CN208752022U - Deep learning machine vision automatic on-line cloth examination device based on raspberry pie - Google Patents
Deep learning machine vision automatic on-line cloth examination device based on raspberry pie Download PDFInfo
- Publication number
- CN208752022U CN208752022U CN201820914131.4U CN201820914131U CN208752022U CN 208752022 U CN208752022 U CN 208752022U CN 201820914131 U CN201820914131 U CN 201820914131U CN 208752022 U CN208752022 U CN 208752022U
- Authority
- CN
- China
- Prior art keywords
- raspberry pie
- cloth
- light source
- deep learning
- machine vision
- 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.)
- Expired - Fee Related
Links
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
Originally practical newly to disclose a kind of deep learning machine vision automatic on-line cloth examination device based on raspberry pie, including machinery frame, the upper horizontal of machinery frame is evenly arranged with several light sources, light source is located at right above cloth to be detected, horizontal homogeneous is gone back above machinery frame is provided with several plane cameras, plane camera is corresponded with light source position and quantity is consistent, plane camera and light source are uniformly distributed relative to the wide cut of cloth to be detected, each plane camera is respectively connected with an image capture controller, all image capture controllers are connect with interchanger, interchanger connect composition data communication network with primary processor raspberry pie again, primary processor raspberry pie is also connected with internet cloud server and display, cloth to be detected is wound on the winder after examining, the utility model solves perching low output existing in the prior art, effect Rate difference and problem with high costs.
Description
Technical field
The utility model belongs to textile industry fabric defects detection equipment technical field, and in particular to one kind is based on raspberry pie
Deep learning machine vision automatic on-line cloth examination device.
Background technique
Cloth inspecting machine, which is that apparel industry production is preceding, carries out the especially big breadth such as cotton, hair, fiber crops, silk, chemical fibre, double width and single width cloth
The special equipment of a set of indispensability of detection, by cloth inspecting machine to cloth to be detected progress assessment of conformity or unqualified, simultaneously
According to the grade for the parameter evaluatings cloth such as the type of fault and number occur.Due to cloth inspecting machine weaving process in play it is important
Role, so the accuracy rate and cost of cloth inspecting machine are most important.It is accomplished manually currently, defect detection mainly passes through, tradition
Artificial perching speed it is unhappy, low output, inefficient.Artificial perching has certain technical requirements and eyesight to want inspection man simultaneously
It asks.Long-term vision is concentrated, and asthenopia is easy to cause, and suffers from occupational disease, and labor intensity is high.Inspection man can only generally detect per hour
To 200 faults or so, it is more than this range, is easy to produce the omission of examination and erroneous judgement.The skills such as computer technology and machine learning in recent years
The rapid development of art, many enterprises throw oneself into the research of automatic on-line perching, due to the limitation of its algorithm, for perching to picture
The inspection of main practical monochrome dyeing fabric.Its high price simultaneously, and the burdensome cost of many enterprises.
Utility model content
The purpose of the utility model is to provide a kind of, and the deep learning machine vision automatic on-line perching based on raspberry pie fills
It sets, solves the problems, such as perching low output existing in the prior art, inefficient and with high costs.
The utility model is the technical scheme adopted is that the deep learning machine vision automatic on-line perching based on raspberry pie
Device, including machinery frame, the upper horizontal of machinery frame are evenly arranged with several light sources, and light source is located at right above cloth to be detected,
Horizontal homogeneous is gone back above machinery frame is provided with several plane cameras, plane camera and light source position one-to-one correspondence and quantity one
It causes, plane camera and light source are uniformly distributed relative to the wide cut of cloth to be detected, and each plane camera is respectively connected with an image
Acquisition controller, all image capture controllers are connect with interchanger, and interchanger connect composition with primary processor raspberry pie again
Data communication network, primary processor raspberry pie are also connected with internet cloud server and display, and cloth to be detected is by examining
After wind on the winder.
The utility model is also characterized by
Plane camera is located at the surface of the horizontal median axis of several light sources.
Light source is annular LED light source.
Plane camera is the area array cameras of COMS+BASLER industrial camera lens.
Image capture controller and primary processor raspberry pie are raspberry pie three generations's Type B.
The utility model has the beneficial effects that the deep learning machine vision automatic on-line cloth examination device based on raspberry pie,
Image is concurrently acquired simultaneously by using a number of camera, and deep learning convolutional neural networks detection algorithm can online certainly
Dynamic detection fault, and save, assess cloth, maximum detection fabric width 330cm, speed of service 400m/min are detected minimum
Fault 0.2mm.The data result of detection is stored into Cloud Server simultaneously, parking of alarming immediately when detecting fault, and right
The classification of fault result, auxiliary operation worker, in the adjustment of process, avoid the waste of cloth by the gross for after, promote the matter of cloth production
Amount.
Detailed description of the invention
Fig. 1 is the structural representation of deep learning machine vision automatic on-line cloth examination device of the utility model based on raspberry pie
Figure.
In figure, 1. primary processor raspberry pies, 2. internet cloud servers, 3. displays, 4. image capture controllers, 5. are handed over
It changes planes, 6. area array cameras, 7. light sources, 8. detected cloth, 9. up- coilers, 10. machinery frames.
Specific embodiment
The utility model is described in detail with reference to the accompanying drawings and detailed description.
Deep learning machine vision automatic on-line cloth examination device of the utility model based on raspberry pie, structure as shown in Figure 1,
Including machinery frame 10, the upper horizontal of machinery frame 10 is evenly arranged with several light sources 7, light source 7 be located at cloth 8 to be detected just on
Side, the top of machinery frame 10 go back horizontal homogeneous and are provided with several plane cameras 6, plane camera 6 and 7 position of light source correspond and
Quantity is consistent, and wide cut of the plane camera 6 with light source 7 relative to cloth 8 to be detected is uniformly distributed, and each plane camera 6 is all connected with
There is an image capture controller 4, all image capture controllers 4 are connect with interchanger 5, interchanger 5 and and primary processor
Raspberry pie 1 connects composition data communication network, and primary processor raspberry pie 1 is also connected with internet cloud server 2 and display 3,
Cloth 8 to be detected is wrapped on up- coiler 9 after examining.
Plane camera 6 is located at the surface of the horizontal median axis of several light sources 7.
Light source 7 is annular LED light source.
Plane camera 6 is the area array cameras of COMS+BASLER industrial camera lens.
Image capture controller 4 and primary processor raspberry pie 1 are raspberry pie three generations's Type B.
Plane camera 6, light source 7 and 4 quantity of image capture controller are 4~6.
Deep learning machine vision automatic on-line cloth examination device of the utility model based on raspberry pie, in artificial intelligence and machine
Under the overall background of the technology development of device vision, deep learning and machine vision are relied on, guarantees the accuracy rate of fabric defects detection,
Automatic on-line perching is realized using the hardware platform of raspberry pie, the specific working principle is as follows:
The underface for the light source 7 for allowing detected cloth 8 to be placed through on machinery frame 10, light source 7 are fixed on machinery frame 10
Face, and with detected 8 keeping parallelism of cloth；Plane camera 6 is fixed in the horizontal median axis of light source 7, plane camera 6 is logical
It crosses CSI bus interface and connect the unit for completing most basic acquisition picture with image capture controller 4, in an identical manner basis
The uniform transversely arranged consistent image capture controller 4 of installation number of the breadth of cloth, so that tested cloth is completely covered；It will
Image capture controller 4 is connected with interchanger 5, and accesses primary processor raspberry pie 1, composition data communication network, completes data
Be exchanged with each other；The digital picture for collecting cloth whole picture reaches primary processor raspberry pie 1 by data communication network, in conjunction with
Primary processor raspberry pie 1, realization handle collected cloth digital picture；Automatic on-line fabric inspecting system is to pass through depth
Learning algorithm, off-line learning training obtain model and realize online perching then by model transplantations into system；Perching it is final
As a result device end is shown in by display 3；Primary processor raspberry pie 1 by perching data result and collects at the same time
Fault data sample be stored in Cloud Server via internet cloud server 2 and BSBC encryption technology, use SSH agreement real
The remote debugging of existing automatic on-line cloth examination device, is conducive to the subsequent update of equipment, provides data for intelligent terminal.Pass through reality
Investigation cloth, which generally requires, detects the fault on its surface, and conventional cloth wide cut is larger, so using 4 tunnel Image Acquisition
System；Image procossing in automatic Cloth Inspecting System uses deep learning convolutional neural networks, has for computing resource biggish
Demand, therefore the collecting work using 4 raspberry pies as the completion image of image capture controller 4, and apply a raspberry pie
As primary server --- primary processor raspberry pie 1 passes through 5 networking of interchanger, the distribution of the automatic on-line fabric inspecting system of realization
Formula hardware platform；Primary processor raspberry pie 1, up- coiler 9, internet cloud server 2, plane camera 6 and light source 7 etc. are mutually close
Cooperation provides for the efficient, stable of automatic on-line fabric inspecting system and accurately best guarantee, people is not only greatly saved in this way
Work detection is time-consuming, and the status for also sufficiently avoiding the current high enterprise of automatic on-line perching equipment price that from can not undertaking not only detects
As a result objective stabilization, and testing cost has been saved, improve detection efficiency.
The utility model realizes the on-line checking of fault using deep learning, in the premise of effective data set off-line training
Condition realizes the automatic on-line detection of fabric defects, and the accuracy rate of testing result is protected at the same time；Meeting system
It operates normally, significantly reduces the cost of automatic on-line perching equipment, realize pushing away on a large scale for automatic on-line perching equipment
Extensively, it is of great significance for the development of textile enterprise；The utility model uses the polishing mode of LED annular light source, effectively
The area array cameras avoided is during acquisition, the non-uniform problem of light, while solving the meeting under special fault type
The practical problem for generating projection, obtains the image information of high quality, is the key condition of automatic on-line cloth examination device；This is practical new
Type uses the OV5647 camera module group of size design mechanical structure and 500W pixel by the camera lens of BASLER industrial camera
The CMOS sensitive chip of mould group is located on the central axes of camera lens simultaneously encapsulation process by dress synthesis, so that minimum fault reaches
0.2mm；The utility model is transversely arranged using several table top array cameras, covers complete cloth cover, and realize 330cm whole picture cloth cover
Online perching；The utility model is concurrently executed using several processes and the thinking of the processing of distributed computing, more efficient benefit
With computing resource, different calculating tasks is distributed in the controller of Image-capturing platform respectively, realizes automatic cloth inspection device
The speed of service of 400m/min；Automatic Cloth Inspecting System in the utility model realizes cloud storage, and detection device will can acquire in real time
Fault sample and perching data be uploaded to cloud space, the monitoring center in factory can be directly accessed cloud communication platform monitoring
Equipment operation condition.Secondly according to the data set collected in cloud platform, producer can improve and rise according to their own needs
Grade equipment, is more advantageous to engineer and efficiently debugs and solve plant issue.Finally, to avoid letting out for cloud platform monitoring information
Leakage, detection data use BSBC encryption technology during transmission, reduce the dependency degree to key, and are promoted conscientiously certainly
The safety of dynamic perching data.
Claims (5)
1. the deep learning machine vision automatic on-line cloth examination device based on raspberry pie, which is characterized in that including machinery frame (10)
The upper horizontal of machinery frame (10) is evenly arranged with several light sources (7), and light source (7) is located at right above cloth to be detected (8), mechanical
Horizontal homogeneous is gone back above frame (10) and is provided with several plane cameras (6), and plane camera (6) and light source (7) position correspond
And quantity is consistent, wide cut of the plane camera (6) with light source (7) relative to cloth to be detected (8) is uniformly distributed, each plane camera
(6) it is respectively connected with an image capture controller (4), all image capture controllers (4) connect with interchanger (5), exchange
Machine (5) connect composition data communication network with primary processor raspberry pie (1) again, and primary processor raspberry pie (1) is also connected with interconnection
Net Cloud Server (2) and display (3), cloth (8) to be detected are wrapped on up- coiler (9) after examining.
2. the deep learning machine vision automatic on-line cloth examination device according to claim 1 based on raspberry pie, feature
It is, the plane camera (6) is located at the surface of the horizontal median axis of several light sources (7).
3. the deep learning machine vision automatic on-line cloth examination device according to claim 1 based on raspberry pie, feature
It is, the light source (7) is annular LED light source.
4. the deep learning machine vision automatic on-line cloth examination device according to claim 1 based on raspberry pie, feature
It is, the plane camera (6) is the area array cameras of COMS+BASLER industrial camera lens.
5. the deep learning machine vision automatic on-line cloth examination device according to claim 1 based on raspberry pie, feature
It is, described image acquisition controller (4) and primary processor raspberry pie (1) are raspberry pie three generations's Type B.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201820914131.4U CN208752022U (en) | 2018-06-13 | 2018-06-13 | Deep learning machine vision automatic on-line cloth examination device based on raspberry pie |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201820914131.4U CN208752022U (en) | 2018-06-13 | 2018-06-13 | Deep learning machine vision automatic on-line cloth examination device based on raspberry pie |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN208752022U true CN208752022U (en) | 2019-04-16 |
Family
ID=66052962
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201820914131.4U Expired - Fee Related CN208752022U (en) | 2018-06-13 | 2018-06-13 | Deep learning machine vision automatic on-line cloth examination device based on raspberry pie |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN208752022U (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340802A (en) * | 2020-03-25 | 2020-06-26 | 嘉兴学院 | Intelligent cloth inspecting machine capable of automatically inspecting cloth by adopting artificial intelligence technology |
| CN111650208A (en) * | 2020-06-01 | 2020-09-11 | 东华大学 | Tour type woven fabric defect on-line detector |
-
2018
- 2018-06-13 CN CN201820914131.4U patent/CN208752022U/en not_active Expired - Fee Related
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340802A (en) * | 2020-03-25 | 2020-06-26 | 嘉兴学院 | Intelligent cloth inspecting machine capable of automatically inspecting cloth by adopting artificial intelligence technology |
| CN111650208A (en) * | 2020-06-01 | 2020-09-11 | 东华大学 | Tour type woven fabric defect on-line detector |
| CN111650208B (en) * | 2020-06-01 | 2021-08-27 | 东华大学 | Tour type woven fabric defect on-line detector |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102967606B (en) | Textile machine fabric defect visual inspection system | |
| CN208752022U (en) | Deep learning machine vision automatic on-line cloth examination device based on raspberry pie | |
| CN202770781U (en) | Fabric defect online detection device based on machine vision | |
| CN111948228A (en) | Industrial production line overall process visual detection system based on 5G communication cloud test | |
| CN112258491A (en) | Three-dimensional simulation method for defects of power equipment | |
| CN105044126A (en) | Visual detection system for large-width continuous surface defects | |
| CN106841237A (en) | A kind of electronic display glass cover plate surfaces defect detecting system and method | |
| CN110796754A (en) | Machine room inspection method based on image processing technology | |
| CN205175925U (en) | Fabric defects real -time detection device | |
| CN102601131A (en) | Steel billet surface quality online detection device | |
| CN110567987A (en) | Feeder terminal substrate detection system and method based on automatic vision technology | |
| CN103245670A (en) | Defect detection device and defect detection method of optical element | |
| CN113988573A (en) | Risk judgment method, system and medium for routing inspection unmanned aerial vehicle based on power system | |
| CN107309178A (en) | Cell phone intelligent automated test device | |
| CN112964732A (en) | Spinning cake defect visual detection system and method based on deep learning | |
| CN203196910U (en) | Automatic detection device for surface printing quality of product packaging box | |
| WO2024037240A1 (en) | Intelligent cloth inspection detection system and method, and computer-readable storage medium and product | |
| CN117114254B (en) | Power grid new energy abnormal data monitoring method and system | |
| CN112906602A (en) | Automatic identification device and identification method for electricity meter of power distribution cabinet based on image processing | |
| CN204125778U (en) | A kind of two-sided cloth inspecting machine | |
| CN103411688B (en) | The Methods of Remote Monitoring of Michelson interference ring and TT&C system thereof | |
| CN202416024U (en) | Automatic fabric inspection machine using hybrid light source | |
| Lin et al. | Enhancing the quality inspection process in the food manufacturing industry through automation | |
| CN117082206A (en) | Wearable intelligent inspection device of photovoltaic power station based on MR+4G technology | |
| CN205910332U (en) | Automatic verification system of acquisition terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190416 Termination date: 20200613 |