CN104063873A - Shaft sleeve part surface defect on-line detection method based on compressed sensing - Google Patents

Shaft sleeve part surface defect on-line detection method based on compressed sensing Download PDF

Info

Publication number
CN104063873A
CN104063873A CN201410320804.XA CN201410320804A CN104063873A CN 104063873 A CN104063873 A CN 104063873A CN 201410320804 A CN201410320804 A CN 201410320804A CN 104063873 A CN104063873 A CN 104063873A
Authority
CN
China
Prior art keywords
image
defect
compressed sensing
matrix
detection
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
Application number
CN201410320804.XA
Other languages
Chinese (zh)
Other versions
CN104063873B (en
Inventor
谢昕
黄志刚
李慧萍
王浩然
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
East China Jiaotong University
Original Assignee
East China Jiaotong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by East China Jiaotong University filed Critical East China Jiaotong University
Priority to CN201410320804.XA priority Critical patent/CN104063873B/en
Publication of CN104063873A publication Critical patent/CN104063873A/en
Application granted granted Critical
Publication of CN104063873B publication Critical patent/CN104063873B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

Disclosed is a shaft sleeve part surface defect on-line detection method based on compressed sensing. Compressed sensing description of a part surface defect image is built through a machine vision and compressed sensing method, and an optical imaging and defect detection model highlighting surface defects is built; a part sample image of typical defects is collected, after denoising and necessary image preprocessing are carried out, sampling frequency adjustment and size normalization are carried out, a sample is trained and a redundant dictionary is built; a proper orthogonal basis decomposition matrix and a random observation matrix are designed, a combined orthogonal matching pursuit algorithm is selected, solution of the minimum norm l0 is converted into the problem of solving the optimal solution to reconstruct a defect image, spare representation of the image to be detected is calculated, and defect recognition is carried out on a part to be detected according to built judgment and recognition standards. An on-line detection system with the functions of feeding, positioning and adjustment, image collection, image processing, defect detection and recognition, part separation and the like is built, and rapid detection on the surface defects of the shaft sleeve part is achieved.

Description

A kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing
Technical field
The present invention relates to a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing, belong to the online technical field of nondestructive testing of mechanical component.
Background technology
Along with progressively shift to China at global manufacturing center, when production and manufacturing capacity constantly expand, people have also proposed more and more higher requirement to product quality.Product surface quality is important component part wherein, its direct use or deep processing important to product.Existence in some application surface imperfection may bring about great losses to user, must carry out strict detection and control.
Model For The Bush-axle Type Parts mainly plays the effects such as support, guiding and location, in all kinds of machines and instrument application very extensive, generally select the materials such as steel, cast iron, bronze or brass to manufacture.In the processes such as the production of part, carrying, assembling, surface, the end face etc. of part located to produce the defects such as various cuts, polishing scratch, pit.The usability to part is produced harmful effect by these defects, and affect rotating accuracy, vibration, noise, sealing and the serviceable life of whole machine.
Domestic many manufacturing enterprises still adopt traditional artificial sampling observation to carry out surface defects of products detection at present, and not only detection speed causes erroneous judgement, undetected slowly but also very easily, cannot meet the requirement of modern enterprise Improving The Quality of Products.In recent years, successively occur methods such as infra-red inspection, ultrasonic Detection Method, EDDY CURRENT and magnetic powder inspection, under industry spot high power pulse noise circumstance, be difficult to reach well-content effect, cannot meet the demand of modern surveying technology to the aspect such as online, accurate, real-time, comprehensive.Along with manufacturing industry is towards the target development of lean production and zero-fault, explore objective, effective, quick, reliable quality control scheme, improve the online detection level of surface defects of parts vision, become many enterprises problem in the urgent need to address.
The existing various defects detection schemes based on machine vision, because large, the on-the-spot noise of image sampling data volume is strong, in the face of different detected objects need to adopt specific figure image intensifying, rim detection, defect Segmentation algorithm, be difficult to meet real-time, the versatility requirement of defects detection.Emerging compressed sensing (Compressed Sensing in recent years, be called for short CS) theory, it is the major transformation to signal sampling theory, be widely used in fields such as system imaging, image co-registration, target identification and image tracking, can predict this technology will exert far reaching influence at part field of non destructive testing.
Summary of the invention
The object of the invention is, detect and have the problems such as collection capacity is large, data user rate is not high, real-time is poor according to existing machinery surface defects of parts, proposed a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing.
Technical scheme of the present invention is, gathers typical defect part sample image, after denoising and necessary pre-service, carries out sample frequency adjustment and size normalization, and training sample is also set up redundant dictionary; The detection platform of design jack catchs type telescoping mechanism, squeezes swollen inner cylinder face and realizes the location clamping to part to be measured while utilizing claw to open; For fear of the damage to inside surface, claw part will be pasted soft ground material.Part to be measured is placed in the detection station of platform, and when detection, axis keeps vertical, and the optical axis of camera lens, camera keeps vertical with axis of workpiece, and the structure that optical device is installed is set near platform.Part image to be measured is sampled, and design random observation matrix, selects united orthogonal matching pursuit algorithm, thereby the rarefaction representation of calculating testing image is realized ONLINE RECOGNITION and the detection of part defect.The intermittent-rotation motion of detection platform adopts motion control card to realize via driver control stepper motor.Sorting mechanism can utilize the break-make of electromagnet to control the switching of door, in detecting, part is produced to new damage, buffering link will be set in the place of componentselected blanking sorted part is carried out to necessary protection.
The present invention has designed a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing:
(1) the Model For The Bush-axle Type Parts sample image of typical defect is carried out to pre-service, comprise image filtering denoising, enhancing, rim detection and defect Segmentation, to obtain defect area;
(2) sample image after treatment is implemented to sampling rate adjusting, the experiment of normalization size, train sample image and set up sample redundant dictionary according to compressed sensing principle;
(3) Model For The Bush-axle Type Parts is placed in detection station, and part image to be measured is pressed to certain frequency sampling, calculates its observing matrix; According to sample redundant dictionary, calculate the rarefaction representation of testing image, thereby realize the identification to part defect to be measured according to the discrimination standard of compressed sensing, and drive sorting mechanism in pick-up unit to carry out online classification.
Realizing the inventive method comprises the following steps:
(1) build high-efficiency high-accuracy detection platform
Design high-precision imaging system and detection system.
When design imaging system, according to selecting factors industrial camera and camera lenses such as accuracy of detection, visual field sizes, through the definite light source of experiment, to obtain the detected image that illumination is even, defect is outstanding.
When design detection system,, according to the requirement of online automatic detection, realize the actions such as automatic charging, transmission, location, clamping, intermittent rotary, discharging and sorting.Intermittent-rotation detection platform and positioning clamping device need to be assembled together, and are organically connected with conveyer simultaneously; The part completing after detection arrives assigned address by discharging, conveying process, completes the sorting of part according to identification judged result control sorting mechanism.In order to adapt to the measurement of different size part, installation and the adjustment link of camera to be set at mechanical part.Design mechanical motion cycle chart, coordinate the action sequence of each mechanism.
(2) rarefaction representation, encoding measurement and the restructing algorithm of the defect image signal of foundation based on compressed sensing
The present invention is directed to the problems such as part image sampling rate is high, efficiency is low, set up rarefaction representation, encoding measurement and the restructing algorithm of the picture signal based on compressed sensing, the quick method that effectively detects surface imperfection, analyze complexity and the stability of compressed sensing algorithm, and be applied to the online detection of surface defects of parts.
(3) surface defects of parts of realizing based on compressed sensing detects online.The present invention describes according to the compressed sensing of part defect image, determines denoising, enhancing, rim detection, the defect Segmentation algorithm of defect sample image; The experiment that sample frequency, normalization adjusted size, the training sample image of enforcement sample image set up redundant dictionary; Setting up part defect detects mathematical model, defect and judges criterion of identification; Design orthogonal basis split-matrix and random observation matrix, solving minimum norm l 0be converted into the suboptimum solution problem of asking, the sparse reconstruction defect image of implementation structure; Calculate the rarefaction representation of testing image, Model For The Bush-axle Type Parts is carried out to accuracy, the robustness of the online test experience of defect analytical algorithm.
The invention has the beneficial effects as follows, the improved method of the present invention can be eliminated surface reflection impact, set up to give prominence to imaging and the defects detection model of defect as target; Realize image registration and the Image Fusion under industrial condition, removing the filtering algorithm of very noisy in image and part outer post surface multiple image is carried out to efficient splicing; The rarefaction representation of the defect image based on compressed sensing; The selection of orthogonal basis split-matrix and random observation matrix and design, the realization of part image restructing algorithm, defect judges the foundation of criterion of identification and the structure of defect on-line detecting system.The present invention's method based on compressed sensing feature at line drawing Model For The Bush-axle Type Parts, shortens the processing time greatly, realizes online and detecting.
The present invention is applicable to mechanical axis cover parts surface imperfection and detects online.
Brief description of the drawings
Fig. 1 is the Technology Roadmap of the inventive method;
Fig. 2 is one of surface defects of parts;
Fig. 3 is two of surface defects of parts;
Fig. 4 is three of surface defects of parts;
In figure, drawing circle place is surface defects of parts place.
Embodiment
The specific embodiment of the present invention as shown in Figure 1.
The present embodiment adopts machine vision and compression sensing method, the compressed sensing of research surface defects of parts image is described, and gathers the part sample image of typical defect, implements after the pre-service of denoising and necessity, carry out sample frequency adjustment and size normalization, training sample is also set up redundant dictionary; Design appropriate orthogonal basis split-matrix and random observation matrix, select united orthogonal matching pursuit algorithm, be converted into and ask suboptimum solution problem with reconstruction defect image solving minimum norm L0, calculate the rarefaction representation of testing image and according to set up judgement criterion of identification, part to be measured carried out to defect recognition, realizing the fast detecting to Model For The Bush-axle Type Parts surface imperfection.
(1) design detection platform
Determining of imaging system light source, illumination methods, industrial camera and coupling camera lens:
The material of Model For The Bush-axle Type Parts is mainly metal, and surface ratio is brighter and cleaner.Under the irradiation of light source, gray scale that flawless measured object surface image presents, color, texture are relatively evenly and without sudden change, the defective surface of tool exists sudden change, and this can be used as the foundation of defect estimation.Because the defects such as cut, polishing scratch, pit present different characteristics of image under different light sources, lighting system, image acquisition mode, can set up thus defect sample storehouse, piece surface typical defect is as shown in Figure 2, Figure 3 and Figure 4.
According to the feature of defect target and detection demand, reduce the impact of surface reflection as far as possible, accurately control the irradiating angle of light, make reflected light can not directly enter camera, be reflected to other direction, the scattered light that defectiveness or tested feature cause enters lens imaging, and method is by experiment determined light source type and illumination methods.The key issue of vision detection system structural design, is exactly the demand according to detection resolution, and the structural parameters such as object distance, focal length and visual field are determined.If the detection resolution of measured surface defect requires as amm, camera pixel resolution is a μ m, and pixel number is p.System magnification β is:
β = 2 b × 10 - 3 a
Visual field FOV can be calculated by following formula, that is:
FOV = b × p × 10 - 3 β
The pass of object distance l and focal distance f is:
f = β × l β + 1
Determine accordingly the camera lens matching, the requirements such as size, the visual field size of simultaneously taking into account tested part and resolution, the industrial camera (not needing capture card) of select tape PCI-Express interface, with obtain illumination evenly, be full of whole visual field, focus on accurately, HD image that target is outstanding.
The design of mechanical system:
According to the detectability of system (ten thousand/year), work number of days (day/year), order of classes or grades at school and working time (hour), can be calculated as follows the required time of piece test (comprising the links such as location, clamping, three intermittent rotaries and discharging), that is:
Consider operation setup time and suitable collating time, t can suitably shorten a bit, can determine accordingly speed, work table rotation speed that part transmits.T is segmented, and the motion cycle chart of design system, ensures the strict sequential order relation of each mechanism action, completes machine driven system design.The corresponding raceway of external cylindrical surface characteristics design in conjunction with part solves the problems such as Model For The Bush-axle Type Parts automatic charging, conveying.
The present embodiment proposes meter jack catchs type telescoping mechanism, utilizes crowded swollen inner cylinder face to realize part to be measured is positioned to clamping in the time that three claws open, and avoids the damage to inside surface, and claw part will be pasted soft ground material.The detection pose of part is upright, and the optical axis of camera lens, camera keeps vertical with axis of workpiece, the structure that optical device is installed will be set near detection platform, for the fine setting of follow-up equipment is provided convenience.The intermittent-rotation motion of detection platform adopts motion control card to realize via driver control stepper motor.Sorting mechanism can utilize the break-make of electromagnet to control the switching of door, in detecting, part is produced to new damage, buffering link will be set in the place of componentselected blanking sorted part is protected.
(2) select sample
According to the defective part image of the tool obtaining from scene, therefrom intercept out defect, noise section, then pick out typical defect and noise image as recognition sample.
(3) pre-service of sample image and set up redundant dictionary
For the serious sample image of noise, first determine that by testing filtering algorithm realizes effective denoising, then strengthen, the pre-service such as rim detection and defect Segmentation, to obtain defect area.The defect image identification of Model For The Bush-axle Type Parts can be described as the linear combination of sample image, has:
Y=Ax
In formula: Y is testing image; A is the dictionary matrix being made up of sample image; X=[x 1, x 2..., x n] ∈ R nfor coefficient vector.In database, with a m dimensional vector y ∈ R mrepresent a sample image, have n to open different sample image y 1, y 2..., y n∈ R m, for a testing image, have:
Y = Σ j = 1 m x j y j
In formula: x jfor linear expression coefficient.Write as matrix form, made [y 1, y 2..., y n] ∈ R m × n, have:
y m×1=A m×nx n×1
In compressed sensing framework, measure and encode and carry out simultaneously, that is to say, can directly obtain the measured value after compression.
The present embodiment compressed sensing mainly comprises:
1. rarefaction representation
Signal, can rarefaction representation under a certain transform domain:
f=ψx
In formula: f represents N × 1 type original signal; ψ is N × N-type transform domain matrix; X is exactly the rarefaction representation of original signal f under transform domain ψ, only has K large coefficient in x, and all the other coefficients can be approximated to be 0.
2. encoding measurement
The object of encoding measurement is to find suitable sample frequency to carry out the sampling of signal, the adaptive linear matrix of matrix right and wrong that this sampling obtains.The key of this part be look for one with the observing matrix φ of the incoherent M × N-type of ψ, wherein M < < N, only need to instead of sample N time to signal sampling M time, sample frequency and data volume all greatly reduce.
y=φf
In formula: the linear projection measured value that y is original signal.
3. signal reconstruction
Signal reconstruction process is generally converted to the optimization problem of a minimum norm.
y = &phi;&Theta;x = &Theta;x
In formula: be called compressed sensing matrix.
Due to signal, x has sparse characteristic, as y and while meeting certain condition, theoretical proof x can be by measured value y by solving optimum l 0norm problem and obtaining,
The successful Application of compressed sensing technology must depend on numerous effective algorithms and realize, and these algorithms are mainly divided into greedy match tracing class algorithm, protruding lax class algorithm and statistics and optimize class algorithm three classes.The minimum l of greedy match tracing class algorithm approximate solution 0norm problem, in each iteration, all select one group of atom of local optimum, typically there is united orthogonal match tracing (Simultaneous Orthogonal Matching Pursuit, SOMP), structure greedy algorithm (Structured Greedy Algorithm, SGA) etc.; Protruding lax class algorithm is to use l 1norm is penalized to punish and is replaced l 0norm punishment, and by solving protruding rule problem reformulation compressed signal, typical algorithm has approximate data (Fast Iterative Shrinkage-Thresholding Algorithm, FISTA), piece coordinate descent algorithm etc.; It is to utilize total probability model and statistical inference instrument to carry out reconstruct sparse signal that statistics is optimized class algorithm, and main algorithm has piece sparse Bayesian study (Block Sparse Bayesian learning, BSBL) etc.
Although solve minimum norm l 0a np problem, but can be formula in actual computation be converted into and ask suboptimum solution, the present invention adopts united orthogonal matching pursuit algorithm, not only can reduce iterations, can also ensure the optimality of each iteration.
The required matrix size of compressed sensing is unified, need to carry out to sample image the normalization of size, sample frequency is closely related with the size of the amount of image information gathering, and adjusts by experiment sample frequency and the image normalization size of image, records one by one testing result.Pretreated sample image is sampled, according to compressed sensing principle, sample image is trained, set up redundant dictionary and judge criterion of identification.
(4) rarefaction representation of part to be measured
Obtain part image to be measured, carry out denoising (for improving Detection accuracy, also can carry out simple image enhancement processing).Calculate the rarefaction representation of part image to be measured by redundant dictionary, judge that according to compressed sensing image criterion of identification just can judge whether defectiveness or noise of testing image.
(5) repetitive operation check testing result
The present embodiment is adjusted sample image pretreatment parameter, checks testing result record; If can not get desirable testing result according to pretreated sample, so again select sample image, then repeat above calculating process, until obtain desirable testing result.

Claims (6)

1. the Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing, is characterized in that, described method step is:
(1) build high-efficiency high-accuracy detection platform, design high-precision imaging and detection system;
(2), efficiency low problem high for part image sampling rate, set up rarefaction representation, encoding measurement and the restructing algorithm of the picture signal based on compressed sensing, the quick method that effectively detects surface imperfection, analyze complexity and the stability of compressed sensing algorithm, and be applied to the online detection of surface defects of parts;
(3) describe according to the compressed sensing of part defect image, determine denoising, enhancing, rim detection, the defect Segmentation algorithm of defect sample image; The experiment that sample frequency, normalization adjusted size, the training sample image of enforcement sample image set up redundant dictionary; Setting up part defect detects mathematical model, defect and judges criterion of identification; Design orthogonal basis split-matrix and random observation matrix, solving minimum norm l 0be converted into the suboptimum solution problem of asking, the sparse reconstruction defect image of implementation structure; Calculate the rarefaction representation of testing image, Model For The Bush-axle Type Parts is carried out to accuracy, the robustness of the online test experience of defect analytical algorithm.
2. a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing according to claim 1, it is characterized in that, described imaging and detection system comprise, light source, industrial camera and coupling camera lens, sorting mechanism, overall mechanical hook-up and detection platform;
Described overall mechanical hook-up adopts jack catchs type telescoping mechanism, utilizes crowded swollen inner cylinder face to realize part to be measured is positioned to clamping in the time that three claws open, and avoids the damage to inside surface, and claw part will be pasted soft ground material; The detection position of part is upright, and the optical axis of camera lens, camera keeps vertical with axis of workpiece, and the structure that optical device is installed is set near detection platform;
The intermittent-rotation motion of described detection platform adopts motion control card to realize via driver control stepper motor; Described jack catchs type telescoping mechanism is arranged in described detection platform;
Described sorting mechanism utilizes the break-make of electromagnet to control the switching of door, in detecting, part is produced to new damage, buffering link will be set in the place of componentselected blanking sorted part is protected.
3. a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing according to claim 1, is characterized in that, described rarefaction representation is, signal, can rarefaction representation under a certain transform domain:
f=ψx
In formula: f represents N × 1 type original signal; ψ is N × N-type transform domain matrix; X is exactly the rarefaction representation of original signal f under transform domain ψ.
4. a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing according to claim 1, it is characterized in that, described encoding measurement is to find suitable sample frequency to carry out the sampling of signal, the adaptive linear matrix of matrix right and wrong that this sampling obtains; Look for one with the observing matrix φ of the incoherent M × N-type of ψ, wherein M < < N, only need to instead of sample N time to signal sampling M time, sample frequency and data volume all greatly reduce;
y=φf
In formula: the linear projection measured value that y is original signal.
5. a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing according to claim 1, is characterized in that, described restructing algorithm process is generally converted to the optimization problem of a minimum norm;
In formula: be called compressed sensing matrix;
Due to signal, x has sparse characteristic, as y and while meeting certain condition, x can be by measured value y by solving optimum l 0norm problem and obtaining,
6. a kind of Model For The Bush-axle Type Parts surface imperfection online test method based on compressed sensing according to claim 1, is characterized in that, the defect image identification of described Model For The Bush-axle Type Parts can be described as the linear combination of sample image, has:
Y=Ax
In formula: Y is testing image; A is the dictionary matrix being made up of sample image; X=[x 1, x 2..., x n] ∈ R nfor coefficient vector; In database, with a m dimensional vector y ∈ R mrepresent a sample image, have n to open different sample image y 1, y 2..., y n∈ R m, for a testing image, have:
In formula: x jfor linear expression coefficient; Write as matrix form, made [y 1, y 2..., y n] ∈ R m × n, have:
y m×1=A m×nx n×1
CN201410320804.XA 2014-07-08 2014-07-08 A kind of Model For The Bush-axle Type Parts surface defect online test method based on compressed sensing Expired - Fee Related CN104063873B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410320804.XA CN104063873B (en) 2014-07-08 2014-07-08 A kind of Model For The Bush-axle Type Parts surface defect online test method based on compressed sensing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410320804.XA CN104063873B (en) 2014-07-08 2014-07-08 A kind of Model For The Bush-axle Type Parts surface defect online test method based on compressed sensing

Publications (2)

Publication Number Publication Date
CN104063873A true CN104063873A (en) 2014-09-24
CN104063873B CN104063873B (en) 2017-11-14

Family

ID=51551565

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410320804.XA Expired - Fee Related CN104063873B (en) 2014-07-08 2014-07-08 A kind of Model For The Bush-axle Type Parts surface defect online test method based on compressed sensing

Country Status (1)

Country Link
CN (1) CN104063873B (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105812299A (en) * 2016-04-22 2016-07-27 中国地质大学(武汉) Channel estimation algorithm and system of wireless sensor network based on joint block sparse reconstruction
CN108288288A (en) * 2018-01-16 2018-07-17 华东交通大学 Accurate shaft size measurement method, the device and system of view-based access control model identification
CN108711151A (en) * 2018-05-22 2018-10-26 广东工业大学 A kind of weld defects detection method, apparatus, equipment, storage medium and system
CN108830909A (en) * 2018-05-24 2018-11-16 武汉精测电子集团股份有限公司 Promote the image preprocessing system and method for period texture image compression ratio
CN109063738A (en) * 2018-07-03 2018-12-21 浙江理工大学 A kind of ceramic water valve plates automatic on-line detection method of compressed sensing
CN109521742A (en) * 2018-12-05 2019-03-26 西安交通大学 A kind of control system and control method for electric rotary body
CN109884180A (en) * 2019-02-14 2019-06-14 昆明理工大学 A kind of sparse current vortex fast imaging detection method of conductive structure defect and system
CN110009618A (en) * 2019-04-02 2019-07-12 浙江大学 A kind of Axle Surface quality determining method and device
CN110044927A (en) * 2019-04-23 2019-07-23 华中科技大学 A kind of detection method of space encoding light field to bend glass surface defect
CN110346141A (en) * 2019-06-20 2019-10-18 燕山大学 Sparse coding Fault Diagnosis of Roller Bearings certainly
CN111879344A (en) * 2020-06-24 2020-11-03 董永康 Fast Brillouin optical time domain analyzer and method based on frequency agility and CS technology
CN112284728A (en) * 2020-09-30 2021-01-29 华南理工大学 Segmented sparse compression and reconstruction method for local fault characteristics of rotary machine
CN112668090A (en) * 2020-12-02 2021-04-16 成都飞机工业(集团)有限责任公司 Engine power split shaft adjusting method based on space orthogonal decomposition
CN113642680A (en) * 2021-10-13 2021-11-12 常州微亿智造科技有限公司 Edge synthesis and hypersphere soft fitting defect detection method
CN113916793A (en) * 2021-09-18 2022-01-11 华南理工大学 Non-contact laser ultrasonic damage detection method and system based on sparse array excitation
CN114937043A (en) * 2022-07-26 2022-08-23 中国工业互联网研究院 Equipment defect detection method, device, equipment and medium based on artificial intelligence
CN114937043B (en) * 2022-07-26 2022-10-25 中国工业互联网研究院 Equipment defect detection method, device, equipment and medium based on artificial intelligence

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103714534A (en) * 2013-12-12 2014-04-09 河海大学 Material surface defect detection method based on compressed sensing

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103714534A (en) * 2013-12-12 2014-04-09 河海大学 Material surface defect detection method based on compressed sensing

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
ALLEN Y. YANG 等: "Towards a robust face recognition system using compressive sensing", 《INTERSPEECH》 *
崔亚楠 等: "基于压缩传感的焊管焊缝X射线图像处理", 《焊接技术》 *
崔亚楠: "基于压缩传感理论的X射线焊缝图像处理", 《中国优秀硕士学位论文全文数据库信息科技辑》 *
杜玉军 等: "基于机器视觉的闭气塞表面缺陷自动检测系统", 《理论与方法》 *
柴汝刚: "基于机器视觉的轮毂在线识别系统研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Cited By (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105812299B (en) * 2016-04-22 2020-05-15 中国地质大学(武汉) Wireless sensor network channel estimation method based on joint block sparse reconstruction
CN105812299A (en) * 2016-04-22 2016-07-27 中国地质大学(武汉) Channel estimation algorithm and system of wireless sensor network based on joint block sparse reconstruction
CN108288288A (en) * 2018-01-16 2018-07-17 华东交通大学 Accurate shaft size measurement method, the device and system of view-based access control model identification
CN108288288B (en) * 2018-01-16 2021-09-10 华东交通大学 Method, device and system for measuring precision shaft dimension based on visual identification
CN108711151A (en) * 2018-05-22 2018-10-26 广东工业大学 A kind of weld defects detection method, apparatus, equipment, storage medium and system
CN108830909A (en) * 2018-05-24 2018-11-16 武汉精测电子集团股份有限公司 Promote the image preprocessing system and method for period texture image compression ratio
CN108830909B (en) * 2018-05-24 2022-08-19 武汉精测电子集团股份有限公司 Image preprocessing system and method for improving compression ratio of periodic texture image
CN109063738A (en) * 2018-07-03 2018-12-21 浙江理工大学 A kind of ceramic water valve plates automatic on-line detection method of compressed sensing
CN109063738B (en) * 2018-07-03 2021-12-21 浙江理工大学 Automatic online detection method for compressed sensing ceramic water valve plate
CN109521742A (en) * 2018-12-05 2019-03-26 西安交通大学 A kind of control system and control method for electric rotary body
CN109884180A (en) * 2019-02-14 2019-06-14 昆明理工大学 A kind of sparse current vortex fast imaging detection method of conductive structure defect and system
CN110009618A (en) * 2019-04-02 2019-07-12 浙江大学 A kind of Axle Surface quality determining method and device
CN110009618B (en) * 2019-04-02 2021-04-20 浙江大学 Shaft part surface quality detection method and device
CN110044927A (en) * 2019-04-23 2019-07-23 华中科技大学 A kind of detection method of space encoding light field to bend glass surface defect
CN110346141A (en) * 2019-06-20 2019-10-18 燕山大学 Sparse coding Fault Diagnosis of Roller Bearings certainly
CN111879344A (en) * 2020-06-24 2020-11-03 董永康 Fast Brillouin optical time domain analyzer and method based on frequency agility and CS technology
CN112284728A (en) * 2020-09-30 2021-01-29 华南理工大学 Segmented sparse compression and reconstruction method for local fault characteristics of rotary machine
CN112284728B (en) * 2020-09-30 2022-03-29 华南理工大学 Segmented sparse compression and reconstruction method for local fault characteristics of rotary machine
CN112668090B (en) * 2020-12-02 2022-04-08 成都飞机工业(集团)有限责任公司 Engine power split shaft adjusting method based on space orthogonal decomposition
CN112668090A (en) * 2020-12-02 2021-04-16 成都飞机工业(集团)有限责任公司 Engine power split shaft adjusting method based on space orthogonal decomposition
CN113916793A (en) * 2021-09-18 2022-01-11 华南理工大学 Non-contact laser ultrasonic damage detection method and system based on sparse array excitation
CN113642680A (en) * 2021-10-13 2021-11-12 常州微亿智造科技有限公司 Edge synthesis and hypersphere soft fitting defect detection method
CN113642680B (en) * 2021-10-13 2022-02-08 常州微亿智造科技有限公司 Edge synthesis and hypersphere soft fitting defect detection method
CN114937043A (en) * 2022-07-26 2022-08-23 中国工业互联网研究院 Equipment defect detection method, device, equipment and medium based on artificial intelligence
CN114937043B (en) * 2022-07-26 2022-10-25 中国工业互联网研究院 Equipment defect detection method, device, equipment and medium based on artificial intelligence

Also Published As

Publication number Publication date
CN104063873B (en) 2017-11-14

Similar Documents

Publication Publication Date Title
CN104063873A (en) Shaft sleeve part surface defect on-line detection method based on compressed sensing
Yang et al. Real-time tiny part defect detection system in manufacturing using deep learning
CN102854191B (en) Real-time visual detection and identification method for high speed rail surface defect
CN109499908A (en) A kind of vision detection system for machine components
CN109584227A (en) A kind of quality of welding spot detection method and its realization system based on deep learning algorithm of target detection
CN102621156B (en) Image-processing-based automatic micro part sorting system
CN106706653A (en) High-speed wide board detection method
CN102914545A (en) Gear defect detection method and system based on computer vision
CN110473178A (en) A kind of open defect detection method and system based on multiple light courcess fusion
CN101576956A (en) On-line character detection method based on machine vision and system thereof
CN110186375A (en) Intelligent high-speed rail white body assemble welding feature detection device and detection method
CN111929314A (en) Wheel hub weld visual detection method and detection system
CN109840900A (en) A kind of line detection system for failure and detection method applied to intelligence manufacture workshop
CN104316530A (en) Part detection method and application
CN104200215A (en) Method for identifying dust and pocking marks on surface of big-caliber optical element
CN103674970B (en) A kind of automatic testing method producing band in batches
CN111103307A (en) Pcb defect detection method based on deep learning
CN107300562B (en) X-ray nondestructive testing method for measuring contact distance of finished relay product
Wu et al. An end-to-end learning method for industrial defect detection
CN110969357A (en) Visual detection method for holes of aluminum alloy machined part
Sun et al. Precision work-piece detection and measurement combining top-down and bottom-up saliency
CN204440490U (en) A kind of Model For The Bush-axle Type Parts surface imperfection on-line measuring device based on compressed sensing
Liang et al. Research on Surface Defect Detection Algorithm of Tube-Type Bottle Based on Machine Vision
Zhang et al. Intelligent detection method for welding seam defects of automobile wheel hub based on yolo
Yu et al. Surface Defect Detection of Hight-speed Railway Hub Based on Improved YOLOv3 Algorithm

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20171114

Termination date: 20180708

CF01 Termination of patent right due to non-payment of annual fee