CN115984843A - Remanufacturing raw material evaluation method and device, storage medium and electronic equipment - Google Patents
Remanufacturing raw material evaluation method and device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN115984843A CN115984843A CN202211559344.7A CN202211559344A CN115984843A CN 115984843 A CN115984843 A CN 115984843A CN 202211559344 A CN202211559344 A CN 202211559344A CN 115984843 A CN115984843 A CN 115984843A
- Authority
- CN
- China
- Prior art keywords
- information
- remanufactured
- raw material
- deep learning
- network
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The application discloses a remanufacturing raw material evaluation method and device, a storage medium and electronic equipment. The method comprises the following steps: acquiring multi-angle image information of a remanufactured raw material; determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilene, an enhanced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet; determining recessive damage information in the remanufactured raw material by laser ultrasonic detection and/or electromagnetic ultrasonic detection; and evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information. The detection speed is improved, the identification effect is better, and the identification types are more.
Description
Technical Field
The present disclosure relates to the field of machine vision and remanufacturing technologies, and in particular, to a remanufacturing raw material evaluation method, apparatus, system, storage medium, and electronic device.
Background
The remanufacturing is used as a part of important technical support of circular economy, and the performance of waste equipment or products can be quickly recovered and even improved by reasonably utilizing an advanced technical process, so that the quality performance of the waste equipment or the products is not lower than the level of original new products. In addition, remanufacturing can utilize the residual values of waste equipment and parts to the maximum extent while realizing efficient recycling of resources, and the full life cycle life of the equipment is prolonged. Particularly, the manufacturing process of the product is greatly shortened, and compared with new product manufacturing, the remanufacturing can obviously reduce carbon emission, so that the double effects of resource recycling and green energy conservation and emission reduction are achieved.
The remanufactured raw materials are large in quantity and various in types, and the damage form and degree are various. Due to the limitation of the identification accuracy and the identification speed, the currently identified objects are limited to the objects with regular shapes, simple structures, obvious characteristics and single judgment criteria.
Therefore, how to intelligently identify the waste parts and how to quickly and accurately classify and shunt the waste parts become a difficult problem in remanufacturing production; moreover, functional upgrading of waste products is the mainstream direction of remanufacturing, but most of the current remanufacturing process designs depend on process test exploration, so that the problems of long time period and high design and shaping cost are urgently solved.
Disclosure of Invention
In view of the above problems, the present application provides a remanufacturing raw material evaluation method, apparatus, system, storage medium, and electronic device. Identifying basic information of an object by using monocular machine vision and binocular machine vision identification technologies, and realizing preliminary classification and distribution of waste parts; the method intelligently identifies the three-dimensional surface morphology of the parts, and researches and identifies the qualitative and quantitative analysis method of the dominant damages such as target corrosion, abrasion, micro-deformation and the like. An acousto-optic-electromagnetic multi-physical-parameter mapping model of near-surface damage of parts is established, an intelligent nondestructive testing technology based on acousto-optic-electromagnetic multi-physical parameters is developed, and intelligent identification of hidden damage such as cracks and fatigue of waste motor parts is achieved. The method comprehensively utilizes the methods of fuzzy comprehensive evaluation, quintuple coefficient evaluation and the like, researches key technologies of failure rate evaluation, aging risk evaluation, remanufacturing value evaluation and the like, and establishes the remanufacturability criterion of the remanufactured raw material component. The problem of excessively depending on manual work in the detection and evaluation process of the remanufactured raw materials is solved.
In a first aspect of the present application, there is provided a remanufactured feedstock evaluation method comprising:
acquiring multi-angle image information of a remanufactured raw material;
determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilene, a reinforced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet;
determining recessive damage information in the remanufactured raw material by laser ultrasonic detection and/or electromagnetic ultrasonic detection;
and evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information.
Further, the explicit impairment information comprises:
first and second dominant impairment information; wherein the content of the first and second substances,
the first dominant impairment information comprises: rust information and/or corrosion information; the second dominant impairment information comprises: wear information and/or deformation information.
Further, the recessive impairment information comprises:
crack information and/or fatigue life information.
Further, the determining the recessive damage information in the remanufactured raw material by laser ultrasonic detection and/or electromagnetic ultrasonic detection includes:
detecting the remanufactured raw material by the laser ultrasonic detection method to obtain first recessive damage information;
detecting the remanufactured raw material by the electromagnetic ultrasonic detection method to obtain second recessive damage information;
and taking the first recessive damage information and the second recessive damage information as the recessive damage information.
Further, the evaluating the remanufactured raw material according to the dominant impairment information and the recessive impairment information by a preset evaluation rule comprises:
judging the dominant damage information and the recessive damage information through a preset remanufacturing evaluation model;
and under the condition that the dominant impairment information meets a first preset condition or the recessive impairment information meets a second preset condition, judging that the remanufacturing raw material cannot be used for remanufacturing.
Further, the acquiring multi-angle image information of the remanufactured raw material comprises:
and acquiring multi-angle image information of the remanufactured raw material acquired by a monocular camera and/or a binocular camera.
In a second aspect of the present application, there is provided a remanufactured feedstock evaluation device comprising:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring multi-angle image information of the remanufactured raw material;
the first determining module is used for determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilenet, an enhanced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet;
the second determination module is used for determining recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method;
and the evaluation module is used for evaluating the remanufactured raw material according to the dominant damage information and the recessive damage information through a preset evaluation rule.
In a third aspect of the present application, there is provided a remanufactured feedstock evaluation system comprising:
the one or more cameras are in communication connection with the evaluation server and are used for acquiring multi-angle image information of the remanufactured raw material and sending the multi-angle image information to the evaluation server;
the evaluation server is used for determining dominant damage information in the multi-angle image information through a preset deep learning network, determining recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method, and evaluating the remanufactured raw material through a preset remanufacturing evaluation model according to the dominant damage information and the recessive damage information, wherein the trunk feature extraction network of the preset deep learning network comprises a Mobilenet, the reinforced feature extraction network of the preset deep learning network comprises a space pyramid network SPP, and the path fusion network of the preset deep learning network comprises a PANET.
In a fourth aspect of the present application, a computer-readable storage medium is provided, storing a computer program, executable by one or more processors, for implementing the method as described above.
In a fifth aspect of the application, an electronic device is provided, comprising a memory and one or more processors, said memory having stored thereon a computer program, said memory and said one or more processors being communicatively connected to each other, the computer program, when executed by said one or more processors, implementing the method as described above.
Compared with the prior art, the technical scheme of the application has the following advantages or beneficial effects:
1. the trunk feature extraction network used by the deep learning network saves resources, avoids invalid calculation, can quickly and effectively extract features, has smaller volume capacity than other trunk feature extraction networks, and improves the detection speed;
2. the deep learning network uses a network module for enhancing feature extraction, so that the feature learning of a data set can be effectively enhanced, and the recognition effect is better;
3. besides surface damage information of remanufactured raw materials, an acoustoelectric and optomagnetic multi-physical-parameter model is utilized, an intelligent nondestructive testing scheme of acoustoelectric and optomagnetic multi-physical parameters is disclosed, deeper recessive damage identification is carried out, and the identification types are more;
4. different from the limit that the conventional target detection only can detect the inherent type of the object, the remanufacturing evaluation model is established, and the remanufacturing evaluation decision of the part is made from the aspect of remanufacturing economic feasibility.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only the embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate exemplary embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application, in which:
FIG. 1 is a flow chart of a remanufactured feedstock evaluation method according to an embodiment of the present disclosure;
FIG. 2 is a flow chart of another remanufactured feedstock evaluation method provided in an embodiment of the present application;
FIG. 3 is a schematic diagram of a backbone feature extraction network architecture;
FIG. 4 is a schematic diagram of a convolutional neural network decomposition structure;
FIG. 5 is a schematic diagram of an enhanced feature extraction network architecture;
FIG. 6 is a schematic diagram of a feature fusion network architecture;
FIG. 7 is a schematic diagram of an electromagnetic ultrasonic test;
FIG. 8 is a schematic illustration of a laser ultrasonic inspection;
fig. 9 is a schematic structural diagram of a remanufactured raw material evaluation device according to an embodiment of the present disclosure;
fig. 10 is a connection block diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The following detailed description will be provided with reference to the accompanying drawings and embodiments, so that how to apply the technical means to solve the technical problems and achieve the corresponding technical effects can be fully understood and implemented. The embodiments and various features in the embodiments of the present application can be combined with each other on the premise of no conflict, and the formed technical solutions are all within the protection scope of the present application.
It should be further noted that, for the convenience of description, only the portions relevant to the present disclosure are shown in the drawings.
Example one
The present embodiment provides a remanufactured raw material evaluation method, and fig. 1 is a flowchart of the remanufactured raw material evaluation method provided in the embodiment of the present application, and as shown in fig. 1, the method of the present embodiment includes:
and step 110, acquiring multi-angle image information of the remanufactured raw material.
Optionally, the remanufacturing raw materials include electromechanical product remanufacturing raw materials, which are taken as an example in this embodiment.
In some embodiments, the obtaining multi-angle image information of remanufactured raw material comprises:
and acquiring multi-angle image information of the remanufactured raw material acquired by a monocular camera and/or a binocular camera.
Optionally, a high-definition monocular high-definition camera or a binocular high-definition camera is used for collecting original image samples of the remanufactured raw materials. Carry out multi-angle, big, automatic shooing in batches to refabrication raw materials sample, ensure to include into the photo with the characteristics of refabrication raw materials under different situations are whole.
For example, a monocular high-definition camera is used for automatically photographing the remanufactured raw material in multiple angles, an original sample is collected, the basic characteristics of the remanufactured raw material are completely included in the photos, and the number of the photos is extremely large; the method comprises the steps of using a binocular camera to carry out multi-angle automatic photographing on remanufacturing raw materials, collecting multi-angle appearance information, integrating images collected by the monocular and binocular cameras, and establishing a three-dimensional model of a target part. And storing all the acquired images and the established three-dimensional model of the target part in an information statistical database. Optionally, the data acquisition process further comprises acquiring an original and complete corrosion-free state of the remanufactured raw material, and the type of the part of the remanufactured raw material can be marked in the manual marking process after the data acquisition. And summarizing and manually labeling the image data acquired by the camera, generating a corresponding three-dimensional model and storing the three-dimensional model as an original new product model in an information statistical database.
And step 120, determining explicit damage information in the multi-angle image information through a preset deep learning Network, wherein a trunk feature extraction Network of the preset deep learning Network comprises Mobilenet, a reinforced feature extraction Network of the preset deep learning Network comprises Spatial Pyramid Network (SPP), and a Path fusion Network of the preset deep learning Network comprises PANet (PANet).
As can be understood by those skilled in the art, the MobileNet network belongs to a lightweight CNN (Convolutional Neural Networks, CNN for short) dedicated to a mobile terminal or an embedded device, and compared with a conventional Convolutional Neural network, model parameters and computation amount are greatly reduced on the premise of a small reduction in accuracy.
In some embodiments, the explicit impairment information comprises:
first and second dominant impairment information; wherein the content of the first and second substances,
the first dominant impairment information comprises: rust information and/or corrosion information; the second dominant impairment information comprises: wear information and/or deformation information.
Optionally, a trained preset deep learning network is used, the three-dimensional morphology of the surface of the remanufactured raw material is identified through the multi-angle image information of the remanufactured raw material, and the dominant damage information such as corrosion, abrasion and deformation is obtained.
For example, when the surface corrosion detection is carried out, a monocular high-definition camera can be used, various deep learning networks and image processing technologies are combined to identify the surface of an object to be detected, and the main identification target is obvious characteristics visible to naked eyes, such as rust, corrosion and the like on the surface of the remanufactured raw material.
For example, when deformation and abrasion detection is performed, a binocular high-definition camera is used, the identification result of the surface corrosion detection is loaded and used, basic information such as dominant damage of an object to be detected, such as abrasion and deformation, is identified, and a qualitative and quantitative analysis method for identifying dominant damage of the object, such as corrosion, abrasion and micro-deformation, is researched.
Optionally, the type of the remanufactured raw material is identified through images acquired by a monocular camera in the data acquisition process, then the corrosion damage is identified, a three-dimensional model is established through images acquired by a binocular camera, and then the three-dimensional model is compared with an original new product model in an information statistical database, so that damage information such as deformation and abrasion is identified.
Optionally, the preset deep learning network includes a combined deep learning network. The trunk feature extraction network combining multiple deep learning networks is a mobilene network, the enhanced feature extraction is a spatial pyramid network SPP and a path fusion network PANet.
Optionally, the trunk feature extraction network may refer to fig. 3, and fig. 3 is a schematic diagram of a structure of the trunk feature extraction network.
MobileNet builds lightweight deep neural networks based on a streamlined architecture using deep separable convolutions, and is mostly used for mobile and embedded vision applications. The network comprises a standard convolutional layer, a thirteen-layer deep convolutional layer and a fourteen-layer network layer. Considering that the standard convolution and the deep convolution have certain similarity, and the network only contains one standard convolution layer in total, the module design is mainly carried out aiming at the characteristic of the deep convolution, and the standard convolution is compatible.
Wherein the core part of the MobileNet is the depth separable convolution. The original convolution operation is decomposed into two smaller operations: deep convolution and point-by-point convolution, the convolution decomposition can refer to fig. 4, and fig. 4 is a schematic diagram of a decomposition structure of a convolution neural network.
Assume that the input feature map size is D B ×D B X I, size of output feature map is D B ×D B X.times.J. Wherein I represents the number of input channels, J represents the number of output channels, D A ×D A Product of the sizes of the convolution kernels, D B ×D B Representing the size product of the mapped feature map. For the standard convolution, the amount of operation is:
D A ×D A ×I×J×D B ×D B
for deep convolution, the amount of operation is:
D A ×D A ×I×D B ×D B
for point-by-point convolution, the amount of operation is:
I×J×D B ×D B
therefore, the total computation of the depth separable convolution is:
D A ×D A ×I×D B ×D B +I×J×D B ×D B
the comparison of the depth separable convolution operand with the standard convolution operand is:
in the above formula, the first term of the numerator is the depth convolution operand, the second term is the point-by-point convolution operand, and the denominator is the standard convolution operand; wherein D is A Is the convolution kernel size; d B Is the input picture size; i is the number of input channels; j is the number of output channels.
The activation function used is ReLU6, the mathematical expression is:
ReLU6＝min(6，max(0，x))
the depth separable convolution first uses depth convolution to separately convolve different input channels, and then uses point-by-point convolution to recombine the previous outputs, once to reduce the computational pressure.
The spatial pyramid structure SPP is composed of three largest pooling layers with different sizes and a jump link, wherein the SPP structure can refer to fig. 5, and fig. 5 is a schematic diagram of a network structure of enhanced features. The input images are reintegrated after being subjected to branch operation of different structures of SPP and then are transmitted to the next layer of network. Different features of the images can be extracted in a different-size pooling layer, and different global features and local features can be fused after operation and combination, so that the detection accuracy is enhanced.
Wherein the input data size is (c, h) in ，w in ) The number of channels, height, and width are indicated.
Kernel size:
step size:
SPP pooling calculation:
h new ＝2*P h +h in
w new ＝2*P w +w in
where ceil () denotes rounding up, floor () denotes rounding down, and n is the pooling number. K h Denotes the height of the nucleus, S h Denotes the step size in the height direction, P h Indicating the filling quantity in the height direction; k w Denotes the width of the nucleus, S w Denotes a step size in the width direction, P w Indicating the number of fills in the width direction; h is a total of new ，w new Respectively, height and width for the next layer of network computation. When the image input is 416 × 416, the size of the pooling layer is 13 × 13 at maximum. The improved algorithm still retains the SPP module with the structure of three maximal pooling of 5 × 5, 9 × 9, 13 × 13 and one hop connection, four in parallelAnd (4) branching.
The traditional fusion algorithm loses more information of bottom layer features in the operation process, and the principle of the PANet as a segmentation algorithm is that the consumption of the bottom layer features in the process of transmitting the bottom layer features to the top layer is reduced by adding a top-down feature fusion path. Fig. 6 is a reference drawing of the PANet structure, and fig. 6 is a schematic diagram of a feature fusion network structure. By connecting in parallel twice, substantially enough underlying features can be retained.
The purpose of using PANet is to promote information flow conduction within the example segmentation framework based on the boxed region. For example, the information path between the underlying feature and the overlying feature is established by enhancing the accurate positioning information flow in the lower layer through the bottom-up path, thereby enhancing the overall feature hierarchy.
The loss function is used for evaluating the degree of the deep learning model with different predicted values and actual values in the training process, and the better the loss function is, the better the performance of the model is. The working process is to calculate the difference value between the calculation result of each iteration of the deep learning network and the true value, and feed the difference value back to the deep learning network, so as to guide the next network training to be carried out in the direction of reducing the difference value.
Alternatively, the loss function used in this embodiment may be composed of 3 parts of classification loss, confidence loss, and coordinate regression loss. Wherein the coordinate regression loss is calculated using mean square error, and the others are calculated using cross entropy.
An alternative coordinate regression loss expression is shown below:
an alternative confidence loss expression is shown below:
an alternative classification loss expression is shown below:
to sum up, an alternative overall loss expression is shown as follows:
Loss＝L coord +L conf +L class
wherein, in the coordinate regression loss expression: lambda [ alpha ] coord As a weight of coordinate loss, x i 、y i To predict the frame center coordinates, w i 、h i Respectively the width and height of a prediction frame, S is a mesh division coefficient of an input picture, B is the number of the prediction frames,
the value of the jth frame in the ith grid is 1 when the jth frame corresponds to the prediction target, otherwise, the value is 0; in the confidence loss expression: />
The value of the jth frame in the ith grid is 1 when the jth frame does not correspond to the prediction target, otherwise, the value is 0, and lambda is noobj To lose weight for confidence, c i Prediction of frame confidence, c' i A true frame confidence; in the classification loss expression: classes is the prediction category, p is the prediction box category probability, and p' is the true box category probability.
The activation function used is h-swish, and an alternative mathematical expression is:
optionally, in the process of training the preset deep learning network, according to the collected multi-angle image information of the remanufactured sample, preliminary image processing is performed, manual marking is performed on the image, specific conditions and existing areas of features are divided, and a final data set is generated. The manual labeling does contain macroscopic dominant lesions and obvious features, and the labeling information is used as a label for training. The process is characterized in that the characteristic regions in the image sample are divided, namely, when the deep learning module learns, the deep learning module informs that the region has the target characteristic, and the region does not have the target characteristic, so that the target characteristic condition is directly defined. The labeling information includes: and the defect positions and defect types (including corrosion, abrasion, deformation and the like) are used as labels to meet the requirements of subsequent deep learning network learning and training. In addition, the combined image processing technology is used for carrying out operations such as image segmentation, binarization, equalization, image filtering and the like on the original samples, and the original samples are expanded under the condition of not damaging the characteristics of the original samples, so that the number of the original samples is doubled, and more characteristics are ensured to be extracted in the subsequent learning. And finally, repeatedly learning and training the manually marked final data set by using a combined deep learning network, extracting enough characteristics, identifying the dominant damage of the data set, and qualitatively and quantitatively analyzing the remanufacturing value of the data set.
Step 130, determining recessive damage information in the remanufactured feedstock by laser ultrasonic detection and/or electromagnetic ultrasonic detection.
In some embodiments, the determining recessive impairment information in the remanufactured feedstock by laser ultrasonication and/or electromagnetic ultrasonication comprises:
detecting the remanufactured raw material by the laser ultrasonic detection method to obtain first recessive damage information;
detecting the remanufactured raw material by the electromagnetic ultrasonic detection method to obtain second recessive damage information;
and taking the first recessive damage information and the second recessive damage information as the recessive damage information.
In some embodiments, the implicit injury information includes:
crack information and/or fatigue life information.
In this embodiment, the remanufactured material was inspected by generating ultrasonic pulses using a pulsed laser. After the sensor emits laser light, a thermo-elastic effect can be generated, or other substances around the tested material can be used for exciting ultrasonic waves. The method has the following advantages:
1) The long-distance detection can be realized, and the attenuation in the propagation process is small;
2) The remanufacturing raw materials do not need to be directly contacted during detection, and the safety factor can be effectively improved for remanufacturing raw materials or scenes of certain dangerous behaviors;
3) The spatial and temporal resolution is high, and the detection resolution is high.
Laser ultrasonic testing can carry out real-time, online detection to the refabrication raw materials under adverse circumstances, through quick ultrasonic scanning formation of image, realizes the intelligent recognition of recessive damage such as refabrication raw materials crackle, fatigue.
Further, because laser ultrasonic's sensitivity is limited, at the in-process that detects refabrication raw materials, slight, not obvious or receive the recessive damage that shelters from a bit, be difficult to by discernment, the information that laser ultrasonic detection returned is also incomplete in addition, so also can combine together with electromagnetic ultrasonic detection method again and use the detection effect that promotes recessive damage.
The electromagnetic ultrasonic technique is to excite and receive ultrasonic waves by means of electromagnetic induction. If high-frequency current is led into a coil close to the surface of the metal material to be detected, induced current with the same frequency is generated in the metal material to be detected, and a constant magnetic field is applied outside the object to be detected, the induced current generates Lorentz force with the same frequency, and the force acts on a metal lattice to be detected to trigger the periodic vibration of the crystal structure of the metal to be detected, so that ultrasonic waves are excited and generated.
Optionally, the electromagnetic ultrasonic detection method may be implemented by an electromagnetic ultrasonic detection module composed of a high-frequency coil, an external magnetic field, and the remanufactured raw material 3. When detection is carried out, the 3 parts participate together to complete the conversion of electromagnetic ultrasound among electricity, magnetism and sound. Through changing the coil structure and the placing position or adjusting the physical parameters of the high-frequency coil, the stress condition of the measured raw material can be quickly changed, so that different types of ultrasonic waves are generated.
Optionally, the processing step after the combination of the laser ultrasonic detection method and the electromagnetic ultrasonic detection method may include: firstly, detecting the remanufactured raw material by using a laser ultrasonic detection method, and then detecting by using an electromagnetic ultrasonic detection method, wherein on the basis, the invisible damage which is not obvious or is ignored by the laser ultrasonic detection method can be detected; and finally, combining the information returned by the laser ultrasonic detection method and the electromagnetic ultrasonic detection method together, and processing the information by a computer to restore complete recessive injury information.
The electromagnetic laser detection method and the laser ultrasonic detection method are combined for use, so that the intelligent identification of the hidden damages such as cracks, fatigue and the like of the remanufactured raw materials can be realized.
Alternatively, the electromagnetic ultrasonic testing can refer to fig. 7, and fig. 7 is a schematic diagram of an electromagnetic ultrasonic testing; the laser ultrasonic inspection can refer to fig. 8, and fig. 8 is a schematic diagram of the laser ultrasonic inspection.
And step 140, evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information.
In some embodiments, the evaluating the remanufactured raw material according to the overt damage information and the implicit damage information by a preset evaluation rule comprises:
judging the dominant damage information and the recessive damage information through a preset remanufacturing evaluation model;
and under the condition that the dominant impairment information meets a first preset condition or the recessive impairment information meets a second preset condition, judging that the remanufacturing raw material cannot be used for remanufacturing.
Optionally, the first predetermined condition includes that the dominant impairment is too severe, and the second predetermined condition includes that the stealth impairment is too severe.
For example, if the preset evaluation rule is satisfied, it is directly determined that the remanufacturing is not possible, and an optional preset evaluation rule includes:
1) Dominant lesions are too severe: for example, corrosion in excess of 75% of the original surface area, complete deformation, wear in excess of 50%, etc.;
2) Stealth damage is too severe: such as excessive internal aging, metal fatigue, deep cracking, etc.
It should be noted that the preset evaluation rule, the first preset condition and the second preset condition may be set according to actual requirements, and specific description is not provided herein.
For facilitating understanding of the technical solution of the present application, reference may also be made to fig. 2, and fig. 2 is a flowchart of another remanufactured raw material evaluation method provided in an embodiment of the present application. In fig. 2, "model loading component information" means: and training the labeled data by using a combined deep learning network, wherein the basic information acquisition of the previous step comprises a process of manually labeling the acquired data.
According to the remanufactured raw material evaluation method provided by the embodiment, basic information of an object is identified by using monocular machine vision and binocular machine vision identification technologies, and preliminary classification and diversion of waste parts are realized; intelligently identifying the three-dimensional surface morphology of the part, and researching and identifying a qualitative and quantitative analysis method of dominant damages such as target corrosion, abrasion, micro-deformation and the like; aiming at the detection of near-surface damage of parts, intelligent non-destructive detection technology based on laser ultrasonic and electromagnetic ultrasonic equipment is utilized to realize intelligent identification of hidden damage such as cracks, fatigue and the like of waste motor parts; and quantitatively evaluating the quality characteristics of the remanufactured parts by combining factors such as comprehensive failure rate, aging risk and the like, making a remanufacturing evaluation decision of the parts from the aspect of remanufacturing economic feasibility, and taking the remanufacturing evaluation decision as a remanufacturing criterion of the remanufactured raw material parts. The problem of excessively depending on manual work in the detection and evaluation process of the remanufactured raw materials is solved. Specifically, the method comprises the following steps: acquiring multi-angle image information of a remanufactured raw material; determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilene, a reinforced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet; determining recessive damage information in the remanufactured raw material by laser ultrasonic detection and/or electromagnetic ultrasonic detection; and evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information. The trunk feature extraction network used by the deep learning network saves resources, avoids invalid calculation, can quickly and effectively extract features, has smaller volume capacity than other trunk feature extraction networks, and improves the detection speed; the deep learning network uses a network module for enhancing feature extraction, so that the feature learning of a data set can be effectively enhanced, and the recognition effect is better; besides surface damage information of remanufactured raw materials, an acoustoelectric and optomagnetic multi-physical-parameter model is utilized, an intelligent nondestructive testing scheme of acoustoelectric and optomagnetic multi-physical parameters is disclosed, deeper recessive damage identification is carried out, and the identification types are more; different from the limit that the conventional target detection can only detect the inherent type of the object, the remanufacturing evaluation model is established, and a remanufacturing evaluation decision of the part is made from the aspect of remanufacturing economic feasibility.
Example two
The present embodiment provides a remanufactured raw material evaluation apparatus, which may be used to execute the embodiments of the method of the present application, and for details not disclosed in the embodiments of the apparatus, please refer to the embodiments of the method of the present application. Fig. 9 is a schematic structural diagram of a remanufactured raw material evaluating apparatus according to an embodiment of the present application, and as shown in fig. 9, a remanufactured raw material evaluating apparatus 900 according to this embodiment includes:
an obtaining module 901, configured to obtain multi-angle image information of a remanufactured raw material;
a first determining module 902, configured to determine dominant impairment information in the multi-angle image information through a preset deep learning network, where a trunk feature extraction network of the preset deep learning network includes Mobilenet, a reinforced feature extraction network of the preset deep learning network includes a spatial pyramid network SPP, and a path fusion network of the preset deep learning network includes PANet;
a second determining module 903, configured to determine recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method;
an evaluation module 904, configured to evaluate the remanufactured raw material according to the dominant impairment information and the recessive impairment information by a preset evaluation rule.
In some embodiments, the explicit impairment information comprises:
first and second dominant impairment information; wherein the content of the first and second substances,
the first dominant impairment information comprises: rust information and/or corrosion information; the second dominant impairment information comprises: wear information and/or deformation information.
In some embodiments, the collateral damage information includes:
crack information and/or fatigue life information.
In some embodiments, the second determining module 903 comprises: a first determining unit, a second determining unit, a confirming unit; wherein the content of the first and second substances,
the first determining unit is used for detecting the remanufactured raw material through the laser ultrasonic detection method to obtain first recessive damage information;
the second determining unit is used for detecting the remanufactured raw material through the electromagnetic ultrasonic detection method to obtain second recessive damage information;
a confirming unit, configured to use the first recessive damage information and the second recessive damage information as the recessive damage information.
In some embodiments, the evaluation module 904 includes a decision unit; wherein the content of the first and second substances,
the judging unit is used for judging the explicit damage information and the implicit damage information through a preset remanufacturing evaluation model; and under the condition that the dominant impairment information meets a first preset condition or the recessive impairment information meets a second preset condition, judging that the remanufacturing raw material cannot be used for remanufacturing.
In some embodiments, the acquisition module 901 is configured to acquire multi-angle image information of the remanufactured raw material acquired by a monocular camera and/or a binocular camera.
Those skilled in the art will appreciate that the configuration shown in fig. 9 is not intended to be limiting of the devices of the embodiments of the present application and may include more or fewer modules/units than those shown, or some modules/units may be combined, or a different arrangement of modules/units.
It should be noted that each of the modules/units may be a functional module or a program module, and may be implemented by software or hardware. For the modules/units implemented by hardware, the above modules/units may be located in the same processor; or the modules/units can be respectively positioned in different processors in any combination.
The device provided by the embodiment comprises: an obtaining module 901, configured to obtain multi-angle image information of a remanufactured raw material; a first determining module 902, configured to determine dominant impairment information in the multi-angle image information through a preset deep learning network, where a trunk feature extraction network of the preset deep learning network includes Mobilenet, a reinforced feature extraction network of the preset deep learning network includes a spatial pyramid network SPP, and a path fusion network of the preset deep learning network includes PANet; a second determining module 903, configured to determine recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method; an evaluation module 904, configured to evaluate the remanufactured raw material according to the dominant impairment information and the recessive impairment information by a preset evaluation rule. The trunk feature extraction network used by the deep learning network saves resources, avoids invalid calculation, can quickly and effectively extract features, has smaller volume capacity than other trunk feature extraction networks, and improves the detection speed; the deep learning network uses a network module for enhancing feature extraction, so that the feature learning of a data set can be effectively enhanced, and the recognition effect is better; besides surface damage information of remanufactured raw materials, an acoustoelectric and optomagnetic multi-physical-parameter model is utilized, an intelligent nondestructive testing scheme of acoustoelectric and optomagnetic multi-physical parameters is disclosed, deeper recessive damage identification is carried out, and the identification types are more; different from the limit that the conventional target detection only can detect the inherent type of the object, the remanufacturing evaluation model is established, and the remanufacturing evaluation decision of the part is made from the aspect of remanufacturing economic feasibility.
EXAMPLE III
This embodiment still provides a refabrication material evaluation system, the system includes:
the one or more cameras are in communication connection with the evaluation server and are used for acquiring multi-angle image information of the remanufactured raw material and sending the multi-angle image information to the evaluation server;
the evaluation server is used for determining dominant damage information in the multi-angle image information through a preset deep learning network, determining recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method, and evaluating the remanufactured raw material through a preset remanufacturing evaluation model according to the dominant damage information and the recessive damage information, wherein the trunk feature extraction network of the preset deep learning network comprises a Mobilenet, the reinforced feature extraction network of the preset deep learning network comprises a space pyramid network SPP, and the path fusion network of the preset deep learning network comprises a PANET.
It is noted that the one or more cameras may include one or more monocular or binocular high definition cameras, and the cameras and the evaluation service may communicate by wired or wireless means.
It can be clearly understood by those skilled in the art that, for convenience and simplicity of description, reference may be made to corresponding processes in the foregoing method embodiments for specific working processes of each device or each module in the system, and repeated descriptions are not repeated in this embodiment.
Example four
The present embodiment further provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the method steps in the foregoing method embodiments can be implemented, and the description of the present embodiment is not repeated herein.
The computer-readable storage medium may also include, among other things, a computer program, a data file, a data structure, etc., either alone or in combination. The computer-readable storage medium or computer program may be specifically designed and understood by those skilled in the art of computer software, or the computer-readable storage medium may be known and available to those skilled in the art of computer software. Examples of computer-readable storage media include: magnetic media such as hard disks, floppy disks, and magnetic tape; optical media, such as CDROM disks and DVDs; magneto-optical media, e.g., optical disks; and hardware devices specifically configured to store and execute computer programs, e.g., read Only Memory (ROM), random Access Memory (RAM), flash memory; or a server, app application mall, etc. Examples of computer programs include both machine code, such as produced by a compiler, and files containing higher level code that may be executed by the computer using an interpreter. The described hardware devices may be configured to act as one or more software modules to perform the operations and methods described above, and vice versa. In addition, the computer-readable storage medium can be distributed over network-coupled computer systems and can store and execute program code or computer programs in a distributed fashion.
EXAMPLE five
Fig. 10 is a connection block diagram of an electronic device according to an embodiment of the present application, and as shown in fig. 10, the electronic device 1000 may include: one or more processors 1001, memory 1002, multimedia components 1003, input/output (I/O) interfaces 1004, and communication components 1005.
Wherein the one or more processors 1001 are configured to perform all or part of the steps of the method embodiments as described above. The memory 1002 is used to store various types of data, which may include, for example, instructions for any application or method in the electronic device, as well as application-related data.
The one or more processors 1001 may be implemented as an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components for performing the methods as in the foregoing method embodiments.
The Memory 1002 may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as a Static Random Access Memory (SRAM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), an Erasable Programmable Read-Only Memory (EPROM), a Programmable Read-Only Memory (PROM), a Read-Only Memory (ROM), a magnetic Memory, a flash Memory, a magnetic disk, or an optical disk.
The multimedia component 1003 may include a screen, which may be a touch screen, and an audio component for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may further be stored in a memory or transmitted through a communication component. The audio assembly further comprises at least one speaker for outputting audio signals.
The I/O interface 1004 provides an interface between the one or more processors 1001 and other interface modules, such as a keyboard, mouse, buttons, and the like. These buttons may be virtual buttons or physical buttons.
The communication component 1005 is used for wired or wireless communication between the electronic device 1000 and other devices. The wired communication includes communication through a network port, a serial port and the like; the wireless communication includes: wi-Fi, bluetooth, near Field Communication (NFC for short), 2G, 3G, 4G, 5G, or a combination of one or more of them. The corresponding communication component 1005 may thus include: wi-Fi module, bluetooth module, NFC module.
In summary, the present application provides a remanufacturing raw material evaluation method, apparatus, system, storage medium, and electronic device. The remanufactured raw material evaluation method adopts monocular machine vision and binocular machine vision identification technologies to identify basic object information and realize primary classification and diversion of waste parts; intelligently identifying the three-dimensional surface morphology of the part, and researching and identifying a qualitative and quantitative analysis method of dominant damages such as target corrosion, abrasion, micro-deformation and the like; aiming at the detection of the near-surface damage of the parts, the intelligent nondestructive detection technology based on laser ultrasonic and electromagnetic ultrasonic equipment is utilized to realize the intelligent identification of hidden damages such as cracks, fatigue and the like of the waste motor parts; and quantitatively evaluating the quality characteristics of the remanufactured parts by combining factors such as comprehensive failure rate, aging risk and the like, and making a remanufacturing evaluation decision of the parts from the aspect of remanufacturing economic feasibility to serve as a remanufacturing criterion of the remanufactured raw material parts. The problem of excessively rely on artifically in the detection evaluation process to the refabrication raw materials is solved. Specifically, the method comprises the following steps: acquiring multi-angle image information of a remanufactured raw material; determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilene, a reinforced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet; determining recessive damage information in the remanufactured raw material by laser ultrasonic detection and/or electromagnetic ultrasonic detection; and evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information. The trunk feature extraction network used by the deep learning network saves resources, avoids invalid calculation, can quickly and effectively extract features, has smaller volume capacity than other trunk feature extraction networks, and improves the detection speed; the deep learning network uses a network module for enhancing feature extraction, so that the feature learning of a data set can be effectively enhanced, and the recognition effect is better; besides surface damage information of remanufactured raw materials, an acoustoelectric and optomagnetic multi-physical-parameter model is utilized, an intelligent nondestructive testing scheme of acoustoelectric and optomagnetic multi-physical parameters is disclosed, deeper recessive damage identification is carried out, and the identification types are more; different from the limit that the conventional target detection only can detect the inherent type of the object, the remanufacturing evaluation model is established, and the remanufacturing evaluation decision of the part is made from the aspect of remanufacturing economic feasibility.
It should be further understood that the method or system disclosed in the embodiments provided in the present application may be implemented in other ways. The method or system embodiments described above are merely illustrative, for example, the flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of methods and apparatus according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, a computer program segment, or a portion of a computer program, which comprises one or more computer programs for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures, or indeed, may be executed substantially concurrently, or in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer programs.
In this application, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, apparatus or device that comprises the element; if the description to "first", "second", etc. is used for descriptive purposes only, it is not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of the technical features indicated; in the description of the present application, the terms "plurality" and "plurality" mean at least two unless otherwise specified; if a server is described, it should be noted that the server may be an independent physical server or terminal, or may be a server cluster formed by a plurality of physical servers, or may be a cloud server capable of providing basic cloud computing services such as a cloud server, a cloud database, a cloud storage, a CDN, and the like; if an intelligent terminal or a mobile device is described in the present application, it should be noted that the intelligent terminal or the mobile device may be a mobile phone, a tablet Computer, a smart watch, a netbook, a wearable electronic device, a Personal Digital Assistant (PDA), an Augmented Reality (AR), a Virtual Reality (VR), a smart television, a smart audio, a Personal Computer (PC), and the like, but is not limited thereto, and the specific form of the intelligent terminal or the mobile device is not particularly limited in the present application.
Finally, it is noted that in the description of the present specification, reference to the description of "one embodiment," "some embodiments," "an example" or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the present application have been shown and described above, it is to be understood that the above embodiments are exemplary and that the description is made only for the sake of understanding the present application and not intended to limit the present application. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the disclosure as defined by the appended claims.
Claims (10)
1. A remanufactured feedstock evaluation method, the method comprising:
acquiring multi-angle image information of remanufactured raw materials;
determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilene, a reinforced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet;
determining recessive damage information in the remanufactured raw material by laser ultrasonic detection and/or electromagnetic ultrasonic detection;
and evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information.
2. The remanufactured feedstock evaluation method of claim 1, wherein the overt impairment information comprises:
first and second dominant impairment information; wherein the content of the first and second substances,
the first dominant impairment information comprises: rust information and/or corrosion information; the second dominant impairment information comprises: wear information and/or deformation information.
3. The remanufactured feedstock evaluation method of claim 1, wherein the latent damage information comprises:
crack information and/or fatigue life information.
4. The remanufactured feedstock evaluation method of claim 1, wherein the determining the recessive damage information in the remanufactured feedstock by laser sonication and/or electromagnetic sonication comprises:
detecting the remanufactured raw material by the laser ultrasonic detection method to obtain first recessive damage information;
detecting the remanufactured raw material by the electromagnetic ultrasonic detection method to obtain second recessive damage information;
and taking the first recessive damage information and the second recessive damage information as the recessive damage information.
5. The remanufactured material evaluation method of claim 1, wherein the evaluating the remanufactured material according to the overt damage information and the recessive damage information through a preset evaluation rule comprises:
judging the dominant damage information and the recessive damage information through a preset remanufacturing evaluation model;
and under the condition that the dominant impairment information meets a first preset condition or the recessive impairment information meets a second preset condition, judging that the remanufacturing raw material cannot be used for remanufacturing.
6. The remanufactured material evaluation method of claim 1, wherein the obtaining multi-angle image information of the remanufactured material comprises:
and acquiring multi-angle image information of the remanufactured raw material acquired by a monocular camera and/or a binocular camera.
7. A remanufactured feedstock evaluation device comprising:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring multi-angle image information of the remanufactured raw material;
the first determining module is used for determining dominant impairment information in the multi-angle image information through a preset deep learning network, wherein a trunk feature extraction network of the preset deep learning network comprises a Mobilenet, an enhanced feature extraction network of the preset deep learning network comprises a spatial pyramid network (SPP), and a path fusion network of the preset deep learning network comprises a PANet;
the second determination module is used for determining recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method;
and the evaluation module is used for evaluating the remanufactured raw material through a preset evaluation rule according to the dominant damage information and the recessive damage information.
8. A remanufactured feedstock evaluation system comprising:
the system comprises one or more cameras, an evaluation server and a control server, wherein the cameras are in communication connection with the evaluation server and used for collecting multi-angle image information of remanufactured raw materials and sending the multi-angle image information to the evaluation server;
the evaluation server is used for determining dominant damage information in the multi-angle image information through a preset deep learning network, determining recessive damage information in the remanufactured raw material through a laser ultrasonic detection method and/or an electromagnetic ultrasonic detection method, and evaluating the remanufactured raw material through a preset remanufacturing evaluation model according to the dominant damage information and the recessive damage information, wherein the trunk feature extraction network of the preset deep learning network comprises a Mobilenet, the reinforced feature extraction network of the preset deep learning network comprises a space pyramid network SPP, and the path fusion network of the preset deep learning network comprises a PANET.
9. A computer-readable storage medium storing a computer program which, when executed by one or more processors, implements the method of any one of claims 1-6.
10. An electronic device comprising a memory and one or more processors, the memory having stored thereon a computer program, the memory and the one or more processors being communicatively connected to each other, the computer program, when executed by the one or more processors, performing the method of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211559344.7A CN115984843A (en) | 2022-12-06 | 2022-12-06 | Remanufacturing raw material evaluation method and device, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211559344.7A CN115984843A (en) | 2022-12-06 | 2022-12-06 | Remanufacturing raw material evaluation method and device, storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115984843A true CN115984843A (en) | 2023-04-18 |
Family
ID=85971369
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211559344.7A Pending CN115984843A (en) | 2022-12-06 | 2022-12-06 | Remanufacturing raw material evaluation method and device, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115984843A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111862067A (en) * | 2020-07-28 | 2020-10-30 | 中山佳维电子有限公司 | Welding defect detection method and device, electronic equipment and storage medium |
| CN111862083A (en) * | 2020-07-31 | 2020-10-30 | 中国矿业大学 | Comprehensive monitoring system and method for steel wire rope state based on vision-electromagnetic detection |
| CN112581430A (en) * | 2020-12-03 | 2021-03-30 | 厦门大学 | Deep learning-based aeroengine nondestructive testing method, device, equipment and storage medium |
| CN113469953A (en) * | 2021-06-10 | 2021-10-01 | 南昌大学 | Transmission line insulator defect detection method based on improved YOLOv4 algorithm |
| CN113834824A (en) * | 2021-09-26 | 2021-12-24 | 河北省特种设备监督检验研究院 | Elevator traction steel belt defect detection device based on vision and electromagnetic technology |
| WO2022105117A1 (en) * | 2020-11-17 | 2022-05-27 | 平安科技（深圳）有限公司 | Method and device for image quality assessment, computer device, and storage medium |
-
2022
- 2022-12-06 CN CN202211559344.7A patent/CN115984843A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111862067A (en) * | 2020-07-28 | 2020-10-30 | 中山佳维电子有限公司 | Welding defect detection method and device, electronic equipment and storage medium |
| CN111862083A (en) * | 2020-07-31 | 2020-10-30 | 中国矿业大学 | Comprehensive monitoring system and method for steel wire rope state based on vision-electromagnetic detection |
| WO2022105117A1 (en) * | 2020-11-17 | 2022-05-27 | 平安科技（深圳）有限公司 | Method and device for image quality assessment, computer device, and storage medium |
| CN112581430A (en) * | 2020-12-03 | 2021-03-30 | 厦门大学 | Deep learning-based aeroengine nondestructive testing method, device, equipment and storage medium |
| CN113469953A (en) * | 2021-06-10 | 2021-10-01 | 南昌大学 | Transmission line insulator defect detection method based on improved YOLOv4 algorithm |
| CN113834824A (en) * | 2021-09-26 | 2021-12-24 | 河北省特种设备监督检验研究院 | Elevator traction steel belt defect detection device based on vision and electromagnetic technology |
Non-Patent Citations (1)
| Title |
|---|
| 徐春广 李卫彬: "《超声波检测基础》", 北京理工大学出版社, pages: 345 - 355 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110779937B (en) | Casting product internal defect intelligent detection device | |
| US11341626B2 (en) | Method and apparatus for outputting information | |
| CN109671071B (en) | Underground pipeline defect positioning and grade judging method based on deep learning | |
| US20130034305A1 (en) | Image-based crack quantification | |
| CN108711148B (en) | Tire defect intelligent detection method based on deep learning | |
| CN115829999A (en) | Insulator defect detection model generation method, device, equipment and storage medium | |
| CN110910343A (en) | Method and device for detecting pavement cracks and computer equipment | |
| CN114240821A (en) | Weld defect detection method based on improved YOLOX | |
| CN109284779A (en) | Object detecting method based on the full convolutional network of depth | |
| CN110009614A (en) | Method and apparatus for output information | |
| Garfo et al. | Defect detection on 3d print products and in concrete structures using image processing and convolution neural network | |
| CN113822880A (en) | Crack identification method based on deep learning | |
| CN115830407A (en) | Cable pipeline fault discrimination algorithm based on YOLOV4 target detection model | |
| CN115439395A (en) | Defect detection method and device for display panel, storage medium and electronic equipment | |
| CN111680575A (en) | Human epithelial cell staining and classifying device, equipment and storage medium | |
| CN111160100A (en) | Lightweight depth model aerial photography vehicle detection method based on sample generation | |
| Bhamare et al. | Quality Inspection of Tire using Deep Learning based Computer Vision | |
| CN117011274A (en) | Automatic glass bottle detection system and method thereof | |
| CN115984843A (en) | Remanufacturing raw material evaluation method and device, storage medium and electronic equipment | |
| CN112052730A (en) | 3D dynamic portrait recognition monitoring device and method | |
| CN116977271A (en) | Defect detection method, model training method, device and electronic equipment | |
| CN116433840A (en) | Method for synthesizing and detecting specific damage image of ballastless track plate fastener | |
| CN111027601A (en) | Plane detection method and device based on laser sensor | |
| CN116109813A (en) | Anchor hole drilling identification method, system, electronic equipment and medium | |
| CN114998222A (en) | Automobile differential shell surface detection method, electronic equipment and medium |
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 |