CN107607546A - Leather defect inspection method, system and device based on photometric stereo vision - Google Patents
Leather defect inspection method, system and device based on photometric stereo vision Download PDFInfo
- Publication number
- CN107607546A CN107607546A CN201710847126.6A CN201710847126A CN107607546A CN 107607546 A CN107607546 A CN 107607546A CN 201710847126 A CN201710847126 A CN 201710847126A CN 107607546 A CN107607546 A CN 107607546A
- Authority
- CN
- China
- Prior art keywords
- mrow
- leather
- image
- mtd
- gray scale
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
The invention discloses leather defect inspection method, system and the device based on photometric stereo vision, the system includes the first acquisition module, first processing module, Second processing module, the 3rd processing module, fourth processing module and the 5th processing module.This method includes the different light source effect in collection position and shoots obtained multiple first images down；After multiple first images are synthesized into depth image, curvature filtering, gray scale stretching, binary conversion treatment are carried out successively, the binary map of leather to be measured is obtained, binary map is handled using SVM, export leather defects detection result.The device includes memory and the processor for performing above-mentioned detection method.By using the present invention, it is capable of the enhancing effect of fast lifting leather defect area, improves the precision and treatment effeciency of leather defects detection.The present invention can be widely applied in leather detection field as leather defect inspection method, system and device based on photometric stereo vision.
Description
Technical field
The present invention relates to defect detecting technique, more particularly to a kind of leather defects detection side based on photometric stereo vision
Method, system and device.
Background technology
Technology word is explained：
Photometric Stereo：Photometric stereo.
SVM：Refer to SVMs.
In suitcase, automotive seat, clothes isocortex process of producing product, quality monitoring is particularly important, and therefore, leather is made
For the raw material of these cortex products, its defects detection and the important process that positioning is exactly quality monitoring.At present, in industry for
The detection of the defects of leather relies primarily on human eye, however, the noisy environment of plant, substantial amounts of detection work and leather defect kind
It is numerous and diverse, these all simple to be difficult to meet Quality Control demand by the detection mode of human eye.Therefore in order to solve these problems,
The defects of there has been proposed using machine vision technique to realize leather, is detected.
The machine vision technique has the advantages of improving production flexibility and automaticity, therefore, utilizes machine vision
The defects of technology is to realize leather is detected, and can apply and not be suitable for the dangerous work environment of manual work at some or manually regard
Feel the occasion for being difficult to meet Quality Control requirement, applicability light；Simultaneously in high-volume industrial processes, machine vision technique is utilized
The defects of to realize leather, is detected, and can greatly improve production efficiency and the automaticity of production, and can improve detection
Precision.However, for the currently used leather defect detecting technique based on machine vision, it still has some shortcomings, example
Such as defect recognition accuracy of detection is low, the low shortcoming of operation treatment effeciency.
The content of the invention
In order to solve the above-mentioned technical problem, the first object of the present invention is to provide a kind of leather based on photometric stereo vision
Defect inspection method, the defects of can improving leather defect detecting technique, detect precision and treatment effeciency.
The second object of the present invention is to provide a kind of leather defect detecting system based on photometric stereo vision, can improve skin
The defects of removing from office defect detecting technique detects precision and treatment effeciency.
The third object of the present invention is to provide a kind of leather defect detecting device based on photometric stereo vision, can improve skin
The defects of removing from office defect detecting technique detects precision and treatment effeciency.
First technical scheme of the present invention is：Leather defect inspection method based on photometric stereo vision, the party
Method comprises the following steps：
Captured obtained multiple first images under the different light source radiation in position are gathered, wherein, described first
Image refers to the image of leather to be measured；
Using photometric stereo vision algorithm, multiple first images collected are synthesized to obtain the second image；
After carrying out curvature filtering to the second image, curvature chart is obtained；
After carrying out gray scale stretching processing to curvature chart, gray scale stretching image is obtained；
After carrying out image binaryzation processing to gray scale stretching image, the binary map of leather to be measured is obtained；
The binary map of leather to be measured is handled using SVMs, exports the testing result of leather to be measured.
Further, the collection under the different light source radiation in position captured obtained multiple first images this
Step, it comprises the following steps：
Fixed camera position, camera is made to be mutually perpendicular to leather detection platform；
Multiple light sources are arranged on the diverse location above leather detection platform, and make the incident ray and leather of light source
Angle between detection platform is 30 °~60 °；
Multiple light sources are controlled to produce different illumination conditions；
Control camera shoots first image under an illumination condition；
The first image obtained to shooting is acquired.
Further, it is described to use photometric stereo vision algorithm, multiple first images collected are synthesized to obtain second
The step for image, it comprises the following steps：
According to multiple first images collected, the surface normal of leather to be measured is calculated；
According to the surface normal being calculated, the orthogonal vectors corresponding to surface normal are calculated；
According to the orthogonal vectors being calculated, generation draws depth image.
Further, it is described curvature filtering is carried out to the second image after, the step for obtaining curvature chart, it is specially：Using
Gaussian curvature wave filter comes after carrying out curvature filtering to the second image, obtains curvature chart.
Further, the expression formula of the Gaussian curvature wave filter is as follows：
Wherein, U is expressed as the second image of input, and H is expressed as the curvature chart of output, and * is expressed as image convolution operation.
Further, described the step for carrying out gray scale stretching processing to curvature chart, it comprises the following steps：
Obtain the minimum gradation value and maximum gradation value of curvature chart；
Using minimum gradation value and maximum gradation value, gray scale stretching is carried out to the gray value of the pixel in curvature chart.
Further, it is described to utilize minimum gradation value and maximum gradation value, the gray value of the pixel in curvature chart is carried out
The step for gray scale stretching, its used gray scale stretching formula are as follows：
Wherein, Y is expressed as the gray value after gray scale stretching, and X is expressed as the gray value before gray scale stretching, XminIt is expressed as most
Small gray value, XmaxIt is expressed as maximum gradation value.
Further, described the step for image binaryzation processing is carried out to gray scale stretching image, its used binaryzation
Conversion formula is as follows：
Threshold=E (S (x)) × 3
Wherein, O (x) is expressed as the binary map of output, and S (x) is expressed as the pixel value of pixel in gray scale stretching image,
Threshold is expressed as threshold value, and E (S (x)) is expressed as the desired value of gray scale stretching image.
Second technical scheme of the present invention is：Leather defect detecting system based on photometric stereo vision, this is
System includes：
First acquisition module, for gathering captured obtained multiple first figures under the different light source radiation in position
Picture, wherein, described first image refers to the image of leather to be measured；
First processing module, for using photometric stereo vision algorithm, multiple first images collected are synthesized
To the second image；
Second processing module, after carrying out curvature filtering to the second image, obtain curvature chart；
3rd processing module, after carrying out gray scale stretching processing to curvature chart, obtain gray scale stretching image；
Fourth processing module, after carrying out image binaryzation processing to gray scale stretching image, obtain the two of leather to be measured
Value figure；
5th processing module, for being handled using SVMs the binary map of leather to be measured, export skin to be measured
The testing result of leather.
3rd technical scheme of the present invention is：Leather defect detecting device based on photometric stereo vision, the dress
Put including：
Memory, for storing at least one program；
Processor, for loading at least one program and performing the above-mentioned leather defect inspection based on photometric stereo vision
Survey method and step.
The beneficial effect of the inventive method, system and device is：The present invention adopts under the light source radiation of diverse location
Collect multiple images of leather to be measured, then using photometric stereo vision algorithm, multiple images collected are synthesized to obtain one
The second image with certain three-dimensional information is opened, then, carries out curvature filtering process, gray scale stretching processing successively to the second image
And after binary conversion treatment, obtain the binary map of leather to be measured, finally using the SVMs trained to leather to be measured two
Value figure is handled, the defects of exporting leather to be measured testing result, it can be seen that, can be quick by using the technology of the present invention
The enhancing effect of leather defect area in Leather Image to be measured is lifted, so as to greatly improve the precision of leather defects detection and work
Efficiency is dealt with, so not only alleviates the workload of staff, and the production quality and effect of leather finish can also be improved
Rate.
Brief description of the drawings
Fig. 1 is a kind of step flow chart of the leather defect inspection method based on photometric stereo vision of the present invention；
Fig. 2 is a kind of structured flowchart of the leather defect detecting system based on photometric stereo vision of the present invention；
Fig. 3 is the specific reality that SVM is built in a kind of leather defect inspection method based on photometric stereo vision of the present invention
Apply a flow chart of steps；
Fig. 4 is the structural representation of photometric stereo experiment porch；
Fig. 5 is the image comparison schematic diagram of Leather Image to be measured under different illumination conditions, composite diagram and testing result.
Embodiment
Embodiment 1
As shown in figure 1, the leather defect inspection method based on photometric stereo vision, this method comprise the following steps：
Captured obtained multiple first images under the different light source radiation in position are gathered, wherein, described first
Image refers to the image of leather to be measured；
Using photometric stereo vision algorithm, multiple first images collected are synthesized to obtain the second image；
After carrying out curvature filtering to the second image, curvature chart is obtained；
After carrying out gray scale stretching processing to curvature chart, gray scale stretching image is obtained；
After carrying out image binaryzation processing to gray scale stretching image, the binary map of leather to be measured is obtained；
The binary map of leather to be measured is handled using SVMs, exports the testing result of leather to be measured.
Preferred embodiment is further used as, the collection is captured under the different light source radiation in position to be obtained
Multiple first images the step for, it comprises the following steps：
Fixed camera position, camera is made to be mutually perpendicular to leather detection platform；
Multiple light sources are arranged on the diverse location above leather detection platform, and make the incident ray and leather of light source
Angle between detection platform is 30 °~60 °；
Multiple light sources are controlled to produce different illumination conditions；
Control camera shoots first image under an illumination condition；
The first image obtained to shooting is acquired.Preferably, the quantity of the first image and the quantity of light source are identical, light
The quantity in source is 4, and/or the model of multiple light sources is identical.
It is further used as preferred embodiment, described to use photometric stereo vision algorithm, will collect multiple the
The step for one image synthesizes to obtain the second image, it comprises the following steps：
According to multiple first images collected, the surface normal of leather to be measured is calculated；
According to the surface normal being calculated, the orthogonal vectors corresponding to surface normal are calculated；
According to the orthogonal vectors being calculated, generation draws depth image.
Be further used as preferred embodiment, it is described curvature filtering is carried out to the second image after, obtain curvature chart this
Step, it is specially：
After carrying out curvature filtering to the second image using Gaussian curvature wave filter, curvature chart is obtained.
Preferred embodiment is further used as, the expression formula of the Gaussian curvature wave filter is as follows：
Wherein, U is expressed as the second image of input, and H is expressed as the curvature chart of output, and * is expressed as image convolution operation.
Preferred embodiment is further used as, described the step for carrying out gray scale stretching processing to curvature chart, it includes
Following steps：
Obtain the minimum gradation value and maximum gradation value of curvature chart；
Using minimum gradation value and maximum gradation value, gray scale stretching is carried out to the gray value of the pixel in curvature chart.
Preferred embodiment is further used as, it is described to utilize minimum gradation value and maximum gradation value, in curvature chart
The gray value of pixel carries out the step for gray scale stretching, and its used gray scale stretching formula is as follows：
Wherein, Y is expressed as the gray value after gray scale stretching, and X is expressed as the gray value before gray scale stretching, XminIt is expressed as most
Small gray value, XmaxIt is expressed as maximum gradation value.
Preferred embodiment is further used as, it is described that this step is handled to gray scale stretching image progress image binaryzation
Suddenly, its used binaryzation conversion formula is as follows：
Threshold=E (S (x)) × 3
Wherein, O (x) is expressed as the binary map of output, and S (x) is expressed as the pixel value of pixel in gray scale stretching image,
Threshold is expressed as threshold value, and E (S (x)) is expressed as the desired value of gray scale stretching image.
Embodiment 2
System corresponding with the above method, as shown in Fig. 2 the leather defect detecting system based on photometric stereo vision, should
System includes：
First acquisition module, for gathering captured obtained multiple first figures under the different light source radiation in position
Picture, wherein, described first image refers to the image of leather to be measured；
First processing module, for using photometric stereo vision algorithm, multiple first images collected are synthesized
To the second image；
Second processing module, after carrying out curvature filtering to the second image, obtain curvature chart；
3rd processing module, after carrying out gray scale stretching processing to curvature chart, obtain gray scale stretching image；
Fourth processing module, after carrying out image binaryzation processing to gray scale stretching image, obtain the two of leather to be measured
Value figure；
5th processing module, for being handled using SVMs the binary map of leather to be measured, export skin to be measured
The testing result of leather.For the first acquisition module and the first to the 5th processing module, they can be program module, or hardware
Circuit, or the apparatus of software and hardware combining.
Preferred embodiment is further used as, first acquisition module includes：
First control submodule, for fixed camera position, camera is made to be mutually perpendicular to leather detection platform；
Second control submodule, for the diverse location being arranged on multiple light sources above leather detection platform, and make
Angle between the incident ray and leather detection platform of light source is 30 °~60 °；
3rd control submodule, for controlling multiple light sources to produce different illumination conditions；
4th control submodule, for controlling camera to shoot first image under an illumination condition；
First collection submodule, the first image for being obtained to shooting are acquired.
Preferred embodiment is further used as, the first processing module includes：
First calculating sub module, for the surface of leather to be measured according to multiple first images for collecting, to be calculated
Normal vector；
Second calculating sub module, for according to the surface normal being calculated, being calculated corresponding to surface normal
Orthogonal vectors；
First generation submodule, for drawing depth image according to the orthogonal vectors being calculated, generation.
Be further used as preferred embodiment, the Second processing module be specifically used for using Gaussian curvature wave filter come
After carrying out curvature filtering to the second image, curvature chart is obtained.Preferably, the table of the Gaussian curvature wave filter described in this embodiment
It is identical with the expression formula of the Gaussian curvature wave filter described in embodiment 1 up to formula.
Preferred embodiment is further used as, the 3rd processing module includes：
3rd calculating sub module, for obtaining the minimum gradation value and maximum gradation value of curvature chart；
First stretching submodule, for utilizing minimum gradation value and maximum gradation value, to the ash of the pixel in curvature chart
Angle value carries out gray scale stretching.
Preferred embodiment is further used as, described first stretches the gray scale stretching formula and reality employed in submodule
The gray scale stretching formula applied employed in example 1 is identical.
It is further used as preferred embodiment, binaryzation conversion formula and reality employed in the fourth processing module
The binaryzation conversion formula applied employed in example 1 is identical.
Embodiment 3
Device corresponding with the above method, the leather defect detecting device based on photometric stereo vision, the device include：
Memory, for storing at least one program；
Processor, for load at least one program and perform described in above-described embodiment 1 based on photometric stereo
The leather defect inspection method step of vision.
Embodiment 4
The present invention is further elaborated with reference to this preferred embodiment.In the present embodiment, the figure of the first image/the 3rd
The quantity of picture and the quantity of light source are preferably 4.
A kind of specific implementation step of the leather defect inspection method based on photometric stereo vision of the present invention is as follows.
Step 1：SVM needed for structure.As shown in figure 3, the sub-step that step 1 specifically includes is as follows.
S101, collection captured 4 obtained the 3rd images under the different light source radiation in position；Described the 3rd
Image refers to the image of leather defects detection sample；
For step S101, it comprises the following steps：
S1011, fixed camera position, make camera be mutually perpendicular to leather detection platform；
S1012,4 light sources are arranged on the diverse location above leather detection platform, and make the incident ray of light source
Angle between leather detection platform is 30 °~60 °；
4 S1013, control light sources produce different illumination conditions；
S1014, control camera shoot the 3rd image under an illumination condition；
S1015, the 3rd image obtained to shooting are acquired, and adopt to obtain 4 the 3rd images altogether；
It can be seen that because the quantity of the 3rd image is 4, therefore, 4 light sources should be controlled to produce 4 different illumination conditions；And
For above-mentioned camera and 4 light sources, they constitute photometric stereo experiment porch, specific as shown in Figure 4；In addition, such as Fig. 4 institutes
Show, the angle between the incident ray and leather detection platform of light source is preferably 40 °；
S102, using photometric stereo vision algorithm, by collect 4 the 3rd images synthesize to obtain one have it is certain
4th image of three-dimensional information；
For step S102, it comprises the following steps：
S1021, according to 4 the 3rd images collecting, calculate leather defects detection sample surface normal；
First, the light intensity of any pixel in surface is made to be expressed as：I=kdNL, wherein, kdFor constant, N is unknown number (institute
The surface normal of demand solution), L is the location parameter of light source, and therefore, the pixel of 4 the 3rd images is embodied as respectively：
I1=kd(Nx*L1x+Ny*L1y+Nz*L1y)
I2=kd(Nx*L2x+Ny*L2y+Nz*L2y)
I3=kd(Nx*L3x+Ny*L3y+Nz*L3y)
I4=kd(Nx*LOx+Ny*L4y+Nz*L4y)
Then, [G]=k is maded[N], then it is N=G/k that above-mentioned equation group is rewritabled；
Finally, N=G/k is utilizeddSurface normal N is drawn to solve；
S1022, the surface normal being calculated according to step S1021, are calculated corresponding to the surface normal
Orthogonal vectors；
It can be seen from orthogonality relation, NV=0, therefore, using the surface normal N being calculated, to solve
Go out surface normal N orthogonal vectors V；
The orthogonal vectors that S1023, basis are calculated, generation draw depth image；Wherein, generated in step S1023
Depth image be the 4th image；
S103, after carrying out curvature filtering to the 4th image using Gaussian curvature wave filter, obtain first curvature figure；Its
In, the expression formula of the Gaussian curvature wave filter is as follows：
In above formula, U is expressed as the image of input, and H is expressed as the curvature chart of output, and * is expressed as image convolution operation；Due to
Curvature filtering is carried out to the 4th image as composite diagram using this Gaussian curvature wave filter, therefore the meter of curvature can be avoided
The geometry flow with complexity is calculated, and importantly, the construction of Gaussian curvature wave filter is very simple, calculating speed is fast, and
In the present invention, it can also lift the enhancing effect of leather defect area on the basis for the treatment of effeciency is improved, and make follow-up skin
The precision for removing from office defects detection result is higher；
S104, after carrying out gray scale stretching processing to first curvature figure, obtain the first gray scale stretching image；
For step S104, it includes：
Gray value corresponding to each pixel is X in S1041, current first curvature figure, obtains the minimum ash of first curvature figure
Angle value XminWith maximum gradation value Xmax；
Specifically, gray value corresponding to each pixel is X in current first curvature figure, travels through all pictures in first curvature figure
The gray value of vegetarian refreshments, obtain the minimum gradation value X of first curvature figureminWith maximum gradation value Xmax, i.e., it is corresponding in image pixel value
Minimum value and maximum；
S1042, utilize the minimum gradation value X drawnminWith maximum gradation value Xmax, to the pixel in first curvature figure
Gray value carries out gray scale stretching；
Specifically, gray scale stretching is that the gray value after 0-255 this tonal range is Y, and Y calculation formula is：
S105, after carrying out image binaryzation processing to the first gray scale stretching image, obtain the two of leather defects detection sample
Value figure, i.e. the second binary map；Binaryzation conversion formula is as follows used by this step：
Threshold=E (S (x)) × 3
Wherein, O (x) is expressed as the binary map of output, and S (x) is expressed as the pixel value of pixel in gray scale stretching image,
Threshold is expressed as threshold value, and E (S (x)) is expressed as the desired value of gray scale stretching image, i.e., view picture gray scale stretching image is flat
Equal gray value；
S106, using the second binary map and the characteristics of image extracted from the second binary map SVM is instructed
Practice；The characteristics of image extracted from the second binary map includes but is not limited to have the spies such as shape, edge, size
Sign；
For step S106, it includes：
S1061, using the second binary map as training input data；
S1062, using the characteristics of image extracted from the second binary map as training output data；
S1063, using training input data and training output data to be trained SVM, obtained SVM after training terminates
For required SVM.
Step 2：Defects detection processing is carried out to leather to be measured using the SVM trained, to realize determining for leather defect
Position, detection, identification.
S201, collection captured 4 obtained the first images under the different light source radiation in position；First figure
Image as referring to leather to be measured；
For step S201, it comprises the following steps：
S2011, fixed camera position, make camera be mutually perpendicular to leather detection platform；
S2012,4 light sources are arranged on the diverse location above leather detection platform, and make the incident ray of light source
Angle between leather detection platform is 30 °~60 °, wherein, the angle is preferably 40 °；
4 S2013, control light sources produce different illumination conditions；
S2014, control camera shoot first image under an illumination condition；
S2015, the first image obtained to shooting are acquired, and adopt to obtain 4 the first images altogether；
It can be seen that because the quantity of the first image is 4, therefore, 4 light sources should be controlled to produce 4 different illumination conditions；
S202, using photometric stereo vision algorithm, by collect 4 the first images synthesize to obtain one have it is certain
Second image of three-dimensional information；
For step S202, it comprises the following steps：
S2021, according to 4 the first images collecting, calculate leather to be measured surface normal；
S2022, the surface normal being calculated according to step S2021, are calculated corresponding to the surface normal
Orthogonal vectors；
S2023, the orthogonal vectors being calculated according to step S2022, generation draw depth image；Wherein, step S2023
Generated in depth image be the second image；
Implementation is specific as shown in step S1021 and S1022 content used by above-mentioned steps S2021 and S2022；
S203, after carrying out curvature filtering to the second image using the Gaussian curvature wave filter described in step S103, obtain
To torsion figure；
S204, after carrying out gray scale stretching processing to torsion figure, obtain the second gray scale stretching image；
For step S204, it includes：
Gray value corresponding to each pixel is X in S2041, current torsion figure, obtains the minimum ash of torsion figure
Angle value XminWith maximum gradation value Xmax；
Specifically, gray value corresponding to each pixel is X in current torsion figure, travels through all pictures in torsion figure
The gray value of vegetarian refreshments, obtain the minimum gradation value X of torsion figureminWith maximum gradation value Xmax, i.e., it is corresponding in image pixel value
Minimum value and maximum；
S2042, utilize the minimum gradation value X drawnminWith maximum gradation value Xmax, using the meter of Y in above-mentioned steps S1042
Formula is calculated, gray scale stretching is carried out to the gray value of the pixel in torsion figure, so as to obtain the second gray scale stretching image；
S205, after carrying out image binaryzation processing to the second gray scale stretching image, obtain the binary map of leather to be measured, i.e., the
One binary map；Binaryzation conversion formula is as shown in the formula in above-mentioned step S105 used by this step；
S206, using the SVM trained in step 1 come the binary map to leather to be measured, i.e. the first binary map, located
Reason, testing result the defects of so as to export leather to be measured.Treated as shown in figure 5, the left side four is opened for what is inputted under different light source positions
The image (i.e. the first image) of leather is surveyed, the top on the right is composite diagram (i.e. the second image), and the lower section on the right is testing result
Schematic diagram.
Obtained by above-mentioned, the photometric stereo visual experiment platform employed in the present invention, acted in the light source of diverse location
Under, multiple images are gathered, a picture with certain three-dimensional information are synthesized in the presence of photometric stereo vision technique, so
The composite diagram has apparent enhancing effect, is advantageous to improve the precision of follow-up leather defects detection；And curvature filters
It is fabulous for the filter effect in the region of surface irregularity, therefore, filtered, can be being improved using curvature on the basis of composite diagram
The enhancing effect of leather defect area can also be lifted on the basis for the treatment of effeciency, makes the accurate of follow-up leather defects detection result
Degree is further enhanced.
The content of the present embodiment is suitable for above-described embodiment 1~3.
Above is the preferable implementation to the present invention is illustrated, but the invention is not limited to the implementation
Example, those skilled in the art can also make a variety of equivalent variations on the premise of without prejudice to spirit of the invention or replace
Change, these equivalent deformations or replacement are all contained in the application claim limited range.
Claims (10)
1. the leather defect inspection method based on photometric stereo vision, it is characterised in that：This method comprises the following steps：
Captured obtained multiple first images under the different light source radiation in position are gathered, wherein, described first image
Refer to the image of leather to be measured；
Using photometric stereo vision algorithm, multiple first images collected are synthesized to obtain the second image；
After carrying out curvature filtering to the second image, curvature chart is obtained；
After carrying out gray scale stretching processing to curvature chart, gray scale stretching image is obtained；
After carrying out image binaryzation processing to gray scale stretching image, the binary map of leather to be measured is obtained；
The binary map of leather to be measured is handled using SVMs, exports the testing result of leather to be measured.
2. the leather defect inspection method based on photometric stereo vision according to claim 1, it is characterised in that：The collection
The step for multiple first images obtained captured by under the different light source radiation in position, it comprises the following steps：
Fixed camera position, camera is made to be mutually perpendicular to leather detection platform；
Multiple light sources are arranged on the diverse location above leather detection platform, and make the incident ray of light source and leather detect
Angle between platform is 30 °~60 °；
Multiple light sources are controlled to produce different illumination conditions；
Control camera shoots first image under an illumination condition；
The first image obtained to shooting is acquired.
3. the leather defect inspection method according to claim 1 or claim 2 based on photometric stereo vision, it is characterised in that：It is described
Using photometric stereo vision algorithm, the step for multiple first images collected are synthesized to obtain the second image, it includes
Following steps：
According to multiple first images collected, the surface normal of leather to be measured is calculated；
According to the surface normal being calculated, the orthogonal vectors corresponding to surface normal are calculated；
According to the orthogonal vectors being calculated, generation draws depth image.
4. the leather defect inspection method according to claim 1 or claim 2 based on photometric stereo vision, it is characterised in that：It is described
After carrying out curvature filtering to the second image, the step for obtaining curvature chart, it is specially：
After carrying out curvature filtering to the second image using Gaussian curvature wave filter, curvature chart is obtained.
5. the leather defect inspection method based on photometric stereo vision according to claim 4, it is characterised in that：The Gauss
The expression formula of curvature wave filter is as follows：
<mrow>
<mi>H</mi>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mrow>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>16</mn>
</mfrac>
</mrow>
</mtd>
<mtd>
<mfrac>
<mn>5</mn>
<mn>16</mn>
</mfrac>
</mtd>
<mtd>
<mrow>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>16</mn>
</mfrac>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mfrac>
<mn>5</mn>
<mn>16</mn>
</mfrac>
</mtd>
<mtd>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</mtd>
<mtd>
<mfrac>
<mn>5</mn>
<mn>16</mn>
</mfrac>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>16</mn>
</mfrac>
</mrow>
</mtd>
<mtd>
<mfrac>
<mn>5</mn>
<mn>16</mn>
</mfrac>
</mtd>
<mtd>
<mrow>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>16</mn>
</mfrac>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>*</mo>
<mi>U</mi>
</mrow>
Wherein, U is expressed as the second image of input, and H is expressed as the curvature chart of output, and * is expressed as image convolution operation.
6. the leather defect inspection method according to claim 1 or claim 2 based on photometric stereo vision, it is characterised in that：It is described
The step for carrying out gray scale stretching processing to curvature chart, it comprises the following steps：
Obtain the minimum gradation value and maximum gradation value of curvature chart；
Using minimum gradation value and maximum gradation value, gray scale stretching is carried out to the gray value of the pixel in curvature chart.
7. the leather defect inspection method based on photometric stereo vision according to claim 6, it is characterised in that：The utilization
Minimum gradation value and maximum gradation value, the step for carrying out gray scale stretching to the gray value of the pixel in curvature chart, it is adopted
Gray scale stretching formula is as follows：
<mrow>
<mi>Y</mi>
<mo>=</mo>
<mfrac>
<mrow>
<mn>255</mn>
<mo>&times;</mo>
<mrow>
<mo>(</mo>
<mi>X</mi>
<mo>-</mo>
<msub>
<mi>X</mi>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mi>X</mi>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>X</mi>
<mi>min</mi>
</msub>
</mrow>
</mfrac>
</mrow>
Wherein, Y is expressed as the gray value after gray scale stretching, and X is expressed as the gray value before gray scale stretching, XminIt is expressed as minimum ash
Angle value, XmaxIt is expressed as maximum gradation value.
8. the leather defect inspection method according to claim 1 or claim 2 based on photometric stereo vision, it is characterised in that：It is described
The step for carrying out image binaryzation processing to gray scale stretching image, its used binaryzation conversion formula is as follows：
<mrow>
<mi>O</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mn>1</mn>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>></mo>
<mi>t</mi>
<mi>h</mi>
<mi>r</mi>
<mi>e</mi>
<mi>s</mi>
<mi>h</mi>
<mi>o</mi>
<mi>l</mi>
<mi>d</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>0</mn>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>o</mi>
<mi>t</mi>
<mi>h</mi>
<mi>e</mi>
<mi>r</mi>
<mi>w</mi>
<mi>i</mi>
<mi>s</mi>
<mi>e</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Threshold=E (S (x)) × 3
Wherein, O (x) is expressed as the binary map of output, and S (x) is expressed as the pixel value of pixel in gray scale stretching image,
Threshold is expressed as threshold value, and E (S (x)) is expressed as the desired value of gray scale stretching image.
9. the leather defect detecting system based on photometric stereo vision, it is characterised in that：The system includes：
First acquisition module, for gathering captured obtained multiple first images under the different light source radiation in position,
Wherein, described first image refers to the image of leather to be measured；
First processing module, for using photometric stereo vision algorithm, multiple first images collected are synthesized to obtain the
Two images；
Second processing module, after carrying out curvature filtering to the second image, obtain curvature chart；
3rd processing module, after carrying out gray scale stretching processing to curvature chart, obtain gray scale stretching image；
Fourth processing module, after carrying out image binaryzation processing to gray scale stretching image, obtain the binary map of leather to be measured；
5th processing module, for being handled using SVMs the binary map of leather to be measured, export leather to be measured
Testing result.
10. the leather defect detecting device based on photometric stereo vision, it is characterised in that：The device includes：
Memory, for storing at least one program；
Processor, for load at least one program and perform claim require described in any one of 1-8 based on photometric stereo
The leather defect inspection method of vision.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710847126.6A CN107607546B (en) | 2017-09-19 | 2017-09-19 | Leather defect detection method, system and device based on photometric stereo vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710847126.6A CN107607546B (en) | 2017-09-19 | 2017-09-19 | Leather defect detection method, system and device based on photometric stereo vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107607546A true CN107607546A (en) | 2018-01-19 |
| CN107607546B CN107607546B (en) | 2020-10-23 |
Family
ID=61061215
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710847126.6A Active CN107607546B (en) | 2017-09-19 | 2017-09-19 | Leather defect detection method, system and device based on photometric stereo vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107607546B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647554A (en) * | 2018-05-15 | 2018-10-12 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of sport footwear air cushion stereoscopic two-dimensional code identification and detection device and method |
| CN109523541A (en) * | 2018-11-23 | 2019-03-26 | 五邑大学 | A kind of metal surface fine defects detection method of view-based access control model |
| CN109829886A (en) * | 2018-12-25 | 2019-05-31 | 苏州江奥光电科技有限公司 | A kind of pcb board defect inspection method based on depth information |
| CN109932367A (en) * | 2019-03-14 | 2019-06-25 | 佛山缔乐视觉科技有限公司 | A kind of curved surface carved image acquisition device |
| CN110298835A (en) * | 2019-07-02 | 2019-10-01 | 广东工业大学 | A kind of leather surface damage testing method, system and relevant apparatus |
| CN110715939A (en) * | 2018-07-12 | 2020-01-21 | 卓峰智慧生态有限公司 | Artificial intelligence-based leather detection method and leather product production method |
| CN110987954A (en) * | 2019-12-30 | 2020-04-10 | 江南大学 | Method and system for eliminating leather surface defect detection blind area |
| CN111624206A (en) * | 2020-07-03 | 2020-09-04 | 东北大学 | Metal surface defect detection method based on linear array camera stereoscopic vision |
| WO2020186623A1 (en) * | 2019-03-18 | 2020-09-24 | 香港纺织及成衣研发中心有限公司 | Method, apparatus and device for processing fur |
| CN113092489A (en) * | 2021-05-20 | 2021-07-09 | 鲸朵(上海)智能科技有限公司 | System and method for detecting appearance defects of battery |
| CN113155852A (en) * | 2021-04-08 | 2021-07-23 | 煤炭科学研究总院 | Transmission band detection method and device and electronic equipment |
| CN113252567A (en) * | 2021-06-08 | 2021-08-13 | 菲特（天津）检测技术有限公司 | Method, system, medium and terminal for rapidly detecting multiple defects on surface of aluminum valve plate |
| CN113781424A (en) * | 2021-09-03 | 2021-12-10 | 苏州凌云光工业智能技术有限公司 | Surface defect detection method, device and equipment |
| CN116559181A (en) * | 2023-07-07 | 2023-08-08 | 杭州灵西机器人智能科技有限公司 | Defect detection method, system, device and medium based on luminosity stereoscopic vision |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102951151A (en) * | 2011-08-24 | 2013-03-06 | 现代摩比斯株式会社 | Lane maintaining auxiliary system for vehicles and method thereof |
| CN103592307A (en) * | 2012-08-17 | 2014-02-19 | 索尼公司 | 3d measuring device, 3d measuring method, program and method of manufacturing substrate |
| CN105787989A (en) * | 2016-03-18 | 2016-07-20 | 山东大学 | Measurement texture geometric feature reconstruction method based on photometric stereo |
-
2017
- 2017-09-19 CN CN201710847126.6A patent/CN107607546B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102951151A (en) * | 2011-08-24 | 2013-03-06 | 现代摩比斯株式会社 | Lane maintaining auxiliary system for vehicles and method thereof |
| CN103592307A (en) * | 2012-08-17 | 2014-02-19 | 索尼公司 | 3d measuring device, 3d measuring method, program and method of manufacturing substrate |
| CN105787989A (en) * | 2016-03-18 | 2016-07-20 | 山东大学 | Measurement texture geometric feature reconstruction method based on photometric stereo |
Non-Patent Citations (2)
| Title |
|---|
| 杨晓波等: "基于光度立体视觉的起皱织物表面形态重建研究", 《东华大学学报（自然科学版）》 * |
| 陆庆庆: "基于人脸图像的性别分类", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647554A (en) * | 2018-05-15 | 2018-10-12 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of sport footwear air cushion stereoscopic two-dimensional code identification and detection device and method |
| CN110715939A (en) * | 2018-07-12 | 2020-01-21 | 卓峰智慧生态有限公司 | Artificial intelligence-based leather detection method and leather product production method |
| CN109523541A (en) * | 2018-11-23 | 2019-03-26 | 五邑大学 | A kind of metal surface fine defects detection method of view-based access control model |
| CN109829886A (en) * | 2018-12-25 | 2019-05-31 | 苏州江奥光电科技有限公司 | A kind of pcb board defect inspection method based on depth information |
| CN109932367A (en) * | 2019-03-14 | 2019-06-25 | 佛山缔乐视觉科技有限公司 | A kind of curved surface carved image acquisition device |
| WO2020186623A1 (en) * | 2019-03-18 | 2020-09-24 | 香港纺织及成衣研发中心有限公司 | Method, apparatus and device for processing fur |
| CN110298835A (en) * | 2019-07-02 | 2019-10-01 | 广东工业大学 | A kind of leather surface damage testing method, system and relevant apparatus |
| CN110298835B (en) * | 2019-07-02 | 2023-07-18 | 广东工业大学 | Leather surface damage detection method, system and related device |
| CN110987954A (en) * | 2019-12-30 | 2020-04-10 | 江南大学 | Method and system for eliminating leather surface defect detection blind area |
| CN110987954B (en) * | 2019-12-30 | 2021-10-22 | 江南大学 | Method and system for eliminating leather surface defect detection blind area |
| CN111624206A (en) * | 2020-07-03 | 2020-09-04 | 东北大学 | Metal surface defect detection method based on linear array camera stereoscopic vision |
| CN111624206B (en) * | 2020-07-03 | 2021-08-31 | 东北大学 | Metal surface defect detection method based on linear array camera stereoscopic vision |
| CN113155852A (en) * | 2021-04-08 | 2021-07-23 | 煤炭科学研究总院 | Transmission band detection method and device and electronic equipment |
| CN113092489A (en) * | 2021-05-20 | 2021-07-09 | 鲸朵(上海)智能科技有限公司 | System and method for detecting appearance defects of battery |
| CN113252567A (en) * | 2021-06-08 | 2021-08-13 | 菲特（天津）检测技术有限公司 | Method, system, medium and terminal for rapidly detecting multiple defects on surface of aluminum valve plate |
| CN113781424A (en) * | 2021-09-03 | 2021-12-10 | 苏州凌云光工业智能技术有限公司 | Surface defect detection method, device and equipment |
| CN113781424B (en) * | 2021-09-03 | 2024-02-27 | 苏州凌云光工业智能技术有限公司 | Surface defect detection method, device and equipment |
| CN116559181A (en) * | 2023-07-07 | 2023-08-08 | 杭州灵西机器人智能科技有限公司 | Defect detection method, system, device and medium based on luminosity stereoscopic vision |
| CN116559181B (en) * | 2023-07-07 | 2023-10-10 | 杭州灵西机器人智能科技有限公司 | Defect detection method, system, device and medium based on luminosity stereoscopic vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107607546B (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107607546A (en) | Leather defect inspection method, system and device based on photometric stereo vision | |
| CN106875381B (en) | Mobile phone shell defect detection method based on deep learning | |
| CN111160269A (en) | Face key point detection method and device | |
| CN109685848A (en) | A kind of neural network coordinate transformation method of three-dimensional point cloud and three-dimension sensor | |
| CN107729872A (en) | Facial expression recognition method and device based on deep learning | |
| CN107194937A (en) | Tongue image partition method under a kind of open environment | |
| CN107845086A (en) | A kind of detection method, system and the device of leather surface conspicuousness defect | |
| CN106355973A (en) | Method and device for guiding drawing | |
| CN110705655A (en) | Tobacco leaf classification method based on coupling of spectrum and machine vision | |
| CN113393461B (en) | Method and system for screening metaphase chromosome image quality based on deep learning | |
| Adem et al. | Defect detection of seals in multilayer aseptic packages using deep learning | |
| CN107958253A (en) | A kind of method and apparatus of image recognition | |
| CN109472774A (en) | A kind of tongue picture picture quality detection method based on deep learning | |
| CN112001901A (en) | Apple defect detection method and system based on convolutional neural network | |
| CN106056161B (en) | A kind of visible detection method for Plane Rotation target | |
| CN114627116B (en) | Fabric defect identification method and system based on artificial intelligence | |
| CN114419011A (en) | Cotton foreign fiber online detection method and system | |
| CN105354405A (en) | Machine learning based immunohistochemical image automatic interpretation system | |
| CN112862744A (en) | Intelligent detection method for internal defects of capacitor based on ultrasonic image | |
| CN108242060A (en) | A kind of method for detecting image edge based on Sobel operators | |
| CN110321922A (en) | A kind of CT image classification method for distinguishing convolutional neural networks based on space correlation | |
| CN110321936A (en) | A method of realizing that picture two is classified based on VGG16 and SVM | |
| CN111144331B (en) | Elbow vein image elbow median vein recognition method and elbow image acquisition device | |
| Khanal et al. | Leather defect detection using semantic segmentation: A hardware platform and software prototype | |
| CN110211122A (en) | A kind of detection image processing method and processing device |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Lin Jianfa Inventor after: Liu Gen Inventor after: Xiao Pan Inventor after: Huang Guancheng Inventor before: Lin Jianfa Inventor before: Liu Gen Inventor before: Wang Han Inventor before: Cai Nian Inventor before: Chen Xindu |
|
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |