CN109919941A - Internal screw thread defect inspection method, device, system, equipment and medium - Google Patents
Internal screw thread defect inspection method, device, system, equipment and medium Download PDFInfo
- Publication number
- CN109919941A CN109919941A CN201910250375.6A CN201910250375A CN109919941A CN 109919941 A CN109919941 A CN 109919941A CN 201910250375 A CN201910250375 A CN 201910250375A CN 109919941 A CN109919941 A CN 109919941A
- Authority
- CN
- China
- Prior art keywords
- screw thread
- internal screw
- image
- defect
- model
- 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
Abstract
The invention discloses a kind of internal screw thread defect inspection method, device, system, computer equipment and storage mediums, first obtain the internal screw thread image for the object that image capture module is sent, the internal screw thread image is the image including the complete internal screw thread information of the object that described image acquisition module is acquired by the component prism and described image acquisition component, it is ensured that the efficiency and precision of internal screw thread Image Acquisition；Further, the internal screw thread image is identified by preset defects detection model, obtains the defects detection information of the object.It ensure that the efficiency and accuracy of the identification of the defects of internal threads image.
Description
Technical field
The present invention relates to automatic field more particularly to a kind of internal screw thread defect inspection method, device, system, equipment and
Medium.
Background technique
It is higher and higher for very multipart quality requirement with technologically continuing to develop.Exist in many components interior
Screw thread, and the quality of internal screw thread quality can have a great impact to the realization of complete machine or function.For example, smart phone camera
Voice coil motor (VCM) be mobile phone important spare part, to adjust camera focal length play the role of it is vital.Wherein,
It will have a direct impact on its performance if the internal screw thread existing defects of voice coil motor, must then be examined before voice coil motor comes into operation
It finds there are open defect individual, avoids the handset capability produced low.
Artificial detection state is also rested on for the open defect detection of the internal screw thread of components at present, not only needs to expend
A large amount of manpower and material resources, and it is also easy to the problem of low detection efficiency, accuracy rate decline occur in testing staff's fatigue.
Summary of the invention
The embodiment of the present invention provides a kind of internal screw thread defect inspection method, device, system, equipment and medium, in solving
The not high problem of threading defects detection efficiency.
A kind of internal screw thread defect inspection method, the internal screw thread defect inspection method are applied to internal screw thread defect detecting device
Image processing module in, the internal screw thread defect detecting device further includes image capture module, described image acquisition module packet
Include component prism and image acquisition component；
The internal screw thread defect inspection method includes:
The internal screw thread image for the object that described image acquisition module is sent is obtained, the internal screw thread image is described image
What acquisition module was acquired by the component prism and described image acquisition component includes the complete internal screw thread information of the object
Image；
The internal screw thread image is identified by preset defects detection model, obtains the defect inspection of the object
Measurement information.
A kind of internal screw thread defect detecting device, including image processing module and image capture module, the internal screw thread defect
Detection device is used to detect the internal screw thread defect of object；
Described image acquisition module is used to acquire the internal screw thread image of the object, and the internal screw thread image is sent
To image processing module, the internal screw thread image is that described image acquisition module is acquired by component prism and image acquisition component
The image including the complete internal screw thread information of the object；
The internal screw thread image is input to preset for receiving the internal screw thread image by described image processing module
It is identified in defects detection model, obtains the defects detection information of the object.
A kind of internal screw thread defect detecting system, including processing unit and above-mentioned internal screw thread defect detecting device；
The processing unit is for detecting the presence of at the object of defect the internal screw thread defect detecting device
Reason.
A kind of computer equipment, including memory, processor and storage are in the memory and can be in the processing
The computer program run on device, the processor realize above-mentioned internal screw thread defects detection side when executing the computer program
Method.
A kind of computer readable storage medium, the computer-readable recording medium storage have computer program, the meter
Calculation machine program realizes above-mentioned internal screw thread defect inspection method when being executed by processor.
In above-mentioned internal screw thread defect inspection method, device, system, computer equipment and storage medium, first obtains image and adopt
Collect the internal screw thread image for the object that module is sent, the internal screw thread image is that described image acquisition module passes through the prism portion
The image including the complete internal screw thread information of the object of part and the acquisition of described image acquisition component, it is ensured that internal screw thread figure
As the efficiency and precision of acquisition；Further, the internal screw thread image is identified by preset defects detection model, is obtained
To the defects detection information of the object.It ensure that the efficiency and accuracy of the identification of the defects of internal threads image.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the schematic diagram of internal screw thread defect detecting device in one embodiment of the invention；
Fig. 2 is a flow chart of internal screw thread defect inspection method in one embodiment of the invention；
Fig. 3 is the exemplary diagram of an internal screw thread image of the invention；
Fig. 4 is another flow chart of internal screw thread defect inspection method in one embodiment of the invention；
Fig. 5 is another flow chart of internal screw thread defect inspection method in one embodiment of the invention；
Fig. 6 is another flow chart of internal screw thread defect inspection method in one embodiment of the invention；
Fig. 7 is another flow chart of internal screw thread defect inspection method in one embodiment of the invention；
Fig. 8 is an application schematic diagram of internal screw thread defect detecting device in one embodiment of the invention；
Fig. 9 is a schematic diagram of computer equipment in one embodiment of the invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall within the protection scope of the present invention.
Internal screw thread defect inspection method provided in an embodiment of the present invention can be applicable to the internal screw thread defects detection dress such as Fig. 1
In setting, wherein image capture module is communicatively coupled with image processing module.Image capture module includes component prism and figure
As acquisition component.Wherein, image processing module can be, but not limited to various CCD cameras, personal computer, laptop, intelligence
It can mobile phone, tablet computer and other equipment with image collecting function.
In one embodiment, it as shown in Fig. 2, providing a kind of internal screw thread defect inspection method, applies in Fig. 1 in this way
Image processing module for be illustrated, include the following steps:
S10: the internal screw thread image for the object that described image acquisition module is sent is obtained, the internal screw thread image is described
What image capture module was acquired by the component prism and described image acquisition component includes the complete internal screw thread of the object
The image of information.
Object is band internal thread component, for example, voice coil motor.Image processing module obtains image capture module hair
The internal screw thread image of the object sent.Wherein, image capture module is cooperated by component prism and image acquisition component, Ke Yitong
Cross an Image Acquisition (primary shooting), so that it may get the internal screw thread figure including the complete internal screw thread information of the object
Picture does not need to carry out multiple Image Acquisition, does not need the adjustment operation for carrying out image, so that it may which guarantee obtains the interior of object yet
Screw image.
Specifically, being used cooperatively by selection component prism and image acquisition component can be realized.Component prism can be
Trigone mirror element, four component prisms, pentaprism component or six component prisms etc. can specifically need to carry out according to available accuracy
Selection.Image acquisition component may include camera, light source and camera lens.For this sentences six component prisms, by reasonably adjusting figure
As the focal length of acquisition component, the distance and image capture module of six component prisms and image acquisition component and the position of object
It sets to guarantee to get clearly internal screw thread image.Six component prisms are used according to component prism, then internal screw thread image can be with
The internal screw thread information of six parts including object, the internal screw thread information of this six parts constitute the complete internal screw thread of object
Information.Illustratively, as shown in figure 3, Fig. 3 is the internal screw thread figure that a width acquires with six component prisms and image acquisition component
Picture.The internal screw thread image is made of 7 parts (middle section and 6 lateral parts), respectively with the bottom surface of six component prisms and 6
A side is corresponding, and obtains by the way that image acquisition component is Polaroid.
S20: identifying the internal screw thread image by preset defects detection model, obtains lacking for the object
Fall into detection information.
Defects detection model is the model that the defect for the internal threads that training obtains in advance is detected.It specifically, can be with
In advance machine learning model is trained to obtain using great amount of samples.Illustratively, acquisition largely has internal screw thread figure in advance
Then the sample image of picture is labeled each sample image, which can be whether the corresponding sample image of mark is deposited
In defect and specific defective locations.It is to be appreciated that sample image can be acquired by above-mentioned image capture module
It arrives.Further, machine learning model is trained using sample image, obtains defects detection model.
In this step, the internal screw thread image is identified by preset defects detection model, obtains the mesh
Mark the defects detection information of object.Specifically, internal screw thread image can be input in defects detection model, object can be obtained
Defects detection information.The defects detection information of object may include existing defects and there is no defects.Further, if it is interior
Screw image existing defects, then defects detection information further includes the location information of defect, which can pass through pixel
Mode embody.
In the present embodiment, the internal screw thread image for the object that image capture module is sent, the internal screw thread figure are first obtained
It is complete including the object as being acquired for described image acquisition module by the component prism and described image acquisition component
The image of whole internal screw thread information, it is ensured that the efficiency and precision of internal screw thread Image Acquisition；Further, pass through preset defect
Detection model identifies the internal screw thread image, obtains the defects detection information of the object.It ensure that internal threads
The efficiency and accuracy of the defects of image identification.
In one embodiment, as shown in figure 4, described carry out the internal screw thread image by preset defects detection model
Identification, obtains the defects detection information of the object, comprising:
S21: image preprocessing is carried out to the internal screw thread image, obtains pretreatment image.
Specifically, pretreatment may include filtering and noise reduction, the increase of Characteristic Contrast degree etc..It is carried out by internal threads image pre-
Processing, can be improved the precision of subsequent detection.
S22: the pretreatment image is normalized, image to be detected is obtained.
A standard size can be preset, then the pretreatment image is normalized, is obtained and gauge
Very little an equal amount of image to be detected.It is to be appreciated that the sample image of corresponding defects detection model also includes a basis
The process that standard size is normalized.
S23: described image to be detected is input in preset defects detection model and is detected, the object is obtained
Defects detection information.
It is obtaining image to be detected and then is being input to image to be detected in preset defects detection model to know
Not to get the defects detection information for arriving the object.
In the present embodiment, pretreatment and normalized are carried out by internal threads image and then passes through defect
Detection model detects the internal screw thread image of object, improves the precision of detection.
In one embodiment, as shown in figure 5, it is described by preset defects detection model to the internal screw thread image into
Row identification, after obtaining the defects detection information of the object, the internal screw thread defect inspection method further include:
S30: if the defects detection information is existing defects, where obtaining defect described in the internal screw thread image
Pixel region.
If the defects detection information is existing defects, shows existing defects in the internal screw thread of object, obtain at this time
Pixel region where defect, preferably to be positioned to defect.It is to be appreciated that defect can be one, it can also be with
It is at least two.And the number in pixel region is identical with the number of defect.Specifically, pixel region can be square
Shape, circle or other shapes, details are not described herein.Preferably, pixel region is rectangle, and pixel region can lead at this time
The pixel information on four vertex of the rectangle is crossed to determine.
S40: according to location information of the defect in the object described in the pixel Area generation.
Specifically, the mapping relations in a pixel region and physical location can be preestablished.Physical location refers to scarce
The specific location being trapped in the internal screw thread of object.Specifically, it can be directed to the internal screw thread of internal screw thread image and object respectively
Construct pixel coordinate system and internal screw thread coordinate system.The mapping for pre-establishing pixel coordinate system and internal screw thread coordinate system again is closed
System.It is alternatively possible to construct corresponding mapping relations according to the positional relationship of image capture module and object, or pass through
Experiment is to determine mapping relations.After obtaining pixel region, so that it may generate the defect in institute according to the mapping relations
State the location information in the internal screw thread of object.
It preferably, further include the process of a pixel region merging technique after obtaining pixel region.In internal screw thread figure
There may be intersections for the corresponding internal screw thread information of every one side in component prism as in.The intersection is generally in every side
Edge in the corresponding internal screw thread information in face.It include overlapping region i.e. in internal screw thread image, if pixel region is located at overlay region
In domain, then pixel region substantially identical in overlapping region is merged, to reduce subsequent calculation amount, improves processing effect
Rate.
Specifically, which can be pre-configured with, and after obtaining pixel region, whether judge pixel region
In overlapping region.Pixel region substantially identical in overlapping region is merged again.The overlapping region can adjust
The focal length of image acquisition component, the distance and image capture module of six component prisms and image acquisition component and object
It is determined when position.
In the present embodiment, by the way that the defects of internal screw thread image to be carried out to the mapping of substantive location information, with preferably
Defective locations are presented.
In one embodiment, as shown in fig. 6, it is described by preset defects detection model to the internal screw thread image into
Row identification, before obtaining the defects detection information of the object, the internal screw thread defect inspection method further include:
S11: internal screw thread sample set is obtained, the internal screw thread sample set includes sample image and each sample image
Labeled data.
Internal screw thread sample set includes the labeled data of a large amount of sample image and each sample image, and labeled data can be with
Corresponding sample image is marked with the presence or absence of defect.Further, specific defect type can also be marked out.Preferably, it marks
Note data further include that the pixel region of existing defects further includes lacking that is, in the labeled data of the sample image of existing defects
Sunken pixel region.Specifically, sample image can be collected by above-mentioned image capture module.In actual production process
In, defective internal screw thread sample is limited, and certain means can be taken to obtain enough samples using existing sample
This image.Such as the multiple angles of sample rotates are acquired into the sample image of all angles in sample plane；For some angle
The brightness for repeatedly adjusting light source, acquires the sample image etc. under each light-source brightness.
S12: classifying to the internal screw thread sample set, obtains training set, verifying collection and test set.
Optionally, internal screw thread sample set is divided into training set, verifying collection and test set by the ratio of 6:2:2.Specifically
Classified Proportion can also be adjusted according to actual needs, and details are not described herein.
S13: the training set being input in neural network model and is trained, and obtains initial detecting model.
S14: the initial detecting model is verified using verifying collection, is verified data.
Initial detecting model is verified by verifying collection, is verified data, verify data may include image slices
Plain precision of prediction, IoU (Intersection over Union) etc..
S15: if the verify data meets preset verifying index, the initial detecting is verified using the test set
The generalization ability of model.
Verifying index can be preset according to actual needs, and verify data and preset verifying index are compared
It is right, if verify data meets preset verifying index, the extensive energy of the initial detecting model is verified using the test set
Power.
Optionally, if verify data does not meet preset verifying index, hyper parameter can be adjusted according to verify data, after
Continuous training initial detecting model, and then repeat to verify initial detecting model using verifying collection, until verify data meets
Preset verifying index.
S16: if the generalization ability of the initial detecting model reaches preset benchmark, defects detection model is obtained.
In the present embodiment, by classification to sample set and training verification process, defects detection is better assured that
The stability of model.
In one embodiment, as shown in fig. 7, the described training set is input in neural network model is trained,
Obtain initial detecting model, comprising:
S131: building neural network model.
Specifically, which can be CNN, FCN, U-net or SegNet etc..
S132: the initial network parameter of the neural network model is configured.
Specifically, initial network parameter may include initial weight, learning rate, momentum value, the number of iterations and single treatment
Image number etc..Wherein, single processed image number can match setting according to GPU cache size.
S133: the sample image in training set is input in the neural network model, the neural network model
Each layer network value layer-by-layer back-propagation from the front to the back, the defective locations being calculated under last network output current network state and
Confidence level.
Sample image in training set is input in the neural network model, each layer of convolutional neural networks model is calculated
Network values, the network values of each layer can obtain using propagated forward algorithm, and each layer network value of neural network model is by preceding
Layer-by-layer back-propagation backward, last network export the defective locations being calculated under current network state and confidence level.
S134: the defective locations being calculated and the confidence level are compared with corresponding labeled data, are obtained
Difference must be compared, each layer network successively propagates forward the comparison difference back to front, adjusts neuron connection weight in network
Value.
S135: optimizing each neuron connection weight of network using the sample image training in the training set in batches,
Deconditioning when network state reaches preset reference or reaches trained the number of iterations, obtains initial detecting model.
In the present embodiment, by the above-mentioned training to neural network model, initial detecting model is better assured that
Training effectiveness and precision.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
In one embodiment, a kind of internal screw thread defect detecting device is provided, which includes image
Processing module and image capture module.Internal screw thread defect inspection method corresponds in image processing module and above-described embodiment.
Detailed description are as follows for each functional unit:
Internal screw thread image acquisition unit, the internal screw thread image of the object for obtaining the transmission of described image acquisition module,
The internal screw thread image be described image acquisition module by the component prism and described image acquisition component acquisition include
The image of the complete internal screw thread information of object.
Defect detection unit is obtained for being identified by preset defects detection model to the internal screw thread image
The defects detection information of the object.
Preferably, defect detection unit includes:
Subelement is pre-processed, for carrying out image preprocessing to the internal screw thread image, obtains pretreatment image.
Normalized subelement obtains image to be detected for the pretreatment image to be normalized.
Defects detection subelement is examined for described image to be detected to be input in preset defects detection model
It surveys, obtains the defects detection information of the object.
Preferably, image processing module is also used to then obtain in described when the defects detection information is existing defects
Pixel region where defect described in screw image；According to defect described in the pixel Area generation in the object
In location information.
Preferably, image processing module is also used to obtain internal screw thread sample set, and the internal screw thread sample set includes sample graph
The labeled data of picture and each sample image；Classify to the internal screw thread sample set, obtain training set, verifying collection and
Test set；The training set is input in neural network model and is trained, initial detecting model is obtained；Using the verifying
Collection verifies the initial detecting model, is verified data；If the verify data meets preset verifying index,
The generalization ability of the initial detecting model is verified using the test set；If the generalization ability of the initial detecting model reaches
Preset benchmark then obtains defects detection model.
Preferably, image processing module is also used to construct neural network model；Configure the initial of the neural network model
Network parameter；Sample image in the training set is input in the neural network model, the neural network model
Each layer network value layer-by-layer back-propagation from the front to the back, the defective locations being calculated under last network output current network state and
Confidence level；The defective locations being calculated and the confidence level are compared with corresponding labeled data, are compared
Difference, each layer network successively propagate forward the comparison difference back to front, adjust neuron connection weight in network；In batches
Secondary each neuron connection weight using the sample image training optimization network in the training set, until network state reaches pre-
If condition or deconditioning when reaching trained the number of iterations, obtain initial detecting model.
Specific about image processing module limits the restriction that may refer to above for internal screw thread defect inspection method,
Details are not described herein.Various pieces in above-mentioned image processing module can be come fully or partially through software, hardware and combinations thereof
It realizes.Above-mentioned various pieces can be embedded in the form of hardware or independently of in the processor in computer equipment, can also be with soft
Part form is stored in the memory in computer equipment, executes the corresponding behaviour of the above various pieces in order to which processor calls
Make.
As shown in figure 8, the internal screw thread defect detecting device includes image processing module 1 and image capture module 2, it is described in
Threading defects detection device is used to detect the internal screw thread defect of object 3.Detailed description are as follows for each functional module:
Described image acquisition module 2 is used to acquire the internal screw thread image of the object 3, and the internal screw thread image is sent out
It send to image processing module 1, the internal screw thread image is that described image acquisition module passes through component prism 22 and image acquisition part
The image including the complete internal screw thread information of the object 3 that part 21 acquires.
The internal screw thread image is input to preset for receiving the internal screw thread image by described image processing module
It is identified in defects detection model, obtains the defects detection information of the object.
Preferably, described image acquisition component 21 includes light source, camera lens and camera, the component prism, the light source, institute
Camera lens and the camera is stated to set gradually along the axial direction of the internal screw thread far from the object.Optionally, the light source
It can be annular light source.
The component prism includes bottom surface and side, and the plane perpendicular is in the axial direction side of the internal screw thread of the object
To the side is made of at least three plane, and the radial section of the component prism is by the bottom surface away from the detection
The direction of object is gradually increased.
Component prism can be prism component, four component prisms, pentaprism component or six component prisms etc., specifically may be used
To be selected according to available accuracy.Image acquisition component can be CCD camera.For this sentences six component prisms, lead to
Cross the focal length for reasonably adjusting image acquisition component, the distance and image capture module of six component prisms and image acquisition component
Guarantee to get clearly internal screw thread image with the position of object.Use six component prisms according to component prism, then it is interior
Screw image may include the internal screw thread information of six parts of object, and the internal screw thread information of this six parts constitutes object
Complete internal screw thread information.
Preferably, the object 3 is placed on objective table 4.
In one embodiment, it provides in a kind of internal screw thread defect detecting system, including processing unit and such as above-described embodiment
The internal screw thread defect detecting device.
The object that processing unit is used to detect the presence of defect to the internal screw thread defect detecting device is handled
Specifically, which can be taken out, removes or mark to the object of existing defects.At one
In specific embodiment, which is robot device, detects the presence of the mesh of defect in interior threading defects detection device
When marking object, the object which will be present defect is taken out.
Preferably, which can also include a suggestion device, which is used for inside
When threading defects detection device detects the presence of the object of defect, prompt information is issued.Specifically, which can be
Text, voice or optical signalling.
In the present embodiment, which can carry out the internal screw thread of object with the presence or absence of defect
Intelligent measurement, and the object of existing defects is automatically processed, improve treatment effeciency.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 9.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is for storing the data that internal screw thread defect inspection method uses in above-described embodiment.The computer is set
Standby network interface is used to communicate with external terminal by network connection.To realize when the computer program is executed by processor
A kind of internal screw thread defect inspection method.
In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, processor realize the internal screw thread in above-described embodiment when executing computer program
Defect inspection method.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program is executed by processor the internal screw thread defect inspection method in above-described embodiment.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function
Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different
Functional unit, module are completed, i.e., the internal structure of described device is divided into different functional unit or module, more than completing
The all or part of function of description.
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations；Although referring to aforementioned reality
Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each
Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features；And these are modified
Or replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all
It is included within protection scope of the present invention.
Claims (10)
1. a kind of internal screw thread defect inspection method, which is characterized in that the internal screw thread defect inspection method is lacked applied to internal screw thread
It falls into the image processing module of detection device, the internal screw thread defect detecting device further includes image capture module, described image
Acquisition module includes component prism and image acquisition component；
The internal screw thread defect inspection method includes:
The internal screw thread image for the object that described image acquisition module is sent is obtained, the internal screw thread image is described image acquisition
The figure including the complete internal screw thread information of the object that module is acquired by the component prism and described image acquisition component
Picture；
The internal screw thread image is identified by preset defects detection model, obtains the defects detection letter of the object
Breath.
2. internal screw thread defect inspection method as described in claim 1, which is characterized in that described to pass through preset defects detection mould
Type identifies the internal screw thread image, obtains the defects detection information of the object, comprising:
Image preprocessing is carried out to the internal screw thread image, obtains pretreatment image；
The pretreatment image is normalized, image to be detected is obtained；
Described image to be detected is input in preset defects detection model and is detected, the defect inspection of the object is obtained
Measurement information.
3. internal screw thread defect inspection method as described in claim 1, which is characterized in that pass through preset defects detection described
Model identifies the internal screw thread image, after obtaining the defects detection information of the object, the internal screw thread defect
Detection method further include:
If the defects detection information is existing defects, the pixel area where defect described in the internal screw thread image is obtained
Domain；
According to location information of the defect in the object described in the pixel Area generation.
4. internal screw thread defect inspection method as described in claim 1, which is characterized in that pass through preset defects detection described
Model identifies the internal screw thread image, before obtaining the defects detection information of the object, the internal screw thread defect
Detection method further include:
Internal screw thread sample set is obtained, the internal screw thread sample set includes the mark number of sample image and each sample image
According to；
Classify to the internal screw thread sample set, obtains training set, verifying collection and test set；
The training set is input in neural network model and is trained, initial detecting model is obtained；
The initial detecting model is verified using verifying collection, is verified data；
If the verify data meets preset verifying index, the general of the initial detecting model is verified using the test set
Change ability；
If the generalization ability of the initial detecting model reaches preset benchmark, defects detection model is obtained.
5. internal screw thread defect inspection method as claimed in claim 4, which is characterized in that described that the training set is input to mind
It is trained in network model, obtains initial detecting model, comprising:
Construct neural network model；
Configure the initial network parameter of the neural network model；
Sample image in the training set is input in the neural network model, each layer net of the neural network model
Layer-by-layer back-propagation, last network export the defective locations being calculated under current network state and confidence to network value from the front to the back
Degree；
The defective locations being calculated and the confidence level are compared with corresponding labeled data, obtain contrast differences
Value, each layer network successively propagate forward the comparison difference back to front, adjust neuron connection weight in network；
In batches using each neuron connection weight of the sample image training optimization network in the training set, until network-like
Deconditioning when state reaches preset condition or reaches trained the number of iterations, obtains initial detecting model.
6. a kind of internal screw thread defect detecting device, which is characterized in that the internal screw thread defect detecting device includes image procossing mould
Block and image capture module, the internal screw thread defect detecting device are used to detect the internal screw thread defect of object；
Described image acquisition module is used to acquire the internal screw thread image of the object, and the internal screw thread image is sent to figure
As processing module, the internal screw thread image is the packet that described image acquisition module is acquired by component prism and image acquisition component
Include the image of the complete internal screw thread information of the object；
The internal screw thread image is input to preset defect for receiving the internal screw thread image by described image processing module
It is identified in detection model, obtains the defects detection information of the object.
7. internal screw thread defect detecting device as claimed in claim 6, which is characterized in that described image acquisition component includes light
Source, camera lens and camera, the component prism, the light source, the camera lens and the camera are along far from the object
The axial direction of screw thread is set gradually；
The component prism includes bottom surface and side, and the plane perpendicular is in the axial direction of the internal screw thread of the object, institute
It states side to be made of at least three plane, the radial section of the component prism is by the bottom surface away from the detectable substance
Direction is gradually increased.
8. a kind of internal screw thread defect detecting system, which is characterized in that in including processing unit and as claimed in claims 6 or 7
Threading defects detection device；
The object that the processing unit is used to detect the presence of defect to the internal screw thread defect detecting device is handled.
9. a kind of computer equipment, including memory, processor and storage are in the memory and can be in the processor
The computer program of upper operation, which is characterized in that the processor realized when executing the computer program as claim 1 to
Any one of 5 internal screw thread defect inspection methods.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In realization internal screw thread defects detection side as described in any one of claim 1 to 5 when the computer program is executed by processor
Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910250375.6A CN109919941A (en) | 2019-03-29 | 2019-03-29 | Internal screw thread defect inspection method, device, system, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910250375.6A CN109919941A (en) | 2019-03-29 | 2019-03-29 | Internal screw thread defect inspection method, device, system, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109919941A true CN109919941A (en) | 2019-06-21 |
Family
ID=66967728
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910250375.6A Pending CN109919941A (en) | 2019-03-29 | 2019-03-29 | Internal screw thread defect inspection method, device, system, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109919941A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110405540A (en) * | 2019-07-09 | 2019-11-05 | 华中科技大学 | A kind of artificial intelligence breaking detection system and method |
| CN110866900A (en) * | 2019-11-05 | 2020-03-06 | 江河瑞通（北京）技术有限公司 | Water body color identification method and device |
| CN111127429A (en) * | 2019-12-24 | 2020-05-08 | 魏志康 | Water conservancy system pipe thread defect detection method based on self-training deep neural network |
| CN112330583A (en) * | 2019-07-16 | 2021-02-05 | 青岛智动精工电子有限公司 | Product defect detection method, device, equipment and storage medium |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102968795A (en) * | 2012-12-03 | 2013-03-13 | 哈尔滨工业大学 | Meteor crater mismatching determination method based on ratio of shaded area to external-contour area |
| CN103761529A (en) * | 2013-12-31 | 2014-04-30 | 北京大学 | Open fire detection method and system based on multicolor models and rectangular features |
| CN103760167A (en) * | 2014-01-15 | 2014-04-30 | 唐山英莱科技有限公司 | High-light-reflection-surface coaxial optical path detection system based on refraction imaging |
| CN106295795A (en) * | 2016-08-09 | 2017-01-04 | 衢州学院 | A kind of copper pipe interior whorl forming qualitative forecasting method based on BP neural network algorithm |
| CN106875381A (en) * | 2017-01-17 | 2017-06-20 | 同济大学 | A kind of phone housing defect inspection method based on deep learning |
| CN107102827A (en) * | 2016-02-23 | 2017-08-29 | 爱思打印解决方案有限公司 | The equipment for improving the method for the quality of image object and performing this method |
| CN107356611A (en) * | 2017-07-13 | 2017-11-17 | 上海大学 | A kind of non-contacting interior threaded surface quality detection device of omnirange |
| CN107492125A (en) * | 2017-07-28 | 2017-12-19 | 哈尔滨工业大学深圳研究生院 | The processing method of automobile fish eye lens panoramic view picture |
| CN107707618A (en) * | 2017-08-24 | 2018-02-16 | 广东欧珀移动通信有限公司 | Method and Related product based on position adjustment download |
| CN207936910U (en) * | 2018-03-24 | 2018-10-02 | 福建省石狮市通达电器有限公司 | Image the auxiliary device of measuring instrument acquisition |
| WO2018208791A1 (en) * | 2017-05-08 | 2018-11-15 | Aquifi, Inc. | Systems and methods for inspection and defect detection using 3-d scanning |
| US20180342050A1 (en) * | 2016-04-28 | 2018-11-29 | Yougetitback Limited | System and method for detection of mobile device fault conditions |
| CN108960413A (en) * | 2018-07-11 | 2018-12-07 | 天津工业大学 | A kind of depth convolutional neural networks method applied to screw surface defects detection |
-
2019
- 2019-03-29 CN CN201910250375.6A patent/CN109919941A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102968795A (en) * | 2012-12-03 | 2013-03-13 | 哈尔滨工业大学 | Meteor crater mismatching determination method based on ratio of shaded area to external-contour area |
| CN103761529A (en) * | 2013-12-31 | 2014-04-30 | 北京大学 | Open fire detection method and system based on multicolor models and rectangular features |
| CN103760167A (en) * | 2014-01-15 | 2014-04-30 | 唐山英莱科技有限公司 | High-light-reflection-surface coaxial optical path detection system based on refraction imaging |
| CN107102827A (en) * | 2016-02-23 | 2017-08-29 | 爱思打印解决方案有限公司 | The equipment for improving the method for the quality of image object and performing this method |
| US20180342050A1 (en) * | 2016-04-28 | 2018-11-29 | Yougetitback Limited | System and method for detection of mobile device fault conditions |
| CN106295795A (en) * | 2016-08-09 | 2017-01-04 | 衢州学院 | A kind of copper pipe interior whorl forming qualitative forecasting method based on BP neural network algorithm |
| CN106875381A (en) * | 2017-01-17 | 2017-06-20 | 同济大学 | A kind of phone housing defect inspection method based on deep learning |
| WO2018208791A1 (en) * | 2017-05-08 | 2018-11-15 | Aquifi, Inc. | Systems and methods for inspection and defect detection using 3-d scanning |
| CN107356611A (en) * | 2017-07-13 | 2017-11-17 | 上海大学 | A kind of non-contacting interior threaded surface quality detection device of omnirange |
| CN107492125A (en) * | 2017-07-28 | 2017-12-19 | 哈尔滨工业大学深圳研究生院 | The processing method of automobile fish eye lens panoramic view picture |
| CN107707618A (en) * | 2017-08-24 | 2018-02-16 | 广东欧珀移动通信有限公司 | Method and Related product based on position adjustment download |
| CN207936910U (en) * | 2018-03-24 | 2018-10-02 | 福建省石狮市通达电器有限公司 | Image the auxiliary device of measuring instrument acquisition |
| CN108960413A (en) * | 2018-07-11 | 2018-12-07 | 天津工业大学 | A kind of depth convolutional neural networks method applied to screw surface defects detection |
Non-Patent Citations (3)
| Title |
|---|
| 李梦园: "深度学习算法在表面缺陷识别中的应用研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 陈永清 等: "内螺纹图像识别技术研究", 《工具技术》 * |
| 韦靖博: "基于DSP图像处理的内螺纹检测系统设计", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110405540A (en) * | 2019-07-09 | 2019-11-05 | 华中科技大学 | A kind of artificial intelligence breaking detection system and method |
| CN110405540B (en) * | 2019-07-09 | 2020-07-10 | 华中科技大学 | Artificial intelligence broken cutter detection system and method |
| CN112330583A (en) * | 2019-07-16 | 2021-02-05 | 青岛智动精工电子有限公司 | Product defect detection method, device, equipment and storage medium |
| CN110866900A (en) * | 2019-11-05 | 2020-03-06 | 江河瑞通（北京）技术有限公司 | Water body color identification method and device |
| CN111127429A (en) * | 2019-12-24 | 2020-05-08 | 魏志康 | Water conservancy system pipe thread defect detection method based on self-training deep neural network |
| CN111127429B (en) * | 2019-12-24 | 2023-06-06 | 河北鹰眼智能科技有限公司 | Water conservancy system pipe thread defect detection method based on self-training deep neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109919941A (en) | Internal screw thread defect inspection method, device, system, equipment and medium | |
| US11341626B2 (en) | Method and apparatus for outputting information | |
| CN109325488A (en) | For assisting the method, device and equipment of car damage identification image taking | |
| CN110245662A (en) | Detection model training method, device, computer equipment and storage medium | |
| CN108875963A (en) | Optimization method, device, terminal device and the storage medium of machine learning model | |
| CN108288027A (en) | A kind of detection method of picture quality, device and equipment | |
| CN107690660A (en) | Image-recognizing method and device | |
| WO2020119419A1 (en) | Image recognition-based testing and apparatus, and computer device and storage medium | |
| CN107563265A (en) | A kind of high beam detection method and device | |
| CN110009614A (en) | Method and apparatus for output information | |
| US11694327B2 (en) | Cross layer common-unique analysis for nuisance filtering | |
| US20220383627A1 (en) | Automatic modeling method and device for object detection model | |
| CN109632809A (en) | Product quality detection method and device | |
| CN109829883A (en) | Product quality detection method and device | |
| CN109919035A (en) | Improve method, apparatus, computer equipment and storage medium that attendance is identified by | |
| CN109657799A (en) | A kind of model tuning method and apparatus based on scene adaptation | |
| CN109978815A (en) | Detection system, information processing unit, evaluation method and storage medium | |
| JP2022514580A (en) | Optical correction by machine learning | |
| JP7372017B2 (en) | Steel component learning device, steel component estimation device, steel type determination device, steel component learning method, steel component estimation method, steel type determination method, and program | |
| CN109754077A (en) | Network model compression method, device and the computer equipment of deep neural network | |
| CN105102607A (en) | Image processing device, program, storage medium, and image processing method | |
| CN115272136B (en) | Certificate photo glasses reflection eliminating method, device, medium and equipment based on big data | |
| CN108154199A (en) | A kind of quick single class object detection method of high-precision based on deep learning | |
| CN111178275A (en) | Fire detection method based on convolutional neural network | |
| CN110705398A (en) | Mobile-end-oriented test paper layout image-text real-time detection 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190621 |