CN117689902A - Forging processing technology of ball cage - Google Patents
Forging processing technology of ball cage Download PDFInfo
- Publication number
- CN117689902A CN117689902A CN202410128698.9A CN202410128698A CN117689902A CN 117689902 A CN117689902 A CN 117689902A CN 202410128698 A CN202410128698 A CN 202410128698A CN 117689902 A CN117689902 A CN 117689902A
- Authority
- CN
- China
- Prior art keywords
- ball cage
- feature map
- feature
- training
- cage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000005242 forging Methods 0.000 title claims abstract description 60
- 238000012545 processing Methods 0.000 title abstract description 24
- 238000005516 engineering process Methods 0.000 title abstract description 13
- 239000013598 vector Substances 0.000 claims abstract description 107
- 230000004927 fusion Effects 0.000 claims abstract description 99
- 238000000034 method Methods 0.000 claims abstract description 58
- 230000008569 process Effects 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims description 92
- 238000001514 detection method Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 15
- 238000010438 heat treatment Methods 0.000 claims description 13
- 238000000605 extraction Methods 0.000 claims description 10
- 230000009467 reduction Effects 0.000 claims description 10
- 238000010586 diagram Methods 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000003825 pressing Methods 0.000 claims description 8
- 238000001816 cooling Methods 0.000 claims description 7
- 238000003754 machining Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 238000004140 cleaning Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000001303 quality assessment method Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 8
- 239000000463 material Substances 0.000 description 7
- 238000013136 deep learning model Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 230000006872 improvement Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 239000011248 coating agent Substances 0.000 description 3
- 238000000576 coating method Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 239000007769 metal material Substances 0.000 description 3
- 238000013441 quality evaluation Methods 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 2
- 239000000956 alloy Substances 0.000 description 2
- 229910045601 alloy Inorganic materials 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000013604 expression vector Substances 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000003801 milling Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003647 oxidation Effects 0.000 description 1
- 238000007254 oxidation reaction Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000004381 surface treatment Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000007514 turning Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a forging processing technology of a ball cage, which acquires a ball cage surface state image of the ball cage to be detected, which is acquired by a camera; extracting features of the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map; fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector; and determining whether the to-be-detected ball cage is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage. Therefore, an efficient quality assessment model can be established, so that automation of the assessment process of the quality of the ball cage is realized, manual intervention is reduced, and efficiency is improved.
Description
Technical Field
The invention relates to the technical field of intelligent forging processing, in particular to a forging processing technology of a ball cage.
Background
The ball cage is an important part for the steering system of the automobile, and can realize the rotation and transmission of the steering shaft. The forging process of the cage is a key factor affecting its performance and life, and therefore, there is a continuous need for optimization and improvement.
The existing ball cage forging processing technology has some problems, such as serious die abrasion, too high cooling speed and the like, so that the ball cage has defects in the aspects of dimensional accuracy, surface quality, mechanical property, wear resistance and the like.
Therefore, an optimized forging process for the ball cage is desired.
Disclosure of Invention
The embodiment of the invention provides a forging processing technology of a ball cage, which is used for acquiring a ball cage surface state image of the ball cage to be detected, which is acquired by a camera; extracting features of the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map; fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector; and determining whether the to-be-detected ball cage is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage. Therefore, an efficient quality assessment model can be established, so that automation of the assessment process of the quality of the ball cage is realized, manual intervention is reduced, and efficiency is improved.
The embodiment of the invention also provides a forging processing technology of the ball cage, which comprises the following steps: heating the blank to obtain a heated blank; placing the heated blank into a die, and pressing the heated blank by using a press to fill the die cavity with the heated blank to obtain a pressed blank; taking the pressed blank out of the die, and cooling, cleaning, heat treating and machining the pressed blank to obtain a ball cage to be detected; performing quality detection on the ball cage to be detected,
The method is characterized by comprising the following steps of:
acquiring a ball cage surface state image of the ball cage to be detected, which is acquired by a camera;
extracting features of the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map;
fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector;
and determining whether the to-be-detected ball cage is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage.
In the above-mentioned forging process of the ball cage, performing feature extraction on the ball cage surface state image to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map, including: performing image preprocessing on the surface state image of the ball cage to obtain a preprocessed surface state image of the ball cage; and respectively extracting shape features, color features and texture features of the preprocessed surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map.
In the forging processing technology of the ball cage, the image preprocessing comprises image noise reduction.
In the foregoing forging process of the gabion, extracting shape features, color features and texture features of the preprocessed surface state image of the gabion to obtain the shape feature map, the color feature map and the texture feature map of the gabion, respectively, including: passing the preprocessed surface state image of the ball cage through a shape feature extractor based on a first convolution layer to obtain the ball cage shape feature map; passing the preprocessed surface state image of the ball cage through a color feature extractor based on a second convolution layer to obtain the color feature map of the ball cage; and passing the preprocessed surface state image of the ball cage through a texture feature extractor based on a plurality of convolution layers to obtain the texture feature map of the ball cage.
In the above-mentioned forging process of the ball cage, fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a multi-dimensional semantic fusion feature vector of the ball cage, including: and fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map by using a sparse feature fusion module to obtain the ball cage multidimensional semantic fusion feature vector.
In the above-mentioned ball cage forging process, the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map are fused by using a sparse feature fusion module to obtain the ball cage multidimensional semantic fusion feature vector, including: the cage shape feature map and the cage color feature map are subjected to a sparsification feature fusion formulaFusing the ball cage texture feature map to obtain the multi-dimensional semantic fusion feature vector of the ball cage; the sparse feature fusion formula is as follows:the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Feature vectors are fused for the multi-dimensional semantics of the ball cage,a cage-shaped feature vector obtained by feature map expansion for the cage-shaped feature map, +.>A color feature vector of the ball cage obtained by expanding the feature map for the color feature map of the ball cage, < > is obtained>A ball cage texture feature vector obtained by expanding the feature map for the ball cage texture feature map, < ->A transformation matrix for said gabion shape feature vector,>a conversion matrix for the color feature vector of the ball cage,>a transformation matrix for the ball cage texture feature vector, < > is provided>To emphasize the sphere cage shape feature vector, +.>To emphasize the color feature vector of the gabion, +. >To emphasize the ball cage texture feature vector.
In the above-mentioned forging process of the ball cage, determining whether the ball cage to be detected is qualified based on the multi-dimensional semantic fusion feature vector of the ball cage includes: and the multi-dimensional semantic fusion feature vector of the ball cage is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the ball cage to be detected is qualified or not.
In the foregoing process for forging a ball cage, passing the multi-dimensional semantic fusion feature vector of the ball cage through a classifier to obtain a classification result, where the classification result is used to indicate whether the ball cage to be detected is qualified, and the method includes: performing full-connection coding on the multi-dimensional semantic fusion feature vectors of the ball cages by using a plurality of full-connection layers of the classifier to obtain coded classification feature vectors; and passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
The forging processing technology of the ball cage further comprises the training steps of: training the shape feature extractor based on the first convolution layer, the color feature extractor based on the second convolution layer, the texture feature extractor based on the multi-layer convolution layer, the sparse feature fusion module and the classifier.
In the foregoing forging process of the ball cage, the training step includes: acquiring training data, wherein the training data comprises training ball cage surface state images of the ball cage to be detected, which are acquired by a camera, and a true value of whether the ball cage to be detected is qualified or not; performing image preprocessing on the training ball cage surface state image to obtain a training preprocessed ball cage surface state image, wherein the image preprocessing comprises image noise reduction; passing the training preprocessed surface state image of the ball cage through the shape feature extractor based on the first convolution layer to obtain a training ball cage shape feature diagram; passing the training preprocessed surface state image of the ball cage through the color feature extractor based on the second convolution layer to obtain a training ball cage color feature map; passing the training preprocessed surface state image of the ball cage through the texture feature extractor based on the multi-layer convolution layers to obtain a training ball cage texture feature map; the sparse feature fusion module is used for fusing the training ball cage shape feature map, the training ball cage color feature map and the training ball cage texture feature map to obtain training ball cage multidimensional semantic fusion feature vectors; passing the training ball cage multidimensional semantic fusion feature vector through a classifier to obtain a classification loss function value; training the shape feature extractor based on the first convolution layer, the color feature extractor based on the second convolution layer, the texture feature extractor based on the multi-layer convolution layer, the sparse feature fusion module and the classifier with the classification loss function value, wherein the training gabion multi-dimensional semantic fusion feature vector is optimized in each iteration of the training.
Compared with the prior art, the forging processing technology of the ball cage acquires the ball cage surface state image of the ball cage to be detected, which is acquired by a camera; extracting features of the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map; fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector; and determining whether the to-be-detected ball cage is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage. Therefore, an efficient quality assessment model can be established, so that automation of the assessment process of the quality of the ball cage is realized, manual intervention is reduced, and efficiency is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. In the drawings:
Fig. 1 is a flowchart of a forging process of a ball cage according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a system architecture of a forging process of a ball cage according to an embodiment of the present invention.
Fig. 3 is a block diagram of a forging system for a ball cage according to an embodiment of the present invention.
Fig. 4 is an application scenario diagram of a forging process of a ball cage according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings. The exemplary embodiments of the present invention and their descriptions herein are for the purpose of explaining the present invention, but are not to be construed as limiting the invention.
As used in the specification and in the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
A flowchart is used in the present invention to describe the operations performed by a system according to embodiments of the present invention. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Various exemplary embodiments, features and aspects of the disclosure will be described in detail below with reference to the drawings. In the drawings, like reference numbers indicate identical or functionally similar elements. Although various aspects of the embodiments are illustrated in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated.
The word "exemplary" is used herein to mean "serving as an example, embodiment, or illustration. Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.
The ball cage is an important component in the steering system of the automobile and plays roles in transmitting the rotation and force of the steering shaft. The cage is typically made of a metallic material, common materials including steel and its alloys, and its structure is typically an outer sphere and an inner sphere, which are connected by a mechanism. The manufacturing process of the cage has an important influence on its performance and life, and therefore there is a continuous need for optimization and improvement.
The manufacturing process of the cage generally comprises the following key steps:
1. and (3) material selection: suitable metallic materials are chosen, typically steel and its alloys, with good strength and wear resistance.
2. Forging and forming: the cage is typically formed using a forging process. In the forging process, the metal blank is heated and then placed in a die, and then subjected to pressure or impact force to generate plastic deformation, so that the preliminary shape of the ball cage is finally obtained.
3. And (3) heat treatment: the forged ball cage needs to be subjected to heat treatment to eliminate internal stress, improve the hardness and strength of the material and improve the overall mechanical properties of the material.
4. Machining: the heat treated ball cage needs to be precisely machined to ensure the dimensional accuracy and the surface quality, including the processes of turning, milling, drilling and the like.
5. Surface treatment: the surface of the cage is often subjected to an anti-corrosion treatment or a coating treatment to improve its corrosion resistance.
In the whole manufacturing process, it is very important to control the technological parameters, material performance and quality detection of each link, and the strength, wear resistance and service life of the ball cage can be improved by continuously optimizing and improving the manufacturing process of the ball cage, so that the reliability and safety of an automobile steering system are improved.
The existing ball cage forging process has some problems, and the problems may cause defects in the aspects of dimensional accuracy, surface quality, mechanical property, wear resistance and the like of the ball cage. Forging of a ball cage typically requires the use of dies, which frequently cause wear, and severe die wear can lead to irregularities or dimensional deviations in the shape of the ball cage, thereby affecting its assembly and use. In the forging process of the ball cage, the cooling speed is too high, so that the internal tissue structure is possibly uneven, and even cracks or deformation are generated, thereby affecting the mechanical property and the wear resistance of the ball cage. If the technological parameters such as temperature, pressure, deformation rate and the like in the forging process are controlled improperly, the internal tissue structure of the ball cage can be unsatisfactory, and the mechanical property and the wear resistance of the ball cage can be influenced. In the heat treatment process of the ball cage, if parameters such as temperature, time and the like are controlled improperly, incomplete or excessive phase change of a tissue can be caused, and the hardness, the strength and the toughness of the ball cage are affected. If the selected metal material is not satisfactory, the mechanical property and the wear resistance of the ball cage may not reach the standards.
In one embodiment of the present invention, fig. 1 is a flowchart of a forging process of a cage according to an embodiment of the present invention. Fig. 2 is a schematic diagram of a system architecture of a forging process of a ball cage according to an embodiment of the present invention. As shown in fig. 1 and 2, the forging process of the gabion according to the embodiment of the present invention includes: 110, heating the blank to obtain a heated blank; 120, placing the heated blank into a mould, and pressing the heated blank by a press machine, so that the heated blank fills a mould cavity to obtain a pressed blank; 130, taking the pressed blank out of the die, and cooling, cleaning, heat treating and machining the pressed blank to obtain a ball cage to be detected; 140, performing quality detection on the to-be-detected ball cage, wherein 140 performs quality detection on the to-be-detected ball cage, and includes: 141, acquiring a ball cage surface state image of the ball cage to be detected, which is acquired by a camera; 142, extracting features of the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map; 143, fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector; 144, determining whether the to-be-detected ball cage is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage.
In said step 110, the blank is heated to obtain a heated blank, the control of the heating temperature and heating time being critical, ensuring that the blank reaches the proper temperature for the subsequent pressing process. By proper heating, the blank can be in a state of easy plastic deformation, which is beneficial to the subsequent pressing process.
In the step 120, the heated blank is placed into a mold, and the heated blank is pressed by a press, so that the heated blank fills the mold cavity to obtain a pressed blank, and the pressure and the action time need to be controlled in the pressing process to ensure that the blank can fully fill the mold cavity to form an ideal ball cage shape. Through proper pressing, the size and the shape of the ball cage can be ensured to meet the requirements, and the compactness and the mechanical property of the ball cage are improved.
In the step 130, the pressed blank is taken out from the mold, and the pressed blank is cooled, cleaned, heat treated and machined to obtain the ball cage to be detected, so that the control of the cooling speed, the cleaning process, the heat treatment parameters and the machining process is very important, and the internal tissue structure, the surface quality and the dimensional accuracy of the ball cage are ensured to meet the requirements. Through reasonable cooling, cleaning, heat treatment and machining, the mechanical property, the wear resistance and the dimensional accuracy of the ball cage can be improved, and the quality of the ball cage can be ensured to reach the standard.
In the step 140, the quality detection is performed on the to-be-detected ball cage, and the quality detection needs to include size measurement, surface quality inspection, mechanical property test and the like, so as to ensure that the quality of the ball cage meets the requirements. The quality problem of the ball cage can be found in time through quality detection, products are ensured to meet the standard, and the reliability and safety of an automobile steering system are improved.
Wherein step 140 further comprises: step 141, acquiring a state image of the surface of the ball cage to be detected, acquired by a camera, wherein the position and parameter setting of the camera are required to be proper, so that the state information of the surface of the ball cage can be accurately captured. Through the acquisition of the surface state image of the ball cage, the digital recording of the surface state of the ball cage can be realized, and a data basis is provided for subsequent analysis and judgment. And 142, carrying out feature extraction on the surface state image of the ball cage to obtain a ball cage shape feature image, a ball cage color feature image and a ball cage texture feature image, wherein the feature extraction algorithm is reasonable in selection and parameter setting, and can ensure that feature information such as shape, color and texture of the surface of the ball cage can be accurately extracted. Through feature extraction, the surface state image of the ball cage can be converted into specific feature data such as a shape, a color, a texture and the like, and a foundation is provided for the generation of a subsequent multidimensional semantic fusion feature vector. Step 143, fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector, wherein the selection and parameter setting of a fusion algorithm are reasonable, so that the information of different features can be fully utilized, and a representative multidimensional semantic fusion feature vector is obtained. Through the generation of the multi-dimensional semantic fusion feature vector of the ball cage, a plurality of feature dimensions on the surface of the ball cage can be combined, and the accuracy of judging the quality of the ball cage is improved. And 144, determining whether the ball cage to be detected is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage, wherein the setting of a qualified standard and the establishment of a judging algorithm are scientific and reliable, and ensuring that whether the quality of the ball cage meets the requirement or not can be accurately judged. Through judging based on the multidimensional semantic fusion feature vector, the automatic and accurate assessment of the quality of the ball cage can be realized, and the quality inspection efficiency and accuracy are improved.
The quality detection of the ball cage is to ensure that the ball cage meets design requirements and has good performance and reliability. The existing quality detection modes mainly comprise the following steps: 1. the surface of the cage is inspected visually or with the aid of tools such as a microscope, magnifying glass, etc., to find possible surface defects, scratches, bubbles, etc. 2. Size measurement: the diameter, roundness, flatness and other dimensions of the ball cage are measured using measuring tools, such as calipers, vernier calipers, projectors and the like, to verify whether they meet the specified dimensional requirements. These detection methods rely on manual operations and judgment, and subjective factors may exist, which may affect the inconsistency and reliability of the results. In addition, a lot of manpower, time and resources are consumed, and the improvement of the production efficiency is limited. Thus, an optimized solution is desired.
Aiming at the technical problems, the technical conception of the application is as follows: and carrying out multidimensional feature extraction on the surface state image of the ball cage to be detected by using an image processing technology based on deep learning so as to excavate the shape, color and texture features of the ball cage from the surface state image, thereby evaluating the quality of the ball cage.
The deep learning model can learn complex image characteristic representation, and can automatically extract rich characteristic information from an image, wherein the rich characteristic information comprises multidimensional characteristics such as shapes, colors, textures and the like. The characteristics extracted by the deep learning model can be subjected to multidimensional fusion, different characteristic dimensions are combined to form more comprehensive and representative ball cage surface characteristic description, and more comprehensive information is provided for quality evaluation. By using the deep learning technology, an efficient quality assessment model can be established, so that automation of the assessment process of the quality of the ball cage is realized, manual intervention is reduced, and efficiency is improved. The deep learning model can learn characteristic representation from a large amount of data, so that accuracy and robustness of quality evaluation can be improved when the surface state image of the ball cage is processed, and the ball cage with different illumination, angles and surface states can be effectively evaluated. The deep learning model can realize real-time processing under the support of hardware, so that the real-time analysis and quality evaluation of the surface state image of the ball cage can be realized, and the quality problem can be found timely.
Based on the above, in the technical scheme of the application, firstly, a ball cage surface state image of a ball cage to be detected, which is acquired by a camera, is acquired; and performing image preprocessing on the surface state image of the ball cage to obtain a preprocessed surface state image of the ball cage, wherein the image preprocessing comprises image noise reduction. The surface visual information of the ball cage to be detected is converted into digital image data through the camera, and the characteristics of the shape, the color, the texture and the like of the ball cage can be obtained from the digital image data in subsequent image processing and analysis, so that quality detection and judgment can be carried out. In addition, the quality of the surface state image of the ball cage can be optimized by carrying out image preprocessing on the surface state image of the ball cage, and noise and interference are reduced so as to obtain clearer and more reliable image data. Among them, image noise reduction is an important step in image preprocessing, and the purpose of the image noise reduction is to reduce noise interference in an image and improve definition and quality of the image. In an actual application scenario, the image noise may be caused by various factors, such as noise existing in the camera itself, ambient light changes, interference in the transmission process, and the like. The image noise reduction algorithm performs smoothing processing on the image through filtering and other technologies, and removes or reduces the influence of noise, so that the quality and usability of the image are improved. Common image noise reduction methods include mean filtering, median filtering, gaussian filtering, and the like.
In a specific embodiment of the present application, feature extraction is performed on the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map, including: performing image preprocessing on the surface state image of the ball cage to obtain a preprocessed surface state image of the ball cage; and respectively extracting shape features, color features and texture features of the preprocessed surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map.
Then, passing the preprocessed surface state image of the ball cage through a shape feature extractor based on a first convolution layer to obtain a shape feature map of the ball cage; passing the preprocessed surface state image of the ball cage through a color feature extractor based on a second convolution layer to obtain a color feature map of the ball cage; and simultaneously, passing the preprocessed surface state image of the ball cage through a texture feature extractor based on a plurality of convolution layers to obtain a texture feature map of the ball cage. Here, multi-dimensional information about the surface of the gabion can be obtained by extracting shape features, color features and texture features in the pre-processed surface state image of the gabion, thereby more fully describing characteristics and quality of the gabion. Specifically, by extracting the shape characteristics, such as the outline, the geometric shape and the like, of the to-be-detected ball cage, whether the size of the ball cage meets the specified requirements or not can be evaluated, and whether the shape defects or deformation and other problems exist or not can be detected. The shape features provide information about the geometry of the cage and can be used to verify that the appearance of the cage meets design requirements. Second, the color characteristics of the cage may reflect the nature of the cage material and the uniformity of the surface. By extracting the color characteristics of the surface of the ball cage, the problems of color deviation, uneven coating and the like, and the quality problems of oxidation, pollution or coating peeling and the like which possibly exist can be detected. In addition, the texture characteristics of the surface of the cage reflect the surface quality and texture uniformity of the cage. By extracting the texture characteristics of the surface of the ball cage, whether the surface has the problems of concave-convex, burr, looseness and the like and the quality problems of possible cracks, damage or abrasion and the like can be detected. In summary, by comprehensively considering the shape feature, the color feature and the texture feature, the quality condition of the ball cage can be comprehensively evaluated, including appearance defects, dimensional deviations, material problems, overall processing quality and the like.
In a specific embodiment of the present application, extracting the shape feature, the color feature and the texture feature of the preprocessed surface state image of the ball cage to obtain the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map, respectively, includes: passing the preprocessed surface state image of the ball cage through a shape feature extractor based on a first convolution layer to obtain the ball cage shape feature map; passing the preprocessed surface state image of the ball cage through a color feature extractor based on a second convolution layer to obtain the color feature map of the ball cage; and passing the preprocessed surface state image of the ball cage through a texture feature extractor based on a plurality of convolution layers to obtain the texture feature map of the ball cage.
It will be appreciated by those of ordinary skill in the art that Convolutional Neural Networks (CNNs) are a type of deep learning model that is widely used in the field of image processing. The core idea of the convolutional neural network is to gradually extract abstract features of images through multi-layer convolution and pooling operations. The convolution layer can extract low-level local features such as information of edges, shapes, colors and the like due to shallower network depth. These low-level local features may be characterized and characterized for the description of shape and color. While more abstract feature information, such as texture, etc., can be extracted by multiple convolution layers. Thus, in the technical scheme of the application, the shape feature extractor, the color feature extractor and the texture feature extractor are respectively constructed by utilizing the convolution layer, so that the shape feature, the color feature and the texture feature contained in the preprocessed surface state image of the ball cage are extracted, and an important data source is provided for judging whether the ball cage to be detected is qualified or not.
It should be appreciated that the quality and characteristics of the cage may be described from various angles, including aspects of shape, color, and texture. These features may provide limited information when analyzed alone, and by fusing them together, the advantages of different features may be taken into account comprehensively, resulting in a more comprehensive and accurate representation of the features. In the technical scheme of the application, it is expected that a sparse feature fusion module is used for fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector. The sparse feature fusion module fully utilizes semantic information expressed by single-mode feature distribution, namely, the spherical cage shape features expressed by the spherical cage shape feature images, the spherical cage color features expressed by the spherical cage color feature images and the texture features expressed by the spherical cage texture feature images to perform fusion and interaction based on association response, and meanwhile, the inhibition effect among different types of information is eliminated, so that the fusion process is more stable.
In one specific embodiment of the present application, fusing the ball cage shape feature map, the ball cage color feature map, and the ball cage texture feature map to obtain a ball cage multi-dimensional semantic fusion feature vector comprises: and fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map by using a sparse feature fusion module to obtain the ball cage multidimensional semantic fusion feature vector.
Further, the cage shape feature map, the cage color feature map and the cage texture feature map are fused by using a sparse feature fusion module to obtain the cage multidimensional semantic fusion feature vector, which comprises: the cage shape feature map and the cage face are subjected to the following sparse feature fusion formulaFusing the color feature map and the ball cage texture feature map to obtain the multi-dimensional semantic fusion feature vector of the ball cage; the sparse feature fusion formula is as follows:the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Fusing feature vectors for the multi-dimensional semantics of the ball cage,>a cage-shaped feature vector obtained by feature map expansion for the cage-shaped feature map, +.>A color feature vector of the ball cage obtained by expanding the feature map for the color feature map of the ball cage, < > is obtained>A ball cage texture feature vector obtained by expanding the feature map for the ball cage texture feature map, < ->A transformation matrix for said gabion shape feature vector,>a conversion matrix for the color feature vector of the ball cage,>a transformation matrix for the ball cage texture feature vector, < > is provided>To emphasize the sphere cage shape feature vector, +.>To emphasize the color feature vector of the gabion, +.>To make the texture characteristic direction of the ball cage obvious Amount of the components.
Further, the multi-dimensional semantic fusion feature vector of the ball cage is passed through a classifier to obtain a classification result, and the classification result is used for indicating whether the ball cage to be detected is qualified or not. Wherein the classifier is a machine learning model that learns the mapping from input features to output labels. By taking the multi-dimensional semantic fusion feature vector of the ball cage as input, the classifier can learn the association between features and the ball cage pass/fail tags. Thus, the classifier can judge the quality state of the ball cage according to the mode and rule of the multi-dimensional semantic fusion feature vector of the ball cage.
In a specific embodiment of the present application, determining whether the to-be-detected ball cage is qualified based on the multi-dimensional semantic fusion feature vector of the ball cage includes: and the multi-dimensional semantic fusion feature vector of the ball cage is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the ball cage to be detected is qualified or not.
Further, the multi-dimensional semantic fusion feature vector of the ball cage is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the ball cage to be detected is qualified or not, and the method comprises the following steps: performing full-connection coding on the multi-dimensional semantic fusion feature vectors of the ball cages by using a plurality of full-connection layers of the classifier to obtain coded classification feature vectors; and passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
In one embodiment of the present application, the forging process of the ball cage further includes a training step of: training the shape feature extractor based on the first convolution layer, the color feature extractor based on the second convolution layer, the texture feature extractor based on the multi-layer convolution layer, the sparse feature fusion module and the classifier. The training step comprises the following steps: acquiring training data, wherein the training data comprises training ball cage surface state images of the ball cage to be detected, which are acquired by a camera, and a true value of whether the ball cage to be detected is qualified or not; performing image preprocessing on the training ball cage surface state image to obtain a training preprocessed ball cage surface state image, wherein the image preprocessing comprises image noise reduction; passing the training preprocessed surface state image of the ball cage through the shape feature extractor based on the first convolution layer to obtain a training ball cage shape feature diagram; passing the training preprocessed surface state image of the ball cage through the color feature extractor based on the second convolution layer to obtain a training ball cage color feature map; passing the training preprocessed surface state image of the ball cage through the texture feature extractor based on the multi-layer convolution layers to obtain a training ball cage texture feature map; the sparse feature fusion module is used for fusing the training ball cage shape feature map, the training ball cage color feature map and the training ball cage texture feature map to obtain training ball cage multidimensional semantic fusion feature vectors; passing the training ball cage multidimensional semantic fusion feature vector through a classifier to obtain a classification loss function value; training the shape feature extractor based on the first convolution layer, the color feature extractor based on the second convolution layer, the texture feature extractor based on the multi-layer convolution layer, the sparse feature fusion module and the classifier with the classification loss function value, wherein the training gabion multi-dimensional semantic fusion feature vector is optimized in each iteration of the training.
In the technical scheme of the application, the training ball cage shape feature map, the training ball cage color feature map and the training ball cage texture feature map respectively express the image shape semantic features, the image color semantic features and the image texture semantic features of the ball cage surface state image after training pretreatment, so that the characteristic distribution information saliency of the training ball cage shape feature map, the training ball cage color feature map and the training ball cage texture feature map, which are respectively based on the specific image semantic features, can be influenced when the sparse feature is fused by using the sparse feature fusion module so as to obtain the training ball cage multidimensional semantic fusion feature vector, and the training ball cage shape feature map, the training ball cage color feature map and the training ball cage texture feature map are considered to be difficultly focused on the characteristic of the multidimensional fusion feature vector in the training process, so that the local distribution speed of the training ball cage expression vector is difficult to be influenced.
Based on the above, the applicant optimizes the training ball cage multidimensional semantic fusion feature vector every time the training ball cage multidimensional semantic fusion feature vector carries out the iteration of classification regression through a classifier, and the training ball cage multidimensional semantic fusion feature vector is expressed as: optimizing the training ball cage multidimensional semantic fusion feature vector in each round of iteration of training by using the following optimization formula to obtain an optimized ball cage multidimensional semantic fusion feature vector; wherein, the optimization formula is:
the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>And->The training ball cage multidimensional semantic fusion feature vector is +.>Square of 1-norm and 2-norm, < ->Is the training ball cage multidimensional semantic fusion feature vector +.>Length of (2), and->Is a weight superparameter,/->Is the characteristic value of the training ball cage multidimensional semantic fusion characteristic vector, < + >>Is the characteristic value of the optimized multi-dimensional semantic fusion characteristic vector of the ball cage, < >>Is the training ball cage multidimensional semantic fusion feature vector, </i >>A logarithmic function with a base of 2 is shown.
Specifically, feature vectors are fused through multidimensional semantics based on the training ball cageGeometric registration of its high-dimensional feature manifold shape is performed with respect to the scale and structural parameters of the training gabion multi-dimensional semantic fusion feature vector +. >Features with rich feature semantic information in the feature set formed by the feature values, namely distinguishable stable interest features which are based on local context information and represent dissimilarity when the classifier classifies, so that the training ball cage multi-dimensional semantic fusion feature vector is realized>And the feature information significance is marked in the classification process, so that the training speed of the classifier is improved.
In summary, a forging process of a ball cage according to an embodiment of the present invention is elucidated, which performs multi-dimensional feature extraction on a cage surface state image of a ball cage to be detected by using an image processing technique based on deep learning, so as to extract shape, color and texture features of the ball cage therefrom, thereby evaluating the quality of the ball cage.
In one embodiment of the present invention, FIG. 3 is a block diagram of a forging system for a cage according to an embodiment of the present invention. As shown in fig. 3, a forging system 200 of a ball cage according to an embodiment of the present invention includes: a blank heating module 210 for heating the blank to obtain a heated blank; a blank pressing module 220, configured to put the heated blank into a mold, press the heated blank with a press, and fill the heated blank into a mold cavity to obtain a pressed blank; a post-press blank processing module 230, configured to take the post-press blank out of the mold, and cool, clean, heat treat, and machine the post-press blank to obtain a ball cage to be inspected; the quality detection module 240 is configured to perform quality detection on the to-be-detected ball cage, and is characterized in that the quality detection module includes: a ball cage surface state image obtaining unit 241, configured to obtain a ball cage surface state image of the ball cage to be detected, which is collected by a camera; a feature extraction unit 242, configured to perform feature extraction on the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map, and a ball cage texture feature map; a fusion unit 243, configured to fuse the ball cage shape feature map, the ball cage color feature map, and the ball cage texture feature map to obtain a multi-dimensional semantic fusion feature vector of the ball cage; and the to-be-detected ball cage qualification judging unit 244 is used for determining whether the to-be-detected ball cage is qualified based on the multi-dimensional semantic fusion feature vector of the ball cage.
Here, it will be understood by those skilled in the art that the specific functions and operations of the respective units and modules in the forging processing system of the above-described ball cage have been described in detail in the above description of the forging processing process of the ball cage with reference to fig. 1 to 2, and thus, repetitive descriptions thereof will be omitted.
As described above, the forging processing system 200 of a ball cage according to an embodiment of the present invention may be implemented in various terminal devices, such as a server or the like for forging processing of a ball cage. In one example, the forging system 200 of the ball cage according to an embodiment of the present invention may be integrated into the terminal device as a software module and/or hardware module. For example, the forging system 200 of the cage may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the forging system 200 of the cage may also be one of a plurality of hardware modules of the terminal equipment.
Alternatively, in another example, the cage forging system 200 and the terminal device may be separate devices, and the cage forging system 200 may be connected to the terminal device via a wired and/or wireless network and communicate interactive information in accordance with a agreed data format.
Fig. 4 is an application scenario diagram of a forging process of a ball cage according to an embodiment of the present invention. As shown in fig. 4, in this application scenario, first, a cage surface state image of the cage to be detected acquired by a camera is acquired (e.g., C as illustrated in fig. 4); the acquired surface state image of the ball cage is then input into a server (e.g., S as illustrated in fig. 4) where a forging algorithm of the ball cage is deployed, wherein the server is capable of processing the surface state image of the ball cage based on the forging algorithm of the ball cage to determine whether the ball cage to be inspected is acceptable.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (10)
1. A forging process of a ball cage, comprising: heating the blank to obtain a heated blank; placing the heated blank into a die, and pressing the heated blank by using a press to fill the die cavity with the heated blank to obtain a pressed blank; taking the pressed blank out of the die, and cooling, cleaning, heat treating and machining the pressed blank to obtain a ball cage to be detected; performing quality detection on the ball cage to be detected,
The method is characterized by comprising the following steps of:
acquiring a ball cage surface state image of the ball cage to be detected, which is acquired by a camera;
extracting features of the surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map;
fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map to obtain a ball cage multidimensional semantic fusion feature vector;
and determining whether the to-be-detected ball cage is qualified or not based on the multi-dimensional semantic fusion feature vector of the ball cage.
2. The forging process of the gabion according to claim 1, wherein the feature extraction of the gabion surface state image to obtain a gabion shape feature map, a gabion color feature map, and a gabion texture feature map, comprises:
performing image preprocessing on the surface state image of the ball cage to obtain a preprocessed surface state image of the ball cage;
and respectively extracting shape features, color features and texture features of the preprocessed surface state image of the ball cage to obtain a ball cage shape feature map, a ball cage color feature map and a ball cage texture feature map.
3. The forging process for a ball cage according to claim 2, wherein the image preprocessing includes image noise reduction.
4. A forging process for a gabion according to claim 3, wherein extracting shape features, color features and texture features of the preprocessed surface state image of the gabion to obtain the gabion shape feature map, the gabion color feature map and the gabion texture feature map, respectively, comprises:
passing the preprocessed surface state image of the ball cage through a shape feature extractor based on a first convolution layer to obtain the ball cage shape feature map;
passing the preprocessed surface state image of the ball cage through a color feature extractor based on a second convolution layer to obtain the color feature map of the ball cage;
and passing the preprocessed surface state image of the ball cage through a texture feature extractor based on a plurality of convolution layers to obtain the texture feature map of the ball cage.
5. The forging process of the ball cage according to claim 4, wherein fusing the ball cage shape feature map, the ball cage color feature map, and the ball cage texture feature map to obtain a ball cage multi-dimensional semantic fusion feature vector comprises:
And fusing the ball cage shape feature map, the ball cage color feature map and the ball cage texture feature map by using a sparse feature fusion module to obtain the ball cage multidimensional semantic fusion feature vector.
6. The forging process of the ball cage according to claim 5, wherein fusing the ball cage shape feature map, the ball cage color feature map, and the ball cage texture feature map using a sparsified feature fusion module to obtain the ball cage multi-dimensional semantic fusion feature vector comprises:
fusing the cage shape feature map, the ball cage color feature map and the ball cage texture feature map by using the following sparse feature fusion formula to obtain the ball cage multidimensional semantic fusion feature vector; the sparse feature fusion formula is as follows:the method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Fusing feature vectors for the multi-dimensional semantics of the ball cage,>a cage-shaped feature vector obtained by feature map expansion for the cage-shaped feature map, +.>A color feature vector of the ball cage obtained by expanding the feature map for the color feature map of the ball cage, < > is obtained>A ball cage texture feature vector obtained by expanding the feature map for the ball cage texture feature map, < - >A transformation matrix for said gabion shape feature vector,>a conversion matrix for the color feature vector of the ball cage,>a transformation matrix for the ball cage texture feature vector, < > is provided>In order to emphasize the spherical cage shape feature vector,to emphasize the color feature vector of the gabion, +.>To emphasize the ball cage texture feature vector.
7. The forging process of the ball cage according to claim 6, wherein determining whether the ball cage to be detected is qualified based on the ball cage multi-dimensional semantic fusion feature vector comprises:
and the multi-dimensional semantic fusion feature vector of the ball cage is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the ball cage to be detected is qualified or not.
8. The forging process of the ball cage according to claim 7, wherein the passing the multi-dimensional semantic fusion feature vector of the ball cage through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the ball cage to be detected is qualified or not, comprises:
performing full-connection coding on the multi-dimensional semantic fusion feature vectors of the ball cages by using a plurality of full-connection layers of the classifier to obtain coded classification feature vectors; and
and the coding classification feature vector is passed through a Softmax classification function of the classifier to obtain the classification result.
9. The forging process for a ball cage according to claim 8, further comprising a training step of: training the shape feature extractor based on the first convolution layer, the color feature extractor based on the second convolution layer, the texture feature extractor based on the multi-layer convolution layer, the sparse feature fusion module and the classifier.
10. The forging process for a ball cage according to claim 9, wherein said training step comprises:
acquiring training data, wherein the training data comprises training ball cage surface state images of the ball cage to be detected, which are acquired by a camera, and a true value of whether the ball cage to be detected is qualified or not;
performing image preprocessing on the training ball cage surface state image to obtain a training preprocessed ball cage surface state image, wherein the image preprocessing comprises image noise reduction;
passing the training preprocessed surface state image of the ball cage through the shape feature extractor based on the first convolution layer to obtain a training ball cage shape feature diagram;
passing the training preprocessed surface state image of the ball cage through the color feature extractor based on the second convolution layer to obtain a training ball cage color feature map;
Passing the training preprocessed surface state image of the ball cage through the texture feature extractor based on the multi-layer convolution layers to obtain a training ball cage texture feature map;
the sparse feature fusion module is used for fusing the training ball cage shape feature map, the training ball cage color feature map and the training ball cage texture feature map to obtain training ball cage multidimensional semantic fusion feature vectors;
passing the training ball cage multidimensional semantic fusion feature vector through a classifier to obtain a classification loss function value;
training the shape feature extractor based on the first convolution layer, the color feature extractor based on the second convolution layer, the texture feature extractor based on the multi-layer convolution layer, the sparse feature fusion module and the classifier with the classification loss function value, wherein the training gabion multi-dimensional semantic fusion feature vector is optimized in each iteration of the training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410128698.9A CN117689902A (en) | 2024-01-31 | 2024-01-31 | Forging processing technology of ball cage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410128698.9A CN117689902A (en) | 2024-01-31 | 2024-01-31 | Forging processing technology of ball cage |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117689902A true CN117689902A (en) | 2024-03-12 |
Family
ID=90126796
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410128698.9A Pending CN117689902A (en) | 2024-01-31 | 2024-01-31 | Forging processing technology of ball cage |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117689902A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102248101A (en) * | 2011-06-15 | 2011-11-23 | 江苏创一精锻有限公司 | Whole forging near net forming process for inner ball cage sliding sleeve |
| CN115456955A (en) * | 2022-08-19 | 2022-12-09 | 燕山大学 | Method for detecting internal burr defect of ball cage dust cover |
| CN115661090A (en) * | 2022-11-01 | 2023-01-31 | 安徽万方织染有限公司 | Intelligent processing technology and system for textile fabric |
| US20230222643A1 (en) * | 2022-01-11 | 2023-07-13 | Bentley Systems, Incorporated | Semantic deep learning and rule optimization for surface corrosion detection and evaluation |
| CN117173147A (en) * | 2023-09-19 | 2023-12-05 | 江西天钢科技有限公司 | Surface treatment equipment and method for steel strip processing |
| CN117348574A (en) * | 2023-11-24 | 2024-01-05 | 佛山市时力涂料科技有限公司 | Intelligent control system and method for paint production line |
-
2024
- 2024-01-31 CN CN202410128698.9A patent/CN117689902A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102248101A (en) * | 2011-06-15 | 2011-11-23 | 江苏创一精锻有限公司 | Whole forging near net forming process for inner ball cage sliding sleeve |
| US20230222643A1 (en) * | 2022-01-11 | 2023-07-13 | Bentley Systems, Incorporated | Semantic deep learning and rule optimization for surface corrosion detection and evaluation |
| CN115456955A (en) * | 2022-08-19 | 2022-12-09 | 燕山大学 | Method for detecting internal burr defect of ball cage dust cover |
| CN115661090A (en) * | 2022-11-01 | 2023-01-31 | 安徽万方织染有限公司 | Intelligent processing technology and system for textile fabric |
| CN117173147A (en) * | 2023-09-19 | 2023-12-05 | 江西天钢科技有限公司 | Surface treatment equipment and method for steel strip processing |
| CN117348574A (en) * | 2023-11-24 | 2024-01-05 | 佛山市时力涂料科技有限公司 | Intelligent control system and method for paint production line |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111815601B (en) | Texture image surface defect detection method based on depth convolution self-encoder | |
| CN110806736B (en) | Method for detecting quality information of forge pieces of die forging forming intelligent manufacturing production line | |
| CN109141232A (en) | A kind of circle plate casting online test method based on machine vision | |
| CN107492098B (en) | It is a kind of based on PCA and CNN high-temperature forging surface defect in position detecting method | |
| CN113239930B (en) | Glass paper defect identification method, system, device and storage medium | |
| CN110992317A (en) | PCB defect detection method based on semantic segmentation | |
| CN109840900B (en) | Fault online detection system and detection method applied to intelligent manufacturing workshop | |
| CN113034482A (en) | Surface roughness detection method based on machine vision and machine learning | |
| CN112017204A (en) | Tool state image classification method based on edge marker graph neural network | |
| CN114565314B (en) | Digital twinning-based hot rolled steel coil end face quality control system and method | |
| CN115018790A (en) | Workpiece surface defect detection method based on anomaly detection | |
| CN112967271B (en) | Casting surface defect identification method based on improved DeepLabv3+ network model | |
| CN117409005B (en) | Defective product detection system and method for plate receiving machine based on image | |
| CN111539951B (en) | Visual detection method for outline size of ceramic grinding wheel head | |
| CN116883393B (en) | Metal surface defect detection method based on anchor frame-free target detection algorithm | |
| Hassan et al. | An artificial intelligent manufacturing process for high-quality low-cost production | |
| CN117689902A (en) | Forging processing technology of ball cage | |
| Shankar et al. | Machine Vision based quality inspection for automotive parts using edge detection technique | |
| Frayman et al. | Machine vision system for automatic inspection of surface defects in aluminum die casting | |
| CN115180476B (en) | Elevator traction sheave race form detection method based on three-dimensional reconstruction | |
| CN116777836A (en) | Multi-mode data driving-based injection molding process product quality detection method | |
| CN115587989A (en) | Workpiece CT image defect detection and segmentation method and system | |
| CN113916884A (en) | Hard turning workpiece surface quality online prediction method | |
| CN118096728B (en) | Machine vision-based part spraying quality detection method | |
| CN117974665B (en) | Metal mold on-line detection method and equipment based on computer vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |