CN107632251A - PCB single board fault detection method - Google Patents
PCB single board fault detection method Download PDFInfo
- Publication number
- CN107632251A CN107632251A CN201710671754.3A CN201710671754A CN107632251A CN 107632251 A CN107632251 A CN 107632251A CN 201710671754 A CN201710671754 A CN 201710671754A CN 107632251 A CN107632251 A CN 107632251A
- Authority
- CN
- China
- Prior art keywords
- single board
- pcb single
- pcb
- failure
- convolutional neural
- 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
Landscapes
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
- Radiation Pyrometers (AREA)
Abstract
The invention provides a kind of PCB single board fault detection method, comprise the following steps：Step 1, PCB single board is fixed on detection means, and be sent to the station to be detected of device；Step 2, the optical sensor in detection means detect PCB single board, and shock module is shaken with certain frequency, induced malfunction；Step 3, PCB single board is in powered operation state, reuse thermal infrared imager and the temperature on PCB single board surface is acquired, generate the Infrared Thermogram of detected sample, and be transmitted to computer；Step 4, computer are pre-processed after obtaining Infrared Thermogram, and are inputted in the convolutional neural networks that have trained and calculated, according to output result judge PCB single board whether failure.The present invention improves PCB single board fault detection efficiency and Detection accuracy, reduces the labor intensity of operating personnel, easy to operate, reliability is high.
Description
Technical field
The present invention relates to a kind of detection method, in particular it relates to a kind of PCB single board fault detection method.
Background technology
PCB (Printed Circuit Board, printed circuit board (PCB)) is the supplier of electronic component electrical connection.PCB
Veneer is a kind of most basic printed circuit board (PCB), and its part is concentrated wherein simultaneously, and wire is then concentrated on another side.In life
During production, the bad error of paster, discrete component or solder joint occurs, it is necessary to carry out fault detect to it in PCB single board.At present
Still the method that artificial detection is used in most of production lines, inspector are carried out one by one by its experience by visual inspection to pcb board
Detection, this method accuracy of detection are influenceed by inspector experience and its degree of fatigue, consume manpower, and detection speed is slow,
The needs of quick detection can not be met, it is more low to result in operating efficiency.From detection speed, detection efficiency and cost control three
From the aspect of individual, the requirement of the production of enterprise's large scale integration is not complyed with.
The content of the invention
For in the prior art the defects of, it is an object of the invention to provide a kind of PCB single board fault detection method, it is improved
PCB single board fault detection efficiency and Detection accuracy, reduce the labor intensity of operating personnel, easy to operate, reliability is high.
According to an aspect of the present invention, there is provided a kind of PCB single board fault detection method, it is characterised in that including following
Step：
Step 1, PCB single board is fixed on detection means, and be sent to the station to be detected of device；
Step 2, the optical sensor in detection means detect PCB single board, and shock module is shaken with certain frequency, is lured
Send out failure；
Step 3, PCB single board is in powered operation state, reuse temperature of the thermal infrared imager to PCB single board surface
It is acquired, generates the Infrared Thermogram of detected sample, and is transmitted to computer；
Step 4, computer are pre-processed after obtaining Infrared Thermogram, and input the convolutional Neural net trained
Calculated in network, according to output result judge PCB single board whether failure.
Preferably, PCB single board to be measured is fixed on objective table by the step 1 by fixing device, then will by conveyer
It is sent to position to be detected.
Preferably, after the step 2 is sent to position to be detected, optical sensor detects PCB single board to be measured, Xiang Zhen
Dynamic model block sends start-up operation instruction, PCB single board to be measured is shaken with certain frequency, induces PCB single board to be measured in advance in reality
Failure caused by the problem of potential in work.
Preferably, after the completion of step 3 vibrations, PCB single board to be measured is made to be in powered operation state, and to infrared heat
Make its start-up operation as instrument sends instruction；Preparation before use is correctly to be connected thermal infrared imager with computer, is allowed to
It is in running order, thermal imager system software in computer is run, adjusts the relevant parameter of thermal imaging system；After being connected to work order,
Thermal infrared imager starts to gather the surface temperature of PCB single board to be measured, obtains Infrared Thermogram and preserves, and transmits it to computer
In.
Preferably, after the step 4 computer receives Infrared Thermogram, it is sent into intelligent diagnostics module and is detected；
In diagnostic phases, first have to carry out pretreatment operation, including gray scale conversion and adjustment size, it is then that pretreated PCB is mono-
The convolutional neural networks that the input of plate thermography trains are calculated, and sample is classified using convolutional neural networks, are classified
As a result it is failure and the not class of failure two, and then screens and be out of order PCB single board and obtain fault rate；It is first complete before diagnostic phases
Into the training of convolutional neural networks；In the training stage, first have to be trained the acquisition of collection, a number of PCB of handmarking
Veneer thermography, labeled as failure and the not class of failure two；Then pretreated sample graph is inputted into convolutional neural networks, carried out
Training；The connection weight and bias of network are constantly updated by error propagated forward and back-propagation algorithm, until
Network effect reaches expected.
Preferably, the purpose that PCB single board is fixed is to be worked in follow-up shock module using fixing device by the step 1
When make PCB single board position keep it is constant；Being to detect on station to be measured using the purpose of optical sensor in the step 2 is
It is no to transmit PCB single board to be measured, so as to send instruction to subsequent module；The effect of shock module is to lure in advance in the step 3
Failure caused by sending out the problem of PCB single board is potential in real work.
Preferably, the purpose that the step 4 makes PCB single board be in powered operation state is：Infrared detection module, fingerprint identification module uses
Passive infrared detection method, Temperature Distribution caused by itself obtains Infrared Thermogram when collection PCB single board works, and this needs PCB
Veneer is in running order lower could to produce heat；PCB single board may expose in the operating condition not to be had under off working state
Some failures, screened out by follow-up detection, ensure the reliability of PCB single board work.
Preferably, the surface temperature of the collection PCB single board to be measured in the step 3 comprises the following steps：By infrared thermal imagery
Instrument is normally connected with computer, is allowed in running order；The software systems of thermal infrared imager in computer are run, adjust its phase
Related parameter；Thermal infrared imager starts to gather the Temperature Distribution on PCB single board surface after being connected to work order, obtains Infrared Thermogram；
Preserve, and transmit medium pending to computer after the completion of thermography collection.
Preferably, convolutional neural networks include training stage and diagnostic phases in the step 4；In the methods of the invention,
The training stage of convolutional neural networks comprises the following steps：The acquisition of training set, a number of PCB single board thermal imagery of handmarking
Figure, labeled as failure and the not class of failure two；PCB single board thermography sample preprocessing, including gray scale conversion and size adjusting；Will be pre-
Sample graph input convolutional neural networks after processing, carry out Training；Pass through error propagated forward and back-propagation algorithm
The weights and bias of network are constantly updated, until training effect reaches expected；The diagnostic phases of convolutional neural networks include with
Lower step：PCB single board thermography sample preprocessing, including gray scale conversion and size adjusting；By pretreated PCB single board thermal imagery
The convolutional neural networks that figure input trains are calculated, and sample are classified using convolutional neural networks, classification results are
Failure and the not class of failure two, and then screen and be out of order PCB single board and obtain fault rate.
Compared with prior art, the present invention has following beneficial effect：
One, PCB single board fault detection method of the present invention, detected in the state of PCB single board powered operation.
In this case, PCB single board can active emitting infrared radiation, whereby it can be detected that it is that may be present in the operating condition
Failure, expand the scope of detection failure.
Two, PCB single board fault detection method of the present invention is asked potential rosin joint of PCB single board etc. using shock module
Topic is induced in advance, eliminates the part incipient fault of PCB single board, improves the PCB single board detected by this method rear
Reliability during continuous use.
Three, PCB single board fault detection method of the present invention is entered using artificial intelligence diagnosis' method to Infrared Thermogram
Row processing, using grader of the convolutional neural networks as failure, with classification speed is fast, accuracy of detection is high, simple to operate
Feature.
Four, the sample delivery of PCB single board fault detection method of the present invention, sample vibrations, infrared collecting, intelligence point
Analysis is automatically performed by computer or detection means, and whole flow process realizes automation, significantly improves PCB single board event
Hinder the efficiency and accuracy of detection of detection, be highly suitable for industrial production etc. for the high system of requirement of real-time.
Brief description of the drawings
The detailed description made by reading with reference to the following drawings to non-limiting example, further feature of the invention,
Objects and advantages will become more apparent upon：
Fig. 1 is the FB(flow block) of detection method；
Fig. 2 is the schematic diagram of detection means of the present invention；
Fig. 3 is the FB(flow block) that infrared module works in detection method；
Fig. 4 is the FB(flow block) that intelligent diagnostics module works in detection method；
Fig. 5 is the operating diagram of convolutional neural networks training stage in detection method.
Embodiment
With reference to specific embodiment, the present invention is described in detail.Following examples will be helpful to the technology of this area
Personnel further understand the present invention, but the invention is not limited in any way.It should be pointed out that the ordinary skill to this area
For personnel, without departing from the inventive concept of the premise, various modifications and improvements can be made.These belong to the present invention
Protection domain.
As shown in figure 1, PCB single board fault detection method of the present invention comprises the following steps：
Step S1, PCB single board is fixed on detection means, and be sent to the station to be detected of device.
Step S2, the optical sensor in detection means detect PCB single board, and shock module is shaken with certain frequency, is lured
Send out failure.
Step S3, PCB single board is in powered operation state, reuse temperature of the thermal infrared imager to PCB single board surface
It is acquired, generates the Infrared Thermogram of detected sample, and is transmitted to computer.
Step S4, computer are pre-processed after obtaining Infrared Thermogram, and input the convolutional Neural net trained
Calculated in network, according to output result judge PCB single board whether failure.
PCB single board to be measured is fixed on objective table by the step 1 by fixing device, then is transmitted by conveyer
To position to be detected.
After the step 2 is sent to position to be detected, optical sensor detects PCB single board to be measured, is sent out to shock module
Go out instruction of starting working, PCB single board to be measured is shaken with certain frequency, induce PCB single board to be measured in advance and dived in real work
The problem of caused by failure.
After the completion of the step 3 vibrations, PCB single board to be measured is set to be in powered operation state, and sent to thermal infrared imager
Instruction makes its start-up operation；Preparation before use is correctly to be connected thermal infrared imager with computer, is allowed to be in work
State, thermal imager system software in computer is run, adjust the relevant parameter of thermal imaging system；After being connected to work order, infrared thermal imagery
Instrument starts to gather the surface temperature of PCB single board to be measured, obtains Infrared Thermogram and preserves, and transmits it in computer.
After the step 4 computer receives Infrared Thermogram, it is sent into intelligent diagnostics module and is detected；In diagnosis rank
Duan Zhong, first have to carry out pretreatment operation, including gray scale conversion and adjustment size, then by pretreated PCB single board thermal imagery
The convolutional neural networks that figure input trains are calculated, and sample are classified using convolutional neural networks, classification results are
Failure and the not class of failure two, and then screen and be out of order PCB single board and obtain fault rate；Before diagnostic phases, convolution is first completed
The training of neutral net；In the training stage, first have to be trained the acquisition of collection, a number of PCB single board heat of handmarking
As figure, labeled as failure and the not class of failure two；Then pretreated sample graph is inputted into convolutional neural networks, carries out supervision
Training；The connection weight and bias of network are constantly updated by error propagated forward and back-propagation algorithm, until network is imitated
Fruit reaches expected.
The purpose that the step 4 makes PCB single board be in powered operation state is：Infrared detection module, fingerprint identification module uses passive red
Outer detection method, Temperature Distribution caused by itself obtains Infrared Thermogram when collection PCB single board works, and this is needed at PCB single board
Heat could be produced under working condition；PCB single board may expose the event not having under off working state in the operating condition
Barrier, is screened out by follow-up detection, it is ensured that the reliability of PCB single board work.
The purpose that PCB single board is fixed is to make PCB when follow-up shock module works using fixing device by the step 1
Veneer position keeps constant；It is to detect whether to transmit on station to be measured using the purpose of optical sensor in the step 2 and treats
PCB single board is surveyed, so as to send instruction to subsequent module；The effect of shock module is to induce PCB in advance mono- in the step 3
Failure caused by the problem of plate is potential in real work, for example refer to only have at solder joint the problem of PCB single board rosin joint or dry joint
A small amount of soldering is lived, and causes loose contact, switch-on and -off under working condition.In the methods of the invention, by shaking PCB single board
The outburst of such failure is induced in advance, can be screened out in follow-up detection module.
As shown in Fig. 2 the schematic diagram for detection means in detection method.1 is computer in figure, and 2 be that optics passes
Sensor, 3 be thermal infrared imager, and 4 be PCB single board to be measured, and 5 be fixing device, and 6 be conveyer, and 7 be objective table, and 8 be vibrations mould
Block.In concrete operations, the preparation of detection can be completed by carrying out hardware connection according to above-mentioned schematic diagram.Shock module will be treated
Survey PCB single board and put objective table and fixation, be sent to position to be detected, it is shaken with certain frequency, induced by shaking
The appearance in advance of the failures such as rosin joint.Infrared detection module, fingerprint identification module makes PCB single board be in powered operation state, uses thermal infrared imager pair
The surface temperature of PCB single board is acquired, and obtains Infrared Thermogram.Intelligent diagnostics module is based on convolutional neural networks to infrared heat
As figure is detected, including following two stages：First stage is the training stage, and training sample is inputted into convolutional neural networks,
Training sample is the pretreated PCB single board thermography with label, constantly updated by training network connection weight and
Bias, until network meets to require；Second stage is diagnostic phases, the convolution that the input of pretreated sample graph is trained
Neutral net is calculated, and sample is classified using convolutional neural networks, and classification results are failure and the not class of failure two, are entered
And screen and be out of order PCB single board and obtain fault rate.The inventive method is realized mono- to PCB by automation, intelligentized means
The detection of plate failure, easy to operate, detection efficiency is high, and reliability is high, versatile.
Illustrate the detection process of the PCB single board fault detection method of the present invention with reference to example.
(1) PCB single board 4 to be measured is fixed on objective table 7 by fixing device 5 first, then transmitted by conveyer 6
To position to be detected.
(2) optical sensor 2 detects PCB single board 4 to be measured, sends start-up operation instruction to shock module 8, makes to be measured
PCB single board 4 is shaken with certain frequency, induces PCB single board 4 to be measured event caused by potential problems possibility in real work in advance
Barrier, such as the problems such as rosin joint.
(3) after the completion of shaking, PCB single board 4 to be measured is made to be in powered operation state, and instruction is sent to thermal infrared imager 3
Make its start-up operation, its workflow block diagram is as shown in Figure 3.Preparation before use is by thermal infrared imager 3 and computer 1
Correct connection, is allowed in running order, runs thermal imager system software in computer 1, adjusts the relevant parameter of thermal imaging system.Connect
To after work order, thermal infrared imager 3 starts to gather the surface temperature of PCB single board 4 to be measured, obtains Infrared Thermogram and preserves, and
Transmit it in computer 1.
(4) after computer 1 receives Infrared Thermogram, it is sent into intelligent diagnostics module and is detected, intelligent diagnostics module work
It is as shown in Figure 4 to make FB(flow block).In diagnostic phases, first have to carry out pretreatment operation, including gray scale conversion and adjustment size,
Then the convolutional neural networks that the input of pretreated PCB single board thermography trains are calculated, utilizes convolutional Neural net
Network is classified to sample, and classification results are failure and the not class of failure two, and then is screened and be out of order PCB single board and obtain failure
Rate.Before diagnostic phases, the training of convolutional neural networks is first completed.In the training stage, first have to be trained the acquisition of collection,
The a number of PCB single board thermography of handmarking, labeled as failure and the not class of failure two.Then by pretreated sample graph
Convolutional neural networks are inputted, carry out Training, the operating diagram of training stage is as shown in Figure 5.By before error to biography
The connection weight and bias that network is constantly updated with back-propagation algorithm are broadcast, until training effect reaches expected.Province in Fig. 5
Contracted notation square frame represents to set varying number and the convolutional layer and sample level of scale according to actual acquisition environment.
The specific embodiment of the present invention is described above.It is to be appreciated that the invention is not limited in above-mentioned
Particular implementation, those skilled in the art can make various deformations or amendments within the scope of the claims, this not shadow
Ring the substantive content of the present invention.
Claims (9)
1. a kind of PCB single board fault detection method, it is characterised in that comprise the following steps：
Step 1, PCB single board is fixed on detection means, and be sent to the station to be detected of device；
Step 2, the optical sensor in detection means detect PCB single board, and shock module is shaken with certain frequency, induce therefore
Barrier；
Step 3, PCB single board is in powered operation state, reuse thermal infrared imager and the temperature on PCB single board surface is carried out
Collection, generates the Infrared Thermogram of detected sample, and be transmitted to computer；
Step 4, computer are pre-processed after obtaining Infrared Thermogram, and are inputted in the convolutional neural networks trained
Calculate, according to output result judge PCB single board whether failure.
2. PCB single board fault detection method according to claim 1, it is characterised in that the step 1 is by fixing device
PCB single board to be measured is fixed on objective table, then position to be detected is sent to by conveyer.
3. PCB single board fault detection method according to claim 1, it is characterised in that the step 2 is sent to be checked
Location postpones, and optical sensor detects PCB single board to be measured, sends start-up operation instruction to shock module, makes PCB single board to be measured
Shaken with certain frequency, induce failure caused by the problem of PCB single board to be measured is potential in real work in advance.
4. PCB single board fault detection method according to claim 1, it is characterised in that after the completion of the step 3 vibrations,
PCB single board to be measured is set to be in powered operation state, and send instruction to thermal infrared imager to make its start-up operation；Preparation before use
Work is correctly to be connected thermal infrared imager with computer, is allowed in running order, and it is soft to run thermal imager system in computer
Part, adjust the relevant parameter of thermal imaging system；After being connected to work order, thermal infrared imager starts to gather the surface temperature of PCB single board to be measured
Degree, obtain Infrared Thermogram and preserve, and transmit it in computer.
5. PCB single board fault detection method according to claim 1, it is characterised in that the step 4 computer receives
To after Infrared Thermogram, it is sent into intelligent diagnostics module and is detected；In diagnostic phases, first have to carry out pretreatment operation, bag
Gray scale conversion and adjustment size are included, then enters the convolutional neural networks that the input of pretreated PCB single board thermography trains
Row is calculated, and sample is classified using convolutional neural networks, and classification results are failure and the not class of failure two, and then filter out event
Barrier PCB single board simultaneously obtains fault rate；Before diagnostic phases, the training of convolutional neural networks is first completed；In the training stage, first
It is trained the acquisition of collection, a number of PCB single board thermography of handmarking, labeled as failure and the not class of failure two；So
Pretreated sample graph is inputted into convolutional neural networks afterwards, carries out Training；By error propagated forward with reversely passing
Connection weight and bias that algorithm constantly updates network are broadcast, until network effect reaches expected.
6. PCB single board fault detection method according to claim 1, it is characterised in that the step 1 is filled using fixed
It is PCB single board position is kept constant when follow-up shock module works to put the purpose that PCB single board is fixed；In the step 2
It is to detect PCB single board to be measured whether is transmitted on station to be measured using the purpose of optical sensor, so as to be sent to subsequent module
Instruction；The effect of shock module causes the problem of induction PCB single board is potential in real work in advance in the step 3
Failure.
7. PCB single board fault detection method according to claim 1, it is characterised in that the step 4 makes at PCB single board
It is in the purpose of powered operation state：Infrared detection module, fingerprint identification module uses passive infrared detection method, when collection PCB single board works certainly
Temperature Distribution obtains Infrared Thermogram caused by body, and this needs PCB single board is in running order lower could produce heat；PCB is mono-
Plate may expose the failure not having under off working state in the operating condition, be screened out by follow-up detection, ensure that PCB is mono-
The reliability of plate work.
8. PCB single board fault detection method according to claim 1, it is characterised in that the collection in the step 3 is treated
The surface temperature for surveying PCB single board comprises the following steps：Thermal infrared imager is normally connected with computer, is allowed to be in work shape
State；The software systems of thermal infrared imager in computer are run, adjust its relevant parameter；Thermal infrared imager is opened after being connected to work order
The Temperature Distribution on beginning collection PCB single board surface, obtains Infrared Thermogram；Preserve, and transmitted to calculating after the completion of thermography collection
Machine is medium pending.
9. PCB single board fault detection method according to claim 1, it is characterised in that convolutional Neural in the step 4
Network includes training stage and diagnostic phases；In the methods of the invention, the training stage of convolutional neural networks comprises the following steps：
The acquisition of training set, a number of PCB single board thermography of handmarking, labeled as failure and the not class of failure two；PCB single board heat
As figure sample preprocessing, including gray scale conversion and size adjusting；Pretreated sample graph is inputted into convolutional neural networks, carried out
Training；The weights and bias of network are constantly updated by error propagated forward and back-propagation algorithm, until training
Effect reaches expected；The diagnostic phases of convolutional neural networks comprise the following steps：PCB single board thermography sample preprocessing, including
Gray scale conversion and size adjusting；The convolutional neural networks that the input of pretreated PCB single board thermography trains are calculated,
Sample is classified using convolutional neural networks, classification results are failure and the not class of failure two, and then it is mono- to screen the PCB that is out of order
Plate simultaneously obtains fault rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710671754.3A CN107632251A (en) | 2017-08-08 | 2017-08-08 | PCB single board fault detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710671754.3A CN107632251A (en) | 2017-08-08 | 2017-08-08 | PCB single board fault detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107632251A true CN107632251A (en) | 2018-01-26 |
Family
ID=61099167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710671754.3A Pending CN107632251A (en) | 2017-08-08 | 2017-08-08 | PCB single board fault detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107632251A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109870617A (en) * | 2018-09-21 | 2019-06-11 | 浙江大学 | A kind of intelligent power plant electrical equipment fault diagnosis method based on width study and infrared image space-time characteristic |
| CN109975686A (en) * | 2019-03-06 | 2019-07-05 | 哈工大机器人（山东）智能装备研究院 | A kind of circuit board short circuit automatic identifying method based on infrared image processing |
| TWI677844B (en) * | 2018-07-13 | 2019-11-21 | 致伸科技股份有限公司 | Product testing system with assistance judgment function and assistance method applied thereto |
| CN112001885A (en) * | 2020-07-14 | 2020-11-27 | 武汉理工大学 | PCB fault detection method, storage medium and system based on machine vision |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101813638A (en) * | 2010-04-30 | 2010-08-25 | 华中科技大学 | Method and device for detecting welding point defect of chip on line |
-
2017
- 2017-08-08 CN CN201710671754.3A patent/CN107632251A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101813638A (en) * | 2010-04-30 | 2010-08-25 | 华中科技大学 | Method and device for detecting welding point defect of chip on line |
Non-Patent Citations (2)
| Title |
|---|
| 张斌 等: "《电子设备环境应力筛选试验研究》", 《计算机测量与控制》 * |
| 王格芳 等: "《印制电路板红外故障诊断仪》", 《仪器仪表学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI677844B (en) * | 2018-07-13 | 2019-11-21 | 致伸科技股份有限公司 | Product testing system with assistance judgment function and assistance method applied thereto |
| CN109870617A (en) * | 2018-09-21 | 2019-06-11 | 浙江大学 | A kind of intelligent power plant electrical equipment fault diagnosis method based on width study and infrared image space-time characteristic |
| CN109870617B (en) * | 2018-09-21 | 2020-08-14 | 浙江大学 | Intelligent power plant electrical equipment fault diagnosis method based on width learning and infrared image space-time characteristics |
| CN109975686A (en) * | 2019-03-06 | 2019-07-05 | 哈工大机器人（山东）智能装备研究院 | A kind of circuit board short circuit automatic identifying method based on infrared image processing |
| CN109975686B (en) * | 2019-03-06 | 2021-04-06 | 哈工大机器人（山东）智能装备研究院 | Circuit board short circuit automatic identification method based on infrared image processing |
| CN112001885A (en) * | 2020-07-14 | 2020-11-27 | 武汉理工大学 | PCB fault detection method, storage medium and system based on machine vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107632251A (en) | PCB single board fault detection method | |
| CN107677700A (en) | PCB single board failure detector | |
| CN107677949A (en) | Integrated circuit batch detector methods | |
| Brunelli et al. | Energy neutral machine learning based iot device for pest detection in precision agriculture | |
| CN108510488B (en) | Four kinds of damage detecting methods of conveyer belt based on residual error network | |
| CN101907453B (en) | Online measurement method and device of dimensions of massive agricultural products based on machine vision | |
| CN107168294A (en) | A kind of unmanned inspection monitoring method of thermoelectricity water system equipment | |
| CN204924967U (en) | Image acquisition mechanism of full -automatic AOI complete machine | |
| CN108761237A (en) | Unmanned plane electric inspection process image vital electrical component diagnoses automatically and labeling system | |
| CN111709333A (en) | Traceability early warning system and health monitoring method based on abnormal excrement of cage-raised chickens | |
| CN109636791A (en) | A kind of Citrus Huanglongbing pathogen detection method based on deep learning, apparatus and system | |
| CN110214120A (en) | Belt checks system and method | |
| EP3594666A1 (en) | Artificial intelligence-based leather inspection method and leather product production method | |
| CN105809674A (en) | Machine vision based die protection apparatus and its functioning method | |
| EP2973239A2 (en) | Network of intelligent machines | |
| CN106960216A (en) | A kind of recognition methods of agricultural machinery working equipment type and device | |
| CN206470205U (en) | Machine vision scolding tin assisted detection system | |
| CN104390982A (en) | Test method for SMT first article inspection | |
| CN104155963A (en) | Vegetable processing production line control system | |
| CN106526448A (en) | Automatic detecting system and detection method for power supply drive board | |
| CN110243826A (en) | A kind of On-line Product detection method and device | |
| Sharma et al. | Krishimitr: A cloud computing and platform for disease detection in agriculture | |
| CN107632250A (en) | Integrated circuit batch detecting device | |
| CN116205835A (en) | Circuit board flaw detection method and device and electronic equipment | |
| CN113011486A (en) | Chicken claw classification and positioning model construction method and system and chicken claw sorting method |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180126 |
|
| RJ01 | Rejection of invention patent application after publication |