CN113743525A - Fabric material identification system and method based on luminosity stereo - Google Patents
Fabric material identification system and method based on luminosity stereo Download PDFInfo
- Publication number
- CN113743525A CN113743525A CN202111074359.XA CN202111074359A CN113743525A CN 113743525 A CN113743525 A CN 113743525A CN 202111074359 A CN202111074359 A CN 202111074359A CN 113743525 A CN113743525 A CN 113743525A
- Authority
- CN
- China
- Prior art keywords
- fabric
- module
- image
- identification
- support vector
- 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.)
- Granted
Links
- 239000004744 fabric Substances 0.000 title claims abstract description 89
- 239000000463 material Substances 0.000 title claims abstract description 43
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 25
- 238000002310 reflectometry Methods 0.000 claims abstract description 24
- 238000000605 extraction Methods 0.000 claims abstract description 18
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 7
- 238000012706 support-vector machine Methods 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 17
- 239000013598 vector Substances 0.000 claims description 14
- 238000013507 mapping Methods 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 2
- 238000002372 labelling Methods 0.000 claims 1
- 238000012360 testing method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 229920000742 Cotton Polymers 0.000 description 1
- 239000004677 Nylon Substances 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 229920001778 nylon Polymers 0.000 description 1
- 229920000728 polyester Polymers 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 210000002268 wool Anatomy 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a fabric material identification system and method based on photometric stereo. The system comprises an acquisition module, an image processing module, a feature extraction module and an identification module. Wherein the acquisition device is provided with light sources and cameras in different directions in a light-tight sealed box. The camera collects images of fabrics made of different materials under different light sources through a shooting window at the bottom end. The image processing module calculates a reflectivity map and a normal map of the image through a photometric stereo method according to the image and corresponding light information, inputs the reflectivity map and the normal map into the feature extraction module to extract features through a convolutional neural network, and finally uses the identification module to classify the material of the fabric. A fabric material identification method based on photometric stereo comprises image acquisition and processing, feature extraction and classification identification, and the method utilizes concave-convex shape information on the surface of a fabric to quickly and accurately obtain the fabric material, thereby saving time and labor.
Description
Technical Field
The invention belongs to the technical field of data processing, relates to an artificial intelligence deep learning data processing method, and particularly relates to a fabric material identification system and method based on photometric stereo.
Background
The fabric material identification and classification technology has wide application scenes in production and life, and comprises the fields of robot design and industrial detection. The technology can provide a means for analyzing the material attribute of the fabric, help is provided for identifying and classifying the material of the fabric, and decision-making efficiency is improved.
Because fabric pictures are affected by various factors such as shape, reflective characteristics, lighting, and viewing angle, identifying fabric material from a picture is a challenging task. The existing method for identifying the fabric material needs to use multiple cameras to obtain multi-angle images, so that the equipment is complex, the using process is complex, the images collected by different cameras relate to the corresponding problem, and the rapid identification of multiple fabrics is difficult to realize. Meanwhile, because the material of the fabric is not only expressed on the color of the fabric, but also more expressed on the concave-convex shape of the surface of the fabric, the microscopic geometry of the surface of the fabric also contains information useful for identifying and classifying the fabric, and the current identification method is lack of acquisition and use of the information.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a fabric material identification system and method based on photometric stereo, which comprises the steps of shooting fabric images under a plurality of different light sources by using an acquisition module, analyzing microcosmic geometric shape information of the surface of the fabric by using an identification module, and identifying the fabric material.
A fabric material identification system based on photometric stereo comprises an acquisition module, an image processing module, a feature extraction module and an identification module.
The acquisition module comprises a light-tight seal box, a camera and a light emitting array. The bottom of the sealed box is provided with a shooting opening, and the camera and the light emitting array are fixed in the sealed box. The shooting direction of the camera is aligned with the shooting port. The light emitting array is used for providing light rays in different directions. Tightly cover the surface fabric surface of treating discernment with collection module, shoot the picture under a plurality of different angles light to a surface fabric to after marking the light direction, transmit for image processing module.
Preferably, the light emitting array is fixed between the camera head and the imaging port of the camera, and includes a plurality of light emitting diodes having different orientations.
Preferably, the acquisition module further comprises a microprocessor fixed inside the sealed box, the microprocessor is used for controlling the light emitting array and the camera, and wirelessly transmitting the obtained picture and the corresponding light direction to the image processing module.
The image processing module calculates the reflectivity and the normal direction of each pixel point in the image according to a plurality of images of the same fabric and corresponding light direction information through a photometric stereo method, and obtains a reflectivity graph and a normal map of a three-dimensional shape after the normal directions are collected.
The characteristic extraction module inputs the reflectivity map and the normal mapping obtained by the image processing module into the convolutional neural network, and extracts characteristic vectors capable of reflecting the three-dimensional microstructure of the surface of the fabric. The identification module identifies the feature vectors obtained by the feature extraction module to complete material classification of the fabric.
Preferably, the convolutional neural network is a VGG-M model initialized by using the disclosed pre-training parameters, the input of the VGG-M model is a reflectivity map and a normal map, and the output is a feature vector.
Preferably, the identification module comprises K support vector machines, and the material classification of the fabric is realized by using a one-to-many support vector machine method. And K is the material type of the fabric to be identified.
A fabric material identification method based on photometric stereo comprises the following steps:
step one, image acquisition
A large number of fabric images are collected, and a plurality of images of the same fabric at the same angle and in different light directions are used as a sample. And marking the label of the sample as a training set according to the material of the fabric.
Preferably, at least 3 images in different light directions are acquired from the same fabric.
Step two, image processing
And (3) calculating the reflectivity and the normal direction of each pixel point in the image according to the light direction information of the sample in the training set by using a photometric stereo method, and summarizing the normal directions to obtain a reflectivity graph and a normal mapping of the three-dimensional shape of the sample.
Step three, feature extraction
And inputting the reflectivity map and the normal line mapping obtained in the step two into the VGG-M model, and outputting the feature vector of the corresponding sample by the VGG-M model.
Step four, classification training
And (4) performing classification training by using K support vector machines aiming at samples of K different labels. Defining the positive class of 1 support vector machine as a label, wherein the rest labels are the negative classes of the support vector machine, and the positive class labels of the K support vector machines are not repeated. And C, respectively inputting the feature vectors obtained in the step three into K support vector machines, and taking the class corresponding to the support vector machine with the maximum classification value as a classification result of the sample.
Step five, identifying the fabric
And (4) collecting a fabric image without a label and to be identified, processing the fabric image according to the step two, inputting the processed fabric image into the VGG-M model for feature extraction, inputting the processed fabric image into the K support vector machines trained in the step four, obtaining labels corresponding to the fabric, and completing identification.
The invention has the following beneficial effects:
the acquisition module in the method is a sealed box with a light source and is matched with a monocular camera for shooting, so that a multi-light source image of the fabric to be recognized can be simply and quickly acquired, and an original image is provided for a subsequent processing module to recognize the fabric material. Compared with the existing similar device, the module has the advantages of low manufacturing cost, small size, light weight and easy use, and the device can be switched between the fabrics to be identified by simply lifting and moving the device; meanwhile, the sealing box is made of a lightproof material, is not influenced by an external light source in use, has low requirements on the use environment, and avoids a complex device structure caused by complex requirements on input pictures.
Drawings
FIG. 1 is a schematic structural diagram of an acquisition module in an embodiment;
FIG. 2 is a schematic view of a camera port of the acquisition module in the embodiment;
FIG. 3 is a flow chart of a fabric material identification method.
Detailed Description
The invention is further explained below with reference to the drawings;
a fabric material identification system based on photometric stereo comprises an acquisition module, an image processing module, a feature extraction module and an identification module.
As shown in fig. 1, the acquisition module comprises a sealed box, a camera 2, a microprocessor circuit board 3 and a light emitting array 5. The sealed box is composed of a square shell 1 which is closed at the top end and is not transparent and a matrix base 4. The square housing 1 is fixed on a rectangular base 4. The camera 2, microprocessor 3 and light emitting array 5 are fixed inside the block shaped housing 1. As shown in fig. 2, a photographing opening is provided in the center of the rectangular base 4. The shooting direction of the camera 2 is opposite to the shooting window in the center of the rectangular base 4. The light emitting array 5 is located between the camera 2 and the photographing window. The light emitting array 5 comprises four sets of differently oriented light emitting diodes. The microprocessor circuit board 3 controls the light emitting diodes to switch on and off states, and light rays in different directions can be provided for the acquisition module. The fabric is tiled on a desktop, the rectangular base 4 is stably placed on the surface of the fabric, the microprocessor circuit board 3 controls the camera 3 to shoot pictures of the same fabric when different light-emitting diodes are on, and the pictures are transmitted to a computer through the universal bus serial interface after the light direction is marked, so that subsequent image processing and identification are carried out.
Photometric stereo is a method in the field of computer vision, which uses three or more pictures from the same viewing angle with different light sources to estimate the reflectivity and normal direction of each pixel point in the image. The image processing module calculates the reflectivity and the normal direction of the fabric according to the pictures and light information collected by the collecting module through a photometric stereo method, and obtains a reflectivity graph and a normal map of a three-dimensional shape after the normal directions are collected, wherein the reflectivity graph and the normal map can reflect the microcollection shape of the surface of the fabric.
The characteristic extraction module inputs the reflectivity map and the normal line mapping obtained by the image processing module into the convolutional neural network to analyze the three-dimensional microstructure of the fabric. The convolutional neural network is a VGG-M model and comprises 5 convolutional layers and 3 full-connection layers, and the initialization parameters in the model use the public pre-training parameters of the VGG-M. The model output may represent a feature vector of the three-dimensional microstructure.
The identification module realizes material classification of the fabric by using a one-to-many support vector machine method, and the identification module comprises K support vector machines, wherein K is the material type of the fabric to be identified. When the one-to-many support vector machine method classifies K classes, K support vector machines are trained, for each support vector machine, only one class exists in a positive example, and the rest K-1 classes are all used as negative classes. And the positive classes of the K support vector machines correspond to the K classes one by one, when the support vector machines are used for prediction, samples to be predicted are respectively input into the K support vector machines, and the class corresponding to the support vector machine with the largest classification value is used as a classification result of the samples. The identification module identifies the feature vectors obtained by the feature extraction module to complete material classification of the fabric.
As shown in fig. 3, a fabric material identification method based on photometric stereo includes the following steps:
step one, image acquisition
The method comprises the steps of collecting images of fabrics made of different materials, and taking 4 images of the same fabric at the same angle and in different light directions as a sample. The label of the sample is marked according to the materials on the fabric component label, and comprises 5 types of cotton cloth, nylon, wool, silk and polyester fiber. The collected images were divided into training and test sets at a 3:1 ratio using a non-return random sampling method.
Step two, image processing
And (3) calculating the reflectivity and the normal direction of each pixel point in the image according to the light direction information of the sample in the training set by using a photometric stereo method, and summarizing the normal directions to obtain a reflectivity graph and a normal map of the three-dimensional shape of the sample.
Step three, feature extraction
And inputting the reflectivity map and the normal line mapping obtained in the step two into the VGG-M model, and outputting the feature vector of the corresponding sample by the VGG-M model.
Step four, classification training
And (3) carrying out classification training on samples of 5 different labels in the training set by using 5 support vector machines. Defining the positive class of 1 support vector machine as a label, the rest 4 labels are the negative classes of the support vector machine, and the positive class labels of 5 support vector machines are not repeated. And (4) respectively inputting the feature vectors obtained in the step three into 5 support vector machines, and taking the class corresponding to the support vector machine with the maximum classification value as a classification result of the sample.
Step five, identifying the fabric
And (4) processing the samples in the test set according to the step two, inputting the samples into a VGG-M model for feature extraction, inputting the samples into K support vector machines trained in the step four to obtain labels corresponding to the fabric, and testing the classification performance of the support vector machines.
Claims (7)
1. The utility model provides a surface fabric material identification system based on luminosity is three-dimensional which characterized in that: the system comprises an acquisition module, an image processing module, a feature extraction module and an identification module;
the acquisition module comprises a lightproof seal box, a camera and a light emitting array; the bottom of the sealed box is provided with a shooting port, and the camera and the light emitting array are fixed in the sealed box; the shooting direction of the camera is aligned with the shooting port; the light emitting array is used for providing light rays in different directions; the bottom of the seal box is tightly and stably placed on the surface of the fabric to be identified, a plurality of pictures under light rays of different angles are shot for one fabric, and the pictures are transmitted to the image processing module after the light ray directions are marked;
the image processing module calculates the reflectivity and the normal direction of each pixel point in the image according to a plurality of images of the same fabric and corresponding light direction information by a photometric stereo method, and obtains a reflectivity graph and a normal map of a three-dimensional shape after the normal directions are summarized;
the characteristic extraction module inputs the reflectivity map and the normal mapping obtained by the image processing module into a convolutional neural network, and extracts characteristic vectors capable of reflecting the three-dimensional microstructure of the surface of the fabric; the identification module identifies the feature vectors obtained by the feature extraction module to complete material classification of the fabric.
2. The system for recognizing fabric materials based on photometric stereo according to claim 1, wherein: the light emitting array is fixed between a camera of the camera and the shooting port and comprises a plurality of light emitting diodes with different orientations.
3. A fabric material identification system based on photometric stereo according to claim 1 or 2 characterized by: the acquisition module also comprises a microprocessor fixed in the sealed box, the microprocessor controls the camera to shoot pictures when the light emitting diodes in different directions in the light emitting array are respectively lighted, and the obtained pictures and the corresponding light direction are transmitted to the image processing module.
4. The system for recognizing fabric materials based on photometric stereo according to claim 1, wherein: the convolutional neural network is a VGG-M model initialized by using the public pre-training parameters; the VGG-M model comprises 5 convolutional layers and 3 fully-connected layers, the input is a reflectivity map and a normal line mapping, and the output is a feature vector.
5. The system for recognizing fabric materials based on photometric stereo according to claim 1, wherein: the identification module performs training on K support vector machines by using a one-to-many method so as to realize material classification of the fabric; and K is the material type of the fabric to be identified.
6. A fabric material identification method based on luminosity stereo is characterized in that: the method uses the identification system of claims 1-5, comprising the steps of:
step one, image acquisition
Collecting a large number of fabric images, and taking a plurality of images of the same fabric at the same angle and in different light directions as a sample; labeling labels of the samples according to the material of the fabric, and enabling the samples and the labels to be in one-to-one correspondence to serve as a training set;
step two, image processing
Calculating the reflectivity and the normal direction of each pixel point in the image according to the light direction information of the sample in the training set by using a photometric stereo method, and summarizing the normal directions to obtain a reflectivity graph and a normal map of the three-dimensional shape of the sample;
step three, feature extraction
Inputting the reflectivity map and the normal mapping obtained in the step two into a VGG-M model, and outputting the feature vector of the corresponding sample by the VGG-M model;
step four, classification training
Performing classification training by using K support vector machines aiming at samples of K different labels; defining the positive class of 1 support vector machine as a label, wherein the rest labels are the negative classes of the support vector machine, and the positive class labels of K support vector machines are not repeated; inputting the feature vectors obtained in the step three into K support vector machines respectively, and taking the category corresponding to the support vector machine with the largest classification value as a classification result of the sample;
step five, identifying the fabric
And (4) collecting a fabric image without a label and to be identified, processing the fabric image according to the step two, inputting the processed fabric image into the VGG-M model for feature extraction, inputting the processed fabric image into the K support vector machines trained in the step four, obtaining labels corresponding to the fabric, and completing identification.
7. The method for recognizing the fabric material based on the photometric stereo as recited in claim 6, wherein: at least 3 images in different light directions are collected from the same fabric.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111074359.XA CN113743525B (en) | 2021-09-14 | 2021-09-14 | Fabric material identification system and method based on luminosity three-dimensional |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111074359.XA CN113743525B (en) | 2021-09-14 | 2021-09-14 | Fabric material identification system and method based on luminosity three-dimensional |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113743525A true CN113743525A (en) | 2021-12-03 |
| CN113743525B CN113743525B (en) | 2024-02-13 |
Family
ID=78738692
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111074359.XA Active CN113743525B (en) | 2021-09-14 | 2021-09-14 | Fabric material identification system and method based on luminosity three-dimensional |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113743525B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2606256A (en) * | 2021-03-12 | 2022-11-02 | Frontier Cool Inc | Fabric informaiton digitization system and method thereof |
| CN116740581A (en) * | 2023-08-16 | 2023-09-12 | 深圳市欢创科技有限公司 | Method for determining material identification model, method for returning to base station and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101069417A (en) * | 2004-11-30 | 2007-11-07 | 松下电器产业株式会社 | Image processing method, image processing apparatus, image processing program, and image file format |
| CN102095731A (en) * | 2010-12-02 | 2011-06-15 | 山东轻工业学院 | System and method for recognizing different defect types in paper defect visual detection |
| WO2021169209A1 (en) * | 2020-02-27 | 2021-09-02 | 平安科技（深圳）有限公司 | Method, apparatus and device for recognizing abnormal behavior on the basis of voice and image features |
-
2021
- 2021-09-14 CN CN202111074359.XA patent/CN113743525B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101069417A (en) * | 2004-11-30 | 2007-11-07 | 松下电器产业株式会社 | Image processing method, image processing apparatus, image processing program, and image file format |
| CN102095731A (en) * | 2010-12-02 | 2011-06-15 | 山东轻工业学院 | System and method for recognizing different defect types in paper defect visual detection |
| WO2021169209A1 (en) * | 2020-02-27 | 2021-09-02 | 平安科技（深圳）有限公司 | Method, apparatus and device for recognizing abnormal behavior on the basis of voice and image features |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2606256A (en) * | 2021-03-12 | 2022-11-02 | Frontier Cool Inc | Fabric informaiton digitization system and method thereof |
| GB2606256B (en) * | 2021-03-12 | 2023-04-26 | Frontier Cool Inc | Fabric information digitization system and method thereof |
| CN116740581A (en) * | 2023-08-16 | 2023-09-12 | 深圳市欢创科技有限公司 | Method for determining material identification model, method for returning to base station and electronic equipment |
| CN116740581B (en) * | 2023-08-16 | 2023-10-27 | 深圳市欢创科技有限公司 | Method for determining material identification model, method for returning to base station and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113743525B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111126325B (en) | Intelligent personnel security identification statistical method based on video | |
| CN105682310B (en) | Combined lighting device and method based on image quality control | |
| US8700498B2 (en) | Feature analyzing apparatus for a surface of an object | |
| CN113743525B (en) | Fabric material identification system and method based on luminosity three-dimensional | |
| CN106909882A (en) | A kind of face identification system and method for being applied to security robot | |
| CN111340798A (en) | Application of deep learning in product appearance flaw detection | |
| Hapsari et al. | Face detection using haar cascade in difference illumination | |
| CN109284738A (en) | Irregular face antidote and system | |
| US20190188451A1 (en) | Lightweight 3D Vision Camera with Intelligent Segmentation Engine for Machine Vision and Auto Identification | |
| Tancik et al. | Flash photography for data-driven hidden scene recovery | |
| CN113128335B (en) | Method, system and application for detecting, classifying and finding micro-living ancient fossil image | |
| KR102503116B1 (en) | Automatic Data Set Generation Device and their Control Method for AI Machine Learning Learning | |
| CN115170792B (en) | Infrared image processing method, device and equipment and storage medium | |
| CN112949452B (en) | Robot low-light environment grabbing detection method based on multitask shared network | |
| CN112070711A (en) | Analysis method of micro-droplets in micro-droplet image detection method | |
| CN107609475B (en) | Pedestrian detection false detection extraction method based on light field camera | |
| CN109597337A (en) | A kind of machine vision intelligent acquisition and control system | |
| KR102704880B1 (en) | System for inspecting product defects by type based on a deep learning model | |
| CN117197789A (en) | Curtain wall frame identification method and system based on multi-scale boundary feature fusion | |
| CN115035082A (en) | YOLOv4 improved algorithm-based aircraft transparency defect detection method | |
| CN114724002A (en) | Method for rapidly detecting contraband | |
| CN110866426A (en) | Pedestrian identification method based on light field camera and deep learning | |
| CN112926702A (en) | Active light source type object material identification system and method | |
| Anwar et al. | Fog classification and accuracy measurement using SVM | |
| CN115690577A (en) | Unmanned aerial vehicle for animal and plant identification and working method thereof |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |