the material line drawing into a primary target picture based on the shooting attributes of the image acquisition device and the material attribute information includes:

[0016] Based on the shooting attributes of the image acquisition device, determining a conversion ratio corresponding to the image pixels captured by the image acquisition device under the thickness information and the actual size;

[0017] According to the size of the material line drawing, the length information, the width information, and the conversion ratio, converting the material line drawing into a primary target picture.

[0018] A method for establishing an adaptive model library provided by the present invention, the region filling of the target region of the primary target picture includes:

[0019] Perform Blob spot analysis on the primary target picture, and obtain the region with the second largest perimeter as the target region;

[0020] Perform region filling on the region with the second largest perimeter.

[0021] A method for establishing an adaptive model library provided by the present invention, after performing image processing on the primary target picture, further includes:

[0022] If the region with the second largest perimeter cannot be filled, obtain the region with the largest area as the target region;

[0023] Perform region filling on the region with the largest area.

[0024] The present invention also provides an apparatus for establishing an adaptive model library, including:

[0025] An acquisition module, configured to acquire material update data of a blanking system, where the material update data includes a material line drawing and material attribute information;

[0026] A conversion module, configured to convert the material line drawing into a primary target picture based on the shooting attributes of an image acquisition device and the material attribute information;

[0027] A filling module, configured to perform region filling on the target region of the primary target picture, perform binary image processing on the filled primary target picture, and obtain a horizontal minimum bounding rectangle region of effective pixels;

[0028] A scaling module, configured to scale the horizontal minimum bounding rectangle region based on a preset rule to obtain a training region of the primary target picture;

[0029] A training module, configured to perform model training on the training region of the primary target picture, and add the obtained material model to the adaptive model library.

[0030] The present invention also provides a method for material recognition, including:

[0031] Collect the material attribute information to be recognized;

[0032] Based on a target error parameter set and the material attribute information to be recognized, screen out a candidate material set in a material database;

[0033] Collect the image information of the material to be recognized, and convert the image information into a model to be detected;

[0034] Traverse each material model in the candidate material set in the adaptive model library, match each material model with the to-be-detected model respectively, and identify the target material. The adaptive model library is established by the method for establishing the adaptive model library described in any one of the above.

[0035] The present invention also provides a material identification device, including:

[0036] An acquisition module, configured to acquire to-be-identified material attribute information;

[0037] A screening module, configured to screen out a candidate material set in the material database based on the target error parameter set and the to-be-identified material attribute information;

[0038] An identification module, configured to acquire the image information of the to-be-identified material and convert the image information into a to-be-detected model; traverse each material model in the candidate material set in the adaptive model library, match each material model with the to-be-detected model respectively, and identify the target material. The adaptive model library is established by the method for establishing the adaptive model library described in any one of the above.

[0039] The present invention also provides a material detection device, including a device body and a control system;

[0040] The control system controls the device body to perform material identification by using the material identification method described in any one of the above.

[0041] According to a material detection device provided by the present invention, the device body includes: a material information acquisition component, a human-computer interaction component, and a detection component;

[0042] The human-computer interaction component is configured to control the material information acquisition component to acquire the material attribute information and image information of the to-be-identified material, and the to-be-identified material is placed on the detection component.

[0043] According to a material detection device provided by the present invention, the material information acquisition component includes an electronic scale, a ranging sensor, and an image acquisition device;

[0044] The electronic scale is configured to acquire the weight information of the to-be-identified material, the ranging sensor is configured to acquire the thickness information of the to-be-identified material, and the image acquisition device is configured to acquire the image information of the to-be-identified material.

[0045] The present invention also provides a data synchronization system, including: a blanking system, a data middle platform, a warehouse management system, and the material detection device described in any one of the above;

[0046] The data middle platform is configured to obtain the material update data of the blanking system and synchronize the material update data to the material detection device;

[0047] The warehouse management system is used to record the material identification results of the material detection equipment and the placement locations of the materials.

[0048] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the method for establishing the adaptive model library as described in any one of the above are implemented.

[0049] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for establishing the adaptive model library as described in any one of the above are implemented.

[0050] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the method for establishing the adaptive model library as described in any one of the above are implemented.

[0051] A method for establishing an adaptive model library, a material identification method, and a material detection device provided by the present invention obtain material update data of a blanking system, where the material update data includes a material line drawing and material attribute information; based on the shooting attributes of an image acquisition device and the material attribute information, convert the material line drawing into a primary target picture; perform area filling on the target area of the primary target picture, perform binary image processing on the filled primary target picture to obtain a horizontal minimum circumscribed rectangle area of effective pixels; based on a preset rule, expand and contract the horizontal minimum circumscribed rectangle area to obtain a training area of the primary target picture; perform model training on the training area of the primary target picture. If the training is successful, add the trained material model to the adaptive model library, and automatically update the adaptive model library by automatically synchronizing the data of the blanking system, without manually taking pictures of materials one by one to update the material model, effectively simplifying the update steps of the adaptive model library and improving the automation level. Description of the Drawings

[0052] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0053] Figure 1 is a schematic flowchart of the method for establishing the adaptive model library provided by the present invention;

[0054] Figure 2 is a schematic structural diagram of the device for establishing the adaptive model library provided by the present invention;

[0055] Figure 3 is a schematic flow chart of the material identification method provided by the present invention;

[0056] Figure 4 is a schematic structural diagram of the material identification device provided by the present invention;

[0057] Figure 5 is a front view of the material detection device provided by the present invention;

[0058] Figure 6 is a top view of the material detection device provided by the present invention;

[0059] Figure 7 is a schematic diagram of the working principle of the material detection device provided by the present invention;

[0060] Figure 8 is a schematic structural diagram of the data synchronization system provided by the present invention;

[0061] Figure 9 is a schematic diagram of the principle of the data synchronization system provided by the present invention;

[0062] Figure 10 is a schematic structural diagram of the electronic device provided by the present invention.

[0063] Reference numerals:

[0064] 1. Mobile detection device; 2. Electronic scale; 3. Distance measurement sensor; 4. CCD camera; 5. Industrial control computer; 6. Switch panel; 7. Printer; 8. Material identification module; 9. Input module; 10. Data processing module; 11. Image processing module; 12. Output module; 13. Weighing equipment parameter module; 14. Thickness measurement equipment parameter module; 15. Printing equipment parameter module; 16. Material database module; 17. Image parameter setting module; 18. Image preprocessing module; 19. Image matching module; 20. Image model library module; 21. Human-machine interaction interface; 22. IO acquisition card; 23. Start identification button; 24. Stop identification button; 25. Switch to the next line button; 26. Print button; 27. Cooling fan; 28. Connector base; 29. Light source; 30. Tempered glass stage; 31. Mouse and keyboard; 32. Monitor. Detailed implementation manners

[0065] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0066] The following will Figures 1 - 10 describe the method for establishing an adaptive model library, the method for identifying materials, and the material detection device of the present invention.

[0067] Figure 1 is a schematic flowchart of the method for establishing an adaptive model library provided by the present invention.

[0068] As Figure 1 shown, a method for establishing an adaptive model library provided by an embodiment of the present invention includes the following steps:

[0069] 101. Obtain the material update data of the blanking system, where the material update data includes the material line drawing and the material attribute information.

[0070] When there is a material update in the blanking system, the blanking system triggers the update of the latest data of the material through the interface. When there is a material update in the blanking system, all the information in the material will be updated, including the name, weight, size, line drawing, etc. of the updated material. The material update data is the data of the blanking system directly synchronized and can be directly read. The detailed process of the material update data in the blanking system can be manual input or other methods. The specific data update process in the blanking system will not be specifically described in this embodiment.

[0071] 102. Based on the shooting attributes of the image acquisition device and the material attribute information, convert the material line drawing into a primary target picture.

[0072] Specifically, the material attribute information includes length information, width information, thickness information, etc. The shooting attributes refer to the shooting attributes after adjustment of the image acquisition device, including information such as the current object distance, focal length, and field of view (FOV) size during shooting. For example, if the image acquisition device is a CCD camera, based on the shooting attributes of the image acquisition device and the material attribute information, the material line drawing is converted into a primary target picture. Specifically, based on the shooting attributes of the image acquisition device, the conversion ratio between the image pixels captured by the image acquisition device and the actual size corresponding to the thickness information is determined. Then, according to the size of the material line drawing, the length information, the width information, and the conversion ratio, combined with specific processing parameters (where the processing parameters are usually less than or equal to 1 and can be intelligently adjusted according to specific actual application scenarios), the material line drawing is converted into a primary target picture. For example, if the image acquisition device is a camera, the size of the picture captured by the camera is S, the object distance WD of the target platform without the material is D, the thickness of material a is H, and the length is L, then the object distance WDa of material a is D - H. Querying the field of view FOV corresponding to WDa is F1, then the size of the primary target image is Sa = S * (L / F1). For example, if the pixels of the target captured by the camera are 4000 * 3000, the object distance WD = 100 mm, and the thickness of material a is 8 mm, then the object distance of material a is 92 mm. Querying the actual field of view corresponding to the object distance of 92 mm is 400 * 300 mm, and the actual size of material a is 200 * 100 mm. Then the size of the picture to be converted is 4000 * (200 / 400) = 2000, and the size of the primary target picture is 2000 * 1500 pixels. To accelerate the processing speed, the pixels of the shooting target will be correspondingly reduced, such as reduced by 50%. Finally, the size of the primary target picture becomes 1000 * 750 pixels.

[0073] 103. Perform area filling on the target area of the primary target picture, and perform binary image processing on the filled primary target picture to obtain the horizontal minimum bounding rectangle area of the effective pixels.

[0074] After obtaining the primary target image, which is also a line drawing, first, the primary target image needs to be intelligently enlarged by a ratio p, where p > 0 is the enlargement ratio factor and can be intelligently adjusted according to the size of the primary target image. After intelligently enlarging the primary target image, the blank edge area of the primary target image is increased to obtain the secondary target image. Increasing the blank edge area ensures that the foreground effective pixels (black) of the material are surrounded by the blank background (white) (at least 3 white pixel points at the edge), otherwise, the edge information of the material may be lost during model training. The secondary target image with the increased blank edge area is used as the primary target image for subsequent image processing. After obtaining the primary target image, the primary target image needs to be filled. Since the primary target image is a line drawing, the area inside the line drawing needs to be filled. For example, for a cube line drawing, the area inside its line drawing is filled to make it closer to a three-dimensional drawing. Then, the filled primary target image is subjected to binary image processing to obtain the horizontal minimum bounding rectangle area of all region sets. Among them, binary image processing refers to converting the gray values of all pixels of the image to 0 or 255 through a dynamic threshold algorithm.

[0075] 104. Based on a preset rule, the horizontal minimum bounding rectangle area is scaled to obtain the training area of the primary target image.

[0076] After obtaining the horizontal minimum bounding rectangle, the horizontal minimum bounding rectangle area is scaled based on a preset rule. Among them, the preset rule can be to scale the horizontal minimum bounding rectangle area proportionally according to the actual required size. For example, if the horizontal minimum bounding rectangle area is 100×90 pixels, the preset rule can be to stretch or shrink from a certain right-angle point of the horizontal minimum bounding rectangle area so that the area size becomes 120×108 pixels, etc. Scaling refers to enlarging or shrinking, that is, intelligently adjusting the size according to the image size based on a certain ratio to obtain the training area of the primary target image.

[0077] 105. The training area of the primary target image is subjected to model training, and the material model is obtained and added to the adaptive model library.

[0078] After obtaining the training area of the primary target image, model training is performed on this training area. If the training is successful, the target model is stored in the library, so that the material model corresponding to the updated material is updated to the adaptive model library. If the training fails, the image is stored in the abnormal image processing folder. In addition, it should be noted that the scaling and filling in the above steps can also be exchanged in order, that is, it is also possible to scale first and then fill, but in this embodiment, the method of filling first and then scaling is preferred.

[0079] To ensure the comprehensiveness of the adaptive model library, the primary target images include the front and back photos of the material. That is, the established adaptive model library can comprehensively complete the identification of the material. In other words, when identifying the material, whether the material is placed face up or face down, the corresponding material model can be queried through the adaptive model library, thereby improving the material identification speed. It can also be understood that in the process of establishing the material model, the model corresponding to one side of the material is first established, and then the material is reversed, and the same process of establishing the material model is executed to ensure the comprehensiveness of the adaptive model library.

[0080] A method for establishing an adaptive model library provided in this embodiment includes: obtaining material update data of the blanking system, where the material update data includes a material line drawing and material attribute information; converting the material line drawing into a primary target image based on the shooting attributes of the image acquisition device and the material attribute information; performing region filling on the target region of the primary target image, performing binary image processing on the filled primary target image to obtain the horizontal minimum bounding rectangle region of the effective pixels; scaling the horizontal minimum bounding rectangle region based on a preset rule to obtain the training region of the primary target image; performing model training on the training region of the primary target image. If the training is successful, add the trained material model to the adaptive model library, and automatically update the adaptive model library by automatically synchronizing the data of the blanking system, without manually taking pictures of each material one by one to update the material model, effectively simplifying the update steps of the adaptive model library and improving the automation level.

[0081] Further, based on the above embodiment, the region filling of the target region of the primary target image in this embodiment includes: performing Blob spot analysis on the primary target image to obtain the region with the second largest perimeter as the target region. Blob refers to a connected region (also called a spot) in the image. Blob spot analysis is to perform connected component extraction and labeling on the binary image after foreground / background separation; perform region filling on the region with the second largest perimeter. Then judge whether the filling is accurately completed. If the filling of the region with the second largest perimeter cannot be completed, that is, if the region with the second largest perimeter cannot be filled, obtain the region with the largest area as the target region; perform region filling on the region with the largest area. That is to say, there are two filling methods. First, fill the region with the second largest perimeter. If the filling is successfully completed, perform image processing on the filled image. If the filling fails, select the region with the largest area to fill, so as to ensure that the filling can be accurately completed.

[0082] Based on the same general inventive concept, the present application also protects an apparatus for establishing an adaptive model library. The apparatus for establishing an adaptive model library provided by the present invention will be described below. The apparatus for establishing an adaptive model library described below can be correspondingly referred to the method for establishing an adaptive model library described above.

[0083] Figure 2 It is a schematic structural diagram of the device for establishing an adaptive model library provided by the present invention.

[0084] Such as Figure 2 shown, an apparatus for establishing an adaptive model library provided by an embodiment of the present invention includes:

[0085] An acquisition module 201, configured to acquire material update data of a blanking system, where the material update data includes a material line drawing and material attribute information;

[0086] A conversion module 202, configured to convert the material line drawing into a primary target picture based on the shooting attributes of an image acquisition device and the material attribute information;

[0087] A filling module 203, configured to perform area filling on a target area of the primary target picture, perform binary image processing on the filled primary target picture, and obtain a horizontal minimum circumscribed rectangle area of effective pixels;

[0088] A scaling module 204, configured to scale the horizontal minimum circumscribed rectangle area based on a preset rule to obtain a training area of the primary target picture;

[0089] A training module 205, configured to perform model training on the training area of the primary target picture, and add the obtained material model to the adaptive model library.

[0090] An apparatus for establishing an adaptive model library provided by the present invention acquires material update data of a blanking system, where the material update data includes a material line drawing and material attribute information; converts the material line drawing into a primary target picture based on the shooting attributes of an image acquisition device and the material attribute information; performs area filling on a target area of the primary target picture, performs binary image processing on the filled primary target picture, and obtains a horizontal minimum circumscribed rectangle area of effective pixels; scales the horizontal minimum circumscribed rectangle area based on a preset rule to obtain a training area of the primary target picture; performs model training on the training area of the primary target picture, and if the training is successful, adds the trained material model to the adaptive model library, automatically updates the adaptive model library by automatically synchronizing the data of the blanking system, and does not need to manually shoot each material one by one to update the material model, effectively simplifying the update steps of the adaptive model library and improving the automation level.

[0091] Further, the filling module 203 in this embodiment is further configured to:

[0092] Increase the blank edge area of the primary target picture to obtain a secondary target picture as the primary target picture.

[0093] Further, in this embodiment, the material attribute information includes: length information, width information, and thickness information; the conversion module 202 is further configured to:

[0094] Based on the shooting attributes of the image acquisition device, determine the conversion ratio between the image pixels captured by the image acquisition device corresponding to the thickness information and the actual size;

[0095] According to the size of the material line drawing, the length information, the width information, and the conversion ratio, convert the material line drawing into a primary target picture.

[0096] Further, the filling module 203 in this embodiment is further configured to:

[0097] Perform Blob spot analysis on the primary target picture, and obtain the region with the second largest perimeter as the target region;

[0098] Perform region filling on the region with the second largest perimeter.

[0099] Further, the filling module 203 in this embodiment is further configured to:

[0100] If the region with the second largest perimeter cannot be filled, obtain the region with the largest area as the target region;

[0101] Perform region filling on the region with the largest area.

[0102] Further, the primary target picture in this embodiment includes: a front photo of the material and / or a back photo of the material.

[0103] Based on the same general inventive concept, the present invention also protects a material recognition method.

[0104] Figure 3 It is a schematic flowchart of the material recognition method provided by the present invention.

[0105] As Figure 3 shown, a material recognition method provided by an embodiment of the present invention includes the following steps:

[0106] 301. Collect the material attribute information to be recognized.

[0107] Specifically, the material to be identified is defined as the material to be recognized. After obtaining the material to be recognized, first obtain the attribute information of the material to be recognized. The attribute information includes weight information, perimeter information, area information, thickness information, etc. The weight information can be obtained through the weighing device of the material detection equipment, and the thickness information can be obtained through the distance sensor. The difference in distance between the two states before and after placing the material to be recognized is used as the thickness of the material to be recognized. The perimeter and area can be collected by obtaining the dimension information of the material to be recognized and then calculating the perimeter and area of the material to be recognized.

[0108] 302. Based on the target error parameter set and the attribute information of the material to be recognized, screen out the candidate material set in the material database.

[0109] When performing material identification, first determine the error parameter set. The error parameter set includes weight error, thickness error, perimeter error, area error, etc. Then, based on the weight error, thickness error, perimeter error, and area error, screen out the candidate material set in the material data. The candidate material set includes all candidate materials that meet the weight, thickness, perimeter, and area of the material to be recognized. For example, initialize the error parameter set K = K0 (including the upper and lower limit errors of weight kw1, kw2; the upper and lower limit errors of perimeter kp1, kp2; the upper and lower limit errors of area ka1, ka2). After obtaining the weight W, thickness H, perimeter P, area A, and the error parameter set K0, then obtain the candidate material set Q1 that meets the error conditions {[W * kw1, W * kw2], [H, H], [P * kp1, P * kp2], [A * ka1, A * ka2]} from the material database. For example, if the error is 0.01%, then the error range of the weight in the candidate material set is W * 0.01%, that is, the final weight range is W - W * 0.01% - -W + W * 0.01%. The others are similar and will not be explained one by one.

[0110] Among them, the initial target error parameter set is only initialized once when the software is opened for the first time and has no value (the error parameter set is a set of upper and lower limit parameters of errors of parameters such as weight, thickness, area, etc. Initially, to ensure that the target can be found, a relatively large value such as 30% will be sacrificed in terms of search time). The target error parameter set refers to the error set after intelligent analysis and adjustment of the actual error parameter data of the accurate search results after a certain number (such as 50, 100, 150 times, etc.) of searches (that is, for example, when searching according to the ±30% error parameter of weight, the difference between the target found and the weight of the photographed target is only 3%, and 3% is the actual error parameter data). For example, based on the data of 150 search results, a total of 3 adjustments have been made. After adjustment, the upper limit of the error weight is 5%, the lower limit is 8%, the upper limit of the area is 12%, and the lower limit of the area is 4%, etc.

[0111] 303. Collect the image information of the material to be recognized, and convert the image information into a model to be detected.

[0112] After obtaining the candidate material set, collect the image information of the material to be recognized. For example, the image information of the material to be recognized can be obtained through a CCD camera, and then image enhancement algorithm processing is performed. The image information of the material to be recognized is scaled according to a preset ratio, and the image information is converted into a recognition model of the material to be recognized. The specific conversion method is to obtain a new picture after expanding or shrinking according to the scaling ratio established by the adaptive model library (such as shrinking by 50%) after collecting the image information. Perform intelligent Blob analysis on this picture to obtain the actual range of the target as the search area. The new target picture and the search area are collectively referred to as the recognition model.

[0113] 304. Traverse each material model in the candidate material set in the adaptive model library, and match each material model with the model to be detected respectively to identify the target material.

[0114] After obtaining the model to be detected of the material to be recognized, search for the target model that matches the model to be detected in the adaptive model library. After finding the target model, the corresponding target material is obtained, thereby completing the accurate recognition of the material to be detected, where the adaptive model library is established by the method for establishing the adaptive model library in any of the above embodiments.

[0115] A material recognition method provided in this embodiment can effectively improve the speed of material recognition by first performing a preliminary screening in the material data based on information such as the weight, thickness, perimeter, and area of the material to be recognized to obtain a candidate material set, and then performing individual matching and recognition from the candidate material set. At the same time, due to the data consistency between the adaptive model library and the blanking system, it can accurately ensure that each material to be detected can be recognized, improving the accuracy and efficiency of material recognition.

[0116] Further, on the basis of the above embodiments, after identifying the target material in this embodiment, the following steps are also included: After the detection is completed, the target material attribute information is compared with the material attribute information to be identified, and then the maximum error parameter value of the target material is found according to the information of the material to be identified as the actual error parameter, and the maximum actual error parameter is added to the error parameter set. The initialization error parameter set is dynamically adjusted according to the size of the data in the error parameter set and the matching speed of each material model with the model to be detected respectively to obtain the target error parameter set. The dynamic adjustment of the error parameter set includes adjusting the initialization error parameter set through a preset rule based on the size of the data in the error parameter set and the matching speed of each material model with the model to be detected respectively, that is, adjusting the value corresponding to each error parameter in the error parameter set. Usually, it is to choose to expand the error parameter. For example, the error parameter K is expanded, where K = K * t (t ≥ 1.5, adjusted according to the actual situation), so as to appropriately expand the error parameter. It should be noted that in practical applications, it may not necessarily be an expansion, but may also be a reduction. If the target material can be accurately identified according to the target error parameter set, but the identification speed is slow, the error parameter is reduced in a certain proportion to reduce the amount of data in the candidate material set and improve the model matching speed. If the target cannot be accurately identified according to the target error parameter, it indicates that the target material is not included in the candidate material set, and the error parameter needs to be adjusted to increase the number of the candidate material set, so as to ensure that the target material can be accurately identified and the material identification speed can be improved.

[0117] Specifically, the dynamic adjustment of the error parameter set includes: obtaining the number of data in the error parameter set; when the number of data meets a preset condition, obtaining the error parameter set corresponding to the material quantity in the material database being the first preset ratio of all the data quantities; based on a preset ratio parameter, scaling the error parameter set that meets the first preset ratio until the error parameter set meets the target preset ratio, and using it as the target error parameter set. For example, the error parameter is adjusted every 50 data, that is, the preset condition is that the error parameter is adjusted every time the quantity in the database increases by 50. The specific adjustment method is to first obtain the total number of all data in the material database, then obtain the error parameter set Kmax corresponding to the current total number of all data, and then query the error parameter set Kf corresponding to half of the total number of all data, that is, the first preset ratio is 50%. Then expand Kf step by step until Kf can meet the condition of querying S% of the data in the material database. The specific size of S can be flexibly adjusted according to the actual situation and can usually be selected between 70 and 100. Thus, by adjusting the number of data triggering the error parameter adjustment and the error range value of the error parameter set, the dynamic adjustment of the error parameter set is carried out, and combined with the number of data in the material database, the error parameter set is dynamically adjusted, so as to ensure both the speed of material identification and the accuracy of material identification.

[0118] Based on the same general inventive concept, the present application also protects a material identification device.

[0119] Figure 4 It is a schematic structural diagram of the material identification device provided by the present invention.

[0120] As Figure 4 shown, a material identification device provided by an embodiment of the present invention includes:

[0121] A collection module 401, configured to collect attribute information of the material to be identified;

[0122] A screening module 402, configured to screen out a candidate material set in the material database based on the target error parameter set and the attribute information of the material to be identified;

[0123] An identification module 403, configured to collect image information of the material to be identified and convert the image information into a model to be detected; traverse each material model in the candidate material set in the adaptive model library, and respectively match each material model with the model to be detected to identify the target material, and the adaptive model library is established by the method for establishing the adaptive model library in any of the above embodiments.

[0124] Further, on the basis of the above embodiment, an error adjustment module is further included in this embodiment, and is used for:

[0125] Dynamically adjust the initialization error parameter set according to the size in the candidate material set and the matching speed of each material model with the model to be detected, so as to obtain the target error parameter set.

[0126] Further, based on the above embodiments, the error adjustment module in this embodiment is further configured to:

[0127] Adjust the initialization error parameter set through a preset rule based on the size of the candidate material set and the matching speed of each material model with the model to be detected.

[0128] Further, based on the above embodiments, the error adjustment module in this embodiment is further configured to:

[0129] Obtain the number of data in the material database;

[0130] When the number of data meets a preset condition, obtain the error parameter set corresponding to the number of materials in the material database being the first preset ratio of all the number of data;

[0131] Based on a preset ratio parameter, scale the error parameter set when it meets the first preset ratio until the error parameter set meets the target preset ratio, and use it as the target error parameter set.

[0132] Further, the error parameter set in this embodiment includes at least one of weight error, perimeter error, and area error.

[0133] Based on the same general inventive concept, the present application also protects a material detection device.

[0134] Figure 5 is the front view of the material detection device provided by the embodiment of the present invention, Figure 6 is the top view of the material detection device provided by the embodiment of the present invention, Figure 7 is the schematic diagram of the working principle of the material detection device provided by the present invention.

[0135] As Figure 5 、 Figure 6 and Figure 7 shown, a material detection device provided by an embodiment of the present invention includes a device body and a control system. The control system controls the device body to perform material identification by using the material identification method in any of the above embodiments.

[0136] Furthermore, the device body in this embodiment includes: a material information acquisition component, a human-computer interaction component, and a detection component; the human-computer interaction component is used to control the material information acquisition component to collect the weight information, thickness information, and image information of the material to be identified, and the material to be identified is placed on the detection component. Among them, the material information acquisition component includes an electronic scale 2, a ranging sensor 3, and an image acquisition device. The image acquisition device can be a CCD camera 4. The electronic scale 2 is used to collect the weight information of the material to be identified, the ranging sensor 3 is used to collect the thickness information of the material to be identified, and the image acquisition device is used to collect the image information of the material to be identified. The human-computer interaction component includes a display 32 and operation buttons; the operation buttons are used to control the start and stop of the material information acquisition component, the switching and printing of the result set. The operation buttons can include a start identification button 23, a stop identification button 24, a switch to the next line button 25, a printing button 26, etc. The display 32 is used to display the results of material detection. The detection component includes: a loading platform 30 and a light source 29; the loading platform 30 is used to place the material to be identified, and the light source 29 is used to irradiate the material to be identified on the loading platform 30. The control system includes: a material database module 16, an image model library module 20, an image processing module 11, and a data processing module 10; the material database module 16 is used to synchronize all material data of the blanking system; the image model library module 20 is used to store all image models corresponding to all material data; the image processing module 11 is used to convert the image information of the material to be identified into a to-be-identified model; the data processing module 10 is used to find a target material model that matches the to-be-identified model among all image models to identify the target material.

[0137] For example, taking the industrial control computer 5 as the entire control system, the industrial control computer 5 integrates a material identification module 8. The material identification module 8 includes an input module 9 for inputting the collected switch signals, weight data, thickness data, and image data, a data processing module 10 for processing the input switch signals, weight data, and thickness data, an image processing module 11 for processing the image data, and a result output module 12 for outputting the processed results. The data processing module 10 includes a weighing device parameter module 13 for setting the parameters of the electronic scale 2, a thickness measuring device parameter module 14 for setting the parameters of the distance measuring sensor 3, a printing device parameter module 15 for setting the parameters of the printer 7, and a material database module 16 for querying the data. The image processing module 11 includes an image parameter setting module 17 for processing the image, an image preprocessing module 18 for preprocessing the image data, an image matching module 19 for matching the image, and a material model library module for processing the drawing data. The weighing device parameter module 13, the thickness measuring device parameter module 14, the printing device parameter module 15, the image parameter setting module 17, and the image matching module 19 are also connected to a human-machine interface 21. The industrial control computer 5 is provided with an I / O interface for connecting the electronic scale 2 to collect the weight data. The industrial control computer 5 is connected to the distance measuring sensor 3 through an IO acquisition card 22 to read the thickness information. The industrial control computer 5 is connected to the switch panel 6 through an IO acquisition card 22 to obtain the input information of the start identification button 23, the stop identification button 24, the switch to the next line button 25, and the printing button 26. The industrial control computer 5 is connected to a cooling fan 27 installed on the side of the mobile loading platform 30. The electronic scale 2 is provided with a connecting piece base 28. The connecting piece base 28 is provided with a backlight source 29. The light source 29 is connected to a light source adjustment controller. The connecting piece base 28 is provided with a tempered glass loading platform 30 for carrying the material. The human-machine interface 21 is a mouse and keyboard 31 for input and a display 32 for displaying information.

[0138] Specifically, the material detection device mainly consists of a detection table, an industrial control computer 5, an IO acquisition card 22, an electronic scale 2, a light source 29, a ranging sensor 3, a switch panel 6, a printer 7, an industrial CCD camera 4, etc. It is used for the identification and marking of unlabeled materials or the verification of labeled materials in the blanking workshop. The material detection device is used in conjunction with hardware and can be designed to run on different operating systems. Before automatically identifying the material, the data center synchronizes the latest material information obtained through the connected blanking system, intelligently processes the material line drawing through the intelligent model training algorithm and puts it into the adaptive model library. The material detection platform synchronizes and updates the material basic information library and the adaptive model library through the data center. When the material detection device is running, the operator places the material to be identified on the tempered glass carrier 30 of the mobile detection device 1, presses the start button on the switch panel 6, and the system starts to detect the material thickness and weight, and enables the backlight source 29 with uniform light. The data processing module 10 inside the industrial control computer 5 obtains the weight and thickness information, and obtains a list of candidate material information that meets the weight and thickness from the material database module 16. At the same time, the CCD camera 4 is triggered to collect image data. The image processing module 11 obtains the image data, enhances and compresses the image through the image preprocessing module 18, traverses the list of candidate material information, finds the corresponding image model in the image model library and matches it with the preprocessed image to obtain a list of matching results. The list of matching results is displayed on the monitor 32 in descending order according to the best similarity weight value, where the best similarity weight value is the weighted average of the similarity and the scaling ratio. Further, operating the switch to the next line 25 can switch the matching result selection, and operating the print button 26 can print the material information according to the customized print format and paste it onto the material to complete the whole process of identifying and marking the material.

[0139] The material detection device of this embodiment comprehensively processes the information of the material, combines the ranging sensor 3, the electronic scale 2, and the industrial CCD camera 4 to automatically identify and detect the material to be identified, performs efficient automated processing, saves a lot of time, and meets the high standards of automation, intelligence, and unmanned operation in the smart factory; it has a fast recognition speed, a high recognition rate, and a wide range of adaptability, and can be applied to any flat material type for laser cutting in the blanking workshop. And it adopts a friendly human-machine interaction interface 21, and the recognition results are displayed on the monitor 32 in descending order according to the best similarity weight value, which is convenient for the user to see the results. The system has a switch control panel button and a printer 7 for printing, and has convenient operation functions such as one-key identification, one-key result switching, and one-key printing to quickly label the material.

[0140] Based on the same general inventive concept, the present application also protects a data synchronization system.

[0141] Figure 8 It is a schematic structural diagram of the data synchronization system provided by the present invention;

[0142] As shown Figure 8 in the figure, a data synchronization system provided in this embodiment includes: a blanking system, a data center, a warehouse management system, and a material detection device as described in any of the above embodiments; the data center is used to obtain the material update data of the blanking system and synchronize the material update data to the material detection device; the warehouse management system is used to record the material identification results of the material detection device and the placement positions of the materials.

[0143] Specifically, the blanking system, the warehouse management system, the data center, and each material detection platform are connected through an internal dedicated high-speed Ethernet channel. When there is a material update in the blanking system, the blanking system triggers the update of the material information and the line drawing to the data center through the interface; after receiving the information transmitted by the blanking system, the data center synchronizes and updates or adds the basic material information to the material information database, and intelligently processes the material line drawing into the required target model through the intelligent model training algorithm and stores it in the adaptive model library. After the update is completed, the data update interface of the material detection platform system is triggered; the material detection platform system synchronizes and updates the basic material information and the adaptive model library of the data center to the local. After receiving the identification signal, the material detection platform screens out the candidate material information from the material information database through information such as the length, width, thickness, area, and perimeter of the material. Further, the target is accurately located from the material adaptive model library. After the target is located, information such as the material placement position is manually input, and the update signal is triggered and the material-related information is uploaded to the warehouse management system; the warehouse management system receives the material update signal and obtains the material-related information from the interface to update the data related to the material inventory.

[0144] Figure 9 It is a schematic diagram of the principle of the data synchronization system provided by the present invention.

[0145] As shown Figure 9As shown in the figure, the basic data preparation is the content completed by the blanking system, that is, to update the material information of the blanking system in real time, including the system material line diagram. Data synchronization processing refers to the data center synchronizing the material update information in the blanking system, storing the updated material basic information in the material information library after synchronization, and at the same time performing operations such as scaling, filling, and flipping on the material line diagram in the blanking system to generate a standard template and adding it to the adaptive model library. Material identification refers to the material detection platform end. The material detection equipment at the material detection platform end updates the material information library and the adaptive model library from the data center, and then performs identification and detection on the material to be identified through the material detection platform end. After successful identification, information such as the storage location is entered and uploaded to the warehouse management system. The final data upload refers to the operation performed at the warehouse management system end, that is, updating the material inventory information in the warehouse system. The material basic information is automatically synchronized with the blanking system through the data center server, and the intelligent model training algorithm is used to perform image processing on the material line diagram to generate an intelligent model for training and storage in the library; after synchronizing the data between the material detection platform and the data center, material identification is performed, and the material detection platform automatically connects the identified information to the warehouse management system to update the material inventory information, connecting the upstream and downstream systems and realizing intelligent and automated integration.

[0146] Figure 10 It is a schematic structural diagram of the electronic device provided by the present invention.

[0147] As Figure 10 shown, the electronic device may include: a processor 1010, a communications interface 1020, a memory 1030, and a communication bus 1040. Among them, the processor 1010, the communication interface 1020, and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 can call the logical instructions in the memory 1030 to execute the method for establishing an adaptive model library, and the method includes: obtaining the material update data of the blanking system, where the material update data includes a material line diagram and material attribute information; converting the material line diagram into a primary target picture based on the shooting attributes of the image acquisition device and the material attribute information; performing area filling on the target area of the primary target picture, and performing binary image processing on the filled primary target picture to obtain the horizontal minimum circumscribed rectangle area of the effective pixels; scaling the horizontal minimum circumscribed rectangle area based on a preset rule to obtain the training area of the primary target picture; performing model training on the training area of the primary target picture to obtain a material model and adding it to the adaptive model library.

[0148] In addition, when the logical instructions in the above-mentioned memory 1030 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0149] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for establishing an adaptive model library provided by the above-mentioned various methods. The method includes: obtaining material update data of the blanking system, where the material update data includes a material line drawing and material attribute information; converting the material line drawing into a primary target picture based on the shooting attributes of the image acquisition device and the material attribute information; performing region filling on the target region of the primary target picture, performing binary image processing on the filled primary target picture to obtain a horizontal minimum circumscribed rectangle region of effective pixels; scaling the horizontal minimum circumscribed rectangle region based on a preset rule to obtain a training region of the primary target picture; performing model training on the training region of the primary target picture to obtain a material model and adding it to the adaptive model library.

[0150] In yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the method for establishing an adaptive model library provided by the above-mentioned various methods. The method includes: obtaining material update data of the blanking system, where the material update data includes a material line drawing and material attribute information; converting the material line drawing into a primary target picture based on the shooting attributes of the image acquisition device and the material attribute information; performing region filling on the target region of the primary target picture, performing binary image processing on the filled primary target picture to obtain a horizontal minimum circumscribed rectangle region of effective pixels; scaling the horizontal minimum circumscribed rectangle region based on a preset rule to obtain a training region of the primary target picture; performing model training on the training region of the primary target picture to obtain a material model and adding it to the adaptive model library.

[0151] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.

[0152] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments or equivalently replace some of the technical features. However, these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for establishing an adaptive model library, characterized in that Including: Obtain the material update data of the blanking system, where the material update data includes a material line drawing and material attribute information, and the material attribute information includes: length information, width information, and thickness information; Based on the shooting attributes of the image acquisition device and the material attribute information, convert the material line drawing into a primary target picture, including: Based on the shooting attributes of the image acquisition device, determine the conversion ratio between the image pixels captured by the image acquisition device corresponding to the thickness information and the actual size; According to the size of the material line drawing, the length information, the width information, and the conversion ratio, convert the material line drawing into a primary target picture; Perform area filling on the target area of the primary target picture, including: Perform Blob spot analysis on the primary target picture to obtain the area with the second-largest perimeter as the target area; Perform area filling on the area with the second-largest perimeter; If the area with the second-largest perimeter cannot be filled, obtain the area with the largest area as the target area; Perform area filling on the area with the largest area; Perform binary image processing on the filled primary target picture to obtain the horizontal minimum bounding rectangle area of the valid pixels; Based on a preset rule, scale the horizontal minimum bounding rectangle area to obtain the training area of the primary target picture; Perform model training on the training area of the primary target picture, and add the material model to the adaptive model library.

2. The method for establishing an adaptive model library according to claim 1, wherein After converting the material line drawing into a primary target picture, it further includes: Increase the blank edge area of the primary target picture to obtain a secondary target picture as the primary target picture.

3. An apparatus for establishing an adaptive model library, characterized in that Including: An acquisition module for obtaining the material update data of the blanking system, where the material update data includes a material line drawing and material attribute information, and the material attribute information includes: length information, width information, and thickness information; A conversion module for converting the material line drawing into a primary target picture based on the shooting attributes of the image acquisition device and the material attribute information, including: Based on the shooting attributes of the image acquisition device, determine the conversion ratio between the image pixels captured by the image acquisition device corresponding to the thickness information and the actual size; According to the size of the material line drawing, the length information, the width information, and the conversion ratio, convert the material line drawing into a primary target picture; A filling module for performing area filling on the target area of the primary target picture, including: Perform Blob spot analysis on the primary target picture to obtain the area with the second-largest perimeter as the target area; Perform area filling on the area with the second-largest perimeter; If the area with the second-largest perimeter cannot be filled, obtain the area with the largest area as the target area; Perform area filling on the area with the largest area; Perform binary image processing on the filled primary target picture to obtain the horizontal minimum bounding rectangle area of the valid pixels; A scaling module for scaling the horizontal minimum bounding rectangle area based on a preset rule to obtain the training area of the primary target picture; A training module for performing model training on the training area of the primary target image, and adding the obtained material model to the adaptive model library.

4. A material identification method, characterized in that, It includes: Collecting the attribute information of the material to be recognized, where the attribute information includes weight information, perimeter information, area information, and thickness information; The weight information is obtained through a weighing device; The thickness information is obtained through a distance measuring sensor; the perimeter information and the area information are obtained through material size calculation; Based on the target error parameter set and the attribute information of the material to be recognized, screening out a candidate material set in the material database; Collecting the image information of the material to be recognized through an image acquisition device, performing image enhancement algorithm processing on the image information, and scaling it according to a preset ratio to be converted into a model to be detected; Traversing each material model in the candidate material set in the adaptive model library, respectively matching each material model with the model to be detected, and identifying the target material, where the adaptive model library is established by the method for establishing the adaptive model library according to any one of claims 1 to 2; Dynamically adjusting the initial error parameter set according to the size of the candidate material set and the matching speed of each material model respectively matching with the model to be detected, to obtain the target error parameter set.

5. A material identification device, characterized in that, It includes: A collection module for collecting the attribute information of the material to be recognized, where the attribute information includes weight information, perimeter information, area information, and thickness information; The weight information is obtained through a weighing device; the thickness information is obtained through a distance measuring sensor; the perimeter information and the area information are obtained through material size calculation; A screening module for screening out a candidate material set in the material database based on the target error parameter set and the attribute information of the material to be recognized; An identification module for collecting the image information of the material to be recognized through an image acquisition device, performing image enhancement algorithm processing on the image information, and scaling it according to a preset ratio to be converted into a model to be detected; traversing each material model in the candidate material set in the adaptive model library, respectively matching each material model with the model to be detected, and identifying the target material, where the adaptive model library is established by the method for establishing the adaptive model library according to any one of claims 1 to 2; An error adjustment module for dynamically adjusting the initial error parameter set according to the size of the candidate material set and the matching speed of each material model respectively matching with the model to be detected, to obtain the target error parameter set.

6. A material detection device, characterized in that, It includes a device body and a control system; The device body includes: A material information collection component, including a weighing device, a distance measuring sensor, and an image acquisition device, for collecting the weight, thickness, and image information of the material to be recognized; A human-computer interaction component, including a display and operation buttons, for displaying the recognition result and receiving user operations; A detection component, including a carrier platform and a light source, for placing the material to be detected and providing illumination; The control system controls the device body to perform material recognition by using the material recognition method according to claim 4.

7. A data synchronization system, characterized in that, It includes: A blanking system, a data center, a warehouse management system, and the material detection device as described in claim 6; The blanking system is used to provide material update data; The data center is used to obtain the material update data of the blanking system and synchronize the material update data to the material detection device; The warehouse management system is used to record the material identification result of the material detection device and the placement location of the material; The material detection device is used to identify materials and update material information.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for establishing the adaptive model library as described in any one of claims 1 to 2.

Images