CN108596931A - A kind of noise robustness hub edge detection algorithm based on Canny operators - Google Patents
A kind of noise robustness hub edge detection algorithm based on Canny operators Download PDFInfo
- Publication number
- CN108596931A CN108596931A CN201810447367.6A CN201810447367A CN108596931A CN 108596931 A CN108596931 A CN 108596931A CN 201810447367 A CN201810447367 A CN 201810447367A CN 108596931 A CN108596931 A CN 108596931A
- Authority
- CN
- China
- Prior art keywords
- image
- hub
- edge
- canny operators
- gradient magnitude
- 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.)
- Withdrawn
Links
- 238000003708 edge detection Methods 0.000 title claims abstract description 20
- 238000004519 manufacturing process Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000009499 grossing Methods 0.000 claims abstract description 5
- 230000002708 enhancing effect Effects 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 description 11
- 238000011161 development Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 230000001154 acute effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of noise robustness hub edge detection algorithm based on Canny operators, includes the following steps:1) the automotive hub front elevation at hub production line scene is shot with video camera, and carries out Gaussian smoothing filter;2) gradient magnitude of the image after calculating smoothly, obtains gradient magnitude image;3) operation in step 1) and step 2) is repeated to gradient magnitude image, obtains the two-wire edge contour of gradient magnitude;4) multi-difference operation refines hub edge image;5) enhancing processing is carried out to hub edge image.The algorithm reduces image fault, avoids the occurrence of false edge, improves the precision and accuracy of the industrial robot wheel hub deburring system based on machine vision.
Description
Technical field
The present invention relates to a kind of Boundary extracting algorithm, specifically a kind of noise robustness hub edge based on Canny operators
Detection algorithm belongs to technical field of image processing.
Background technology
Burr is the inevitable phenomenon generated in intermetallic composite coating.In automotive hub industry, in process due to wheel hub
It will produce burr phenomena so that the quality of product can be affected, and restrict the development of automotive hub enterprise.In recent years, to vapour
The main stream approach that wheel hub carries out deburring is to carry out teaching programming with industrial robot, it is allowed to drive workpiece or cutter, edge
Fixed track movement, to realize the polishing to robot edge.This method is for a kind of certain fixing model, style, ruler
It is suitable for very little wheel hub, but when a variety of automotive wheels for occurring different styles, different model on a flow production line
When hub, it is necessary to carry out teaching programming to robot again, this undoubtedly reduces the real-time and rapidity of the entire production line, no
Meet the rate of economic development and industry demand of current fast development.
With today's society the degree of automation and science and technology continuous development, machine vision technique increasingly by
The concern of people.Industrial robot deburring system based on machine vision is by machine vision technique and image processing techniques knot
Altogether, robot oneself is allowed to possess " eyes ", realization polishes to the automotive hub of different styles and model.
The edge extracting of automotive hub is basis and the core of entire machine vision algorithm.Vapour is only accurately extracted
The profile of wheel hub could carry out robot trajectory planning according to the coordinate of profile point and then control the movement of robot.Image
Edge be usually image gray scale or color occur acute variation place.The edge detection operator of mainstream has at present
Canny operators, Roberts operators, Prewitt operators, Krisch operators, Sobel operators, Laplacian operators and LOG are calculated
Son.Wherein, Canny operators are considered as one of most successful edge detection method.In order to refine edge in Canny operators,
Non-maxima suppression has been carried out to gradient assignment, the purpose is to find the local maximum of image all pixels point, but it is traditional
Canny operators when carrying out non-maxima suppression operation, due to contrast points selection have prodigious randomness, so can produce
Raw false edge, especially when, there are when noise, even more will produce a large amount of false edge in image.
In conclusion existing Canny edge detection methods are to the industrial robot wheel hub deburring based on machine vision
The precision and accuracy of system are all influential.
Invention content
The technical problem to be solved by the present invention is to:For deficiency of the above-mentioned Canny operators in edge detection, one is proposed
Edge detection method of the kind based on the improved accurate extraction Noise image outline of Canny operators.Image fault is reduced, is avoided out
Existing false edge, to improve the precision and accuracy of the industrial robot wheel hub deburring system based on machine vision.
The technical proposal for solving the technical problem of the invention is:A kind of noise robustness wheel hub based on Canny operators
Edge detection algorithm, the edge detection algorithm include the following steps:
1) the automotive hub front elevation at hub production line scene is shot with video camera, and carries out Gaussian smoothing filter;
2) gradient magnitude of the image after calculating smoothly, obtains gradient magnitude image;
3) operation in step 1) and step 2) is repeated to gradient magnitude image, obtains the two-wire edge wheel of gradient magnitude
It is wide;
4) multi-difference operation refines hub edge image;
5) enhancing processing is carried out to hub edge image.
Further, step 1) by formula (1) construct two-dimensional Gaussian function G (x, y), with its to original image f (x, y) into
Row convolution operation, obtains the smoothed image g (x, y) of formula (2), and concrete operation step is as follows:
G (x, y)=G (x, y) * f (x, y) (2)
Wherein x, y indicate that the cross of original image pixels point, ordinate value, σ are the standard deviation of Gaussian function respectively.
Further, the gradient magnitude of the image after step 2) is calculated smoothly using Canny operators.
Further, it is using the circular of Canny operators:
Using the first-order partial derivative finite difference of 2 × 2 window sizes, calculate by the filtered image ladder of Gaussian filter
Degree vector:
gx/ 2 (3) (x, y)=[g (x, y+1)-g (x, y)+g (x+1, y+1)-g (x+1, y)]
gy/ 2 (4) (x, y)=[g (x, y)-g (x+1, y)+g (x, y+1)-g (x+1, y+1)]
gx(x, y) and gy(x, y) expression smoothed image g (x, y) is in the gradient vector in the directions x and y respectively, after obtaining smoothly
Image g (x, y) gradient magnitude M (x, y):
Further, the circular of step 3) is:
gM(x, y)=G (x, y) * M (x, y) (7)
Wherein gM(x, y) is the gradient magnitude image obtained after Gaussian smoothing, gMx(x, y) and gMy(x, y) is respectively
Gradient magnitude image after smooth the directions x and y gradient vector, M ' (x, y) be it is smooth after gradient magnitude image gradient
Amplitude, that is, original image Quadratic Pressure Gradient amplitude.
Further, step 4) carries out calculus of differences to the M (x, y) and M ' (x, y) that are obtained in step 2 and step 3, obtains
Edge M " (x, y) in formula (9) after refinement:
M " (x, y)=M (x, y)-M ' (x, y) (9)
Further, step 4) is in order to further obtain thinner edge, and repeat the above steps progress multi-difference operation,
Obtain a plurality of refinement edge of respective numbers.
Further, step 5) by a plurality of refinement edge obtained in step 4) be added to obtain enhanced edge MF (x,
y)。
Further, the pixel value that 0 is less than in MF (x, y) is set to 0 by step 5), you can obtains final hub edge wheel
Exterior feature is embodied as:
Description of the drawings
Fig. 1 is the implementation flow chart of original Canny operators.
Fig. 2 is the general frame implementation flow chart of the method for the present invention.
Fig. 3 is the practical automotive hub image of a width that industrial camera takes.
Fig. 4 (a) is the edge of automotive hub in the Fig. 4 obtained using tradition Canny operators, in traditional Canny operators,
Take Gaussian filter parameter σ=2.
Fig. 4 (b) is the edge of automotive hub in the Fig. 4 obtained using improvement Canny operators proposed by the present invention, in this hair
In bright, Gaussian filter parameter σ=2 are equally taken.
Fig. 5 (a) is the noisy acoustic image to be obtained after 10 Gaussian noise to the automotive hub addition standard deviation in Fig. 4.
Fig. 5 (b) is the noisy acoustic image to be obtained after 20 Gaussian noise to the automotive hub addition standard deviation in Fig. 4.
Fig. 6 (a) is the edge of Noise automotive hub in the Fig. 6 (a) obtained using tradition Canny operators, in tradition
In Canny operators, Gaussian filter parameter σ=2 are taken.
Fig. 6 (b) is the side of Noise automotive hub in the Fig. 6 (a) obtained using improvement Canny operators proposed by the present invention
Edge equally takes Gaussian filter parameter σ=2 in the present invention.
Fig. 7 (a) is the edge of Noise automotive hub in the Fig. 6 (b) obtained using tradition Canny operators, in tradition
In Canny operators, Gaussian filter parameter σ=2 are taken.
Fig. 7 (b) is the side of Noise automotive hub in the Fig. 6 (b) obtained using improvement Canny operators proposed by the present invention
Edge equally takes Gaussian filter parameter σ=2 in the present invention.
Specific implementation mode
Technical scheme of the present invention is described in further detail with reference to embodiments, the present embodiment is in skill of the present invention
Implemented under the premise of art scheme, gives detailed embodiment and specific operating process, but the protection model of the present invention
It encloses and is not limited to following embodiments.
The present embodiment is to carry out edge extracting to the automotive hub image that certain automotive hub production line photographs arrives, and is extracted
Journey is as shown in Fig. 2, specifically include following steps:
1) the automotive hub figure f (x, y) that hub production line scene is shot with video camera, as shown in figure 3, and carrying out height to it
This smothing filtering, construction two-dimensional Gaussian function G (x, y) carry out convolution operation to original image f (x, y) with it, are smoothly schemed
As g (x, y), concrete operation step is as follows:
G (x, y)=G (x, y) * f (x, y)
2) it uses Canny operators to use the first-order partial derivative finite difference of 2 × 2 window sizes, calculates and pass through gaussian filtering
The filtered image gradient vector of device:
gx(x, y)=[g (x, y+1)-g (x, y)+g (x+1, y+1)-g (x+1, y)]/2
gy(x, y)=[g (x, y)-g (x+1, y)+g (x, y+1)-g (x+1, y+1)]/2
Then the gradient magnitude M (x, y) of the image g (x, y) after can obtaining smoothly:
3) operation in step 1) and step 2) is repeated to gradient magnitude image M (x, y), obtains the two-wire item of gradient magnitude
Edge contour M ' (x, y), concrete operation step are:
gM(x, y)=G (x, y) * M (x, y)
4) calculus of differences, the edge after being refined are carried out to the M (x, y) and M ' (x, y) that are obtained in step 2 and step 3
M " (x, y):
M " (x, y)=M (x, y)-M ' (x, y)
In order to further obtain thinner edge, repeats the above steps and carry out multi-difference operation.This experiment carries out three times
Calculus of differences obtains three thinning edges:M " (x, y), M " ' (x, y), M " " (x, y).
5) enhancing processing is carried out to hub edge image, concrete operations are the three thinning edge phases that will be obtained in step 4)
Add to obtain edge MF (x, y):
MF (x, y)=M " (x, y)+M " ' (x, y)+M " " (x, y)
Then the pixel value for being less than 0 in MF (x, y) is set to 0, you can final hub edge profile is obtained, it is specific to indicate
For:
In order to verify the validity of this algorithm, calculated respectively with original Canny operators and improvement Canny proposed by the present invention
Son acts on the automotive hub image of Fig. 3, shown in edge image such as Fig. 4 (a) and Fig. 4 (b) of extraction.Two figures, which compare, can be seen that
The hub edge extracted with the present invention is more accurate, reduces false edge, and can reach same with original Canny operators
Refinement precision and marginal signal strength.Further, in order to verify the noise robustness of this algorithm, to the automotive hub of Fig. 3
Image adds the Gaussian noise that standard deviation is 10 and 20, and obtained noisy image such as Fig. 5 (a) and Fig. 5 (b) are shown.In order to compare
Original Canny operators and improvement Canny operators proposed by the present invention calculate the edge extracting effect of noisy acoustic image with two kinds
Method carries out edge extracting, edge contour figure such as Fig. 6 and Fig. 7 institutes of extraction to the noisy wheel hub image of Fig. 5 (a) He Fig. 5 (b) respectively
Show.Compare (a) of Fig. 6 and Fig. 7, (b) can be obtained, improvement Canny operators proposed by the present invention remain able to accurately noisy image
Ground extracts edge, has noise robustness, and original Canny operators can then extract a large amount of false edge.
In order to quantitatively assess operator of the present invention, it is 5,10,15 and to add standard deviation respectively to the automotive hub image of Fig. 3
20 Gaussian noise, and more original Canny operators and the Canny operators proposed by the present invention that improve examine the edge of noisy image
Effect is surveyed, this experimental selection Y-PSNR (PSNR) and quality factor (FOM) are used as measurement index.Specifically, PSNR values are got over
Greatly, it is meant that the contour images distortion detected is smaller;And FOM values are bigger, and it is more preferable, false to represent obtained edge connectivity
Edge is less.Assessment result is as shown in table 1, and as can be known from the table data, compared with original Canny operators, use is proposed by the present invention
The edge contour image fault that innovatory algorithm obtains is less, and has better edge connectivity and less false edge.
Table 1
Claims (9)
1. a kind of noise robustness hub edge detection algorithm based on Canny operators, which is characterized in that the edge detection algorithm
Include the following steps:
1) the automotive hub front elevation at hub production line scene is shot with video camera, and carries out Gaussian smoothing filter;
2) gradient magnitude of the image after calculating smoothly, obtains gradient magnitude image;
3) operation in step 1) and step 2) is repeated to gradient magnitude image, obtains the two-wire edge contour of gradient magnitude;
4) multi-difference operation refines hub edge image;
5) enhancing processing is carried out to hub edge image.
2. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 1, feature exist
In step 1) constructs two-dimensional Gaussian function G (x, y) by formula (1), carries out convolution operation to original image f (x, y) with it, obtains
To the smoothed image g (x, y) of formula (2), concrete operation step is as follows:
G (x, y)=G (x, y) * f (x, y) (2)
Wherein x, y indicate that the cross of original image pixels point, ordinate value, σ are the standard deviation of Gaussian function respectively.
3. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 1, feature exist
In the gradient magnitude of image of the step 2) using the calculating of Canny operators after smooth.
4. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 3, feature
It is that the circular using Canny operators is:
Using the first-order partial derivative finite difference of 2 × 2 window sizes, calculate by the filtered image gradient of Gaussian filter to
Amount:
gx/ 2 (3) (x, y)=[g (x, y+1)-g (x, y)+g (x+1, y+1)-g (x+1, y)]
gy/ 2 (4) (x, y)=[g (x, y)-g (x+1, y)+g (x, y+1)-g (x+1, y+1)]
gx(x, y) and gy(x, y) indicates smoothed image g (x, y) in the gradient vector in the directions x and y, the figure after obtaining smoothly respectively
As the gradient magnitude M (x, y) of g (x, y):
5. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 1, feature exist
In the circular of step 3) is:
gM(x, y)=G (x, y) * M (x, y) (7)
Wherein gM(x, y) is the gradient magnitude image obtained after Gaussian smoothing, gMx(x, y) and gMy(x, y) is respectively smooth
Gradient magnitude image afterwards the directions x and y gradient vector, M ' (x, y) be it is smooth after gradient magnitude image gradient magnitude,
The namely Quadratic Pressure Gradient amplitude of original image.
6. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 1, feature exist
In step 4) carries out calculus of differences to the M (x, y) and M ' (x, y) that are obtained in step 2 and step 3, obtains in formula (9) after refinement
Edge M " (x, y):
M " (x, y)=M (x, y)-M ' (x, y) (9)
7. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 6, feature exist
In step 4) repeats the above steps to further obtain thinner edge and carries out multi-difference operation, obtain respective numbers
A plurality of refinement edge.
8. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 7, feature exist
In step 5) is added a plurality of refinement edge obtained in step 4) to obtain enhanced edge MF (x, y).
9. a kind of noise robustness hub edge detection algorithm based on Canny operators according to claim 8, feature exist
In the pixel value for being less than 0 in MF (x, y) is set to 0 by step 5), you can is obtained final hub edge profile, is embodied as:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810447367.6A CN108596931A (en) | 2018-05-11 | 2018-05-11 | A kind of noise robustness hub edge detection algorithm based on Canny operators |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810447367.6A CN108596931A (en) | 2018-05-11 | 2018-05-11 | A kind of noise robustness hub edge detection algorithm based on Canny operators |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108596931A true CN108596931A (en) | 2018-09-28 |
Family
ID=63636725
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810447367.6A Withdrawn CN108596931A (en) | 2018-05-11 | 2018-05-11 | A kind of noise robustness hub edge detection algorithm based on Canny operators |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108596931A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110838127A (en) * | 2019-10-30 | 2020-02-25 | 合肥工业大学 | Feature image edge detection method for intelligent automobile |
CN111445512A (en) * | 2020-06-17 | 2020-07-24 | 浙江大学 | Hub parameter feature extraction method in complex production line background |
CN112163587A (en) * | 2020-09-30 | 2021-01-01 | 北京环境特性研究所 | Feature extraction method and device of target object and computer readable medium |
CN113505811A (en) * | 2021-06-10 | 2021-10-15 | 常州理工科技股份有限公司 | Machine vision imaging method for hub production |
CN115880289A (en) * | 2023-02-21 | 2023-03-31 | 深圳普菲特信息科技股份有限公司 | Steel coil burr identification method, system and medium based on big data processing |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080293329A1 (en) * | 2007-05-21 | 2008-11-27 | Applied Materials, Inc. | Methods and apparatus for identifying a substrate edge profile and adjusting the processing of the substrate according to the identified edge profile |
CN102184552A (en) * | 2011-05-11 | 2011-09-14 | 上海理工大学 | Moving target detecting method based on differential fusion and image edge information |
CN103927526A (en) * | 2014-04-30 | 2014-07-16 | 长安大学 | Vehicle detecting method based on Gauss difference multi-scale edge fusion |
CN105548185A (en) * | 2016-01-08 | 2016-05-04 | 浙江科技学院 | Automobile wheel hub screw hole recognition method based on machine vision and shielding method and system |
-
2018
- 2018-05-11 CN CN201810447367.6A patent/CN108596931A/en not_active Withdrawn
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080293329A1 (en) * | 2007-05-21 | 2008-11-27 | Applied Materials, Inc. | Methods and apparatus for identifying a substrate edge profile and adjusting the processing of the substrate according to the identified edge profile |
CN102184552A (en) * | 2011-05-11 | 2011-09-14 | 上海理工大学 | Moving target detecting method based on differential fusion and image edge information |
CN103927526A (en) * | 2014-04-30 | 2014-07-16 | 长安大学 | Vehicle detecting method based on Gauss difference multi-scale edge fusion |
CN105548185A (en) * | 2016-01-08 | 2016-05-04 | 浙江科技学院 | Automobile wheel hub screw hole recognition method based on machine vision and shielding method and system |
Non-Patent Citations (2)
Title |
---|
张巍等: "多孔介质中甲烷水合物边界的CT图像识别技术", 《CT理论与应用研究》 * |
杨飞等: "一种自动的轮毂识别分类方法", 《微计算机信息》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110838127A (en) * | 2019-10-30 | 2020-02-25 | 合肥工业大学 | Feature image edge detection method for intelligent automobile |
CN111445512A (en) * | 2020-06-17 | 2020-07-24 | 浙江大学 | Hub parameter feature extraction method in complex production line background |
CN112163587A (en) * | 2020-09-30 | 2021-01-01 | 北京环境特性研究所 | Feature extraction method and device of target object and computer readable medium |
CN113505811A (en) * | 2021-06-10 | 2021-10-15 | 常州理工科技股份有限公司 | Machine vision imaging method for hub production |
CN115880289A (en) * | 2023-02-21 | 2023-03-31 | 深圳普菲特信息科技股份有限公司 | Steel coil burr identification method, system and medium based on big data processing |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108596931A (en) | A kind of noise robustness hub edge detection algorithm based on Canny operators | |
CN107220988B (en) | Part image edge extraction method based on improved canny operator | |
CN109034017B (en) | Head pose estimation method and machine readable storage medium | |
CN104299008B (en) | Vehicle type classification method based on multi-feature fusion | |
CN104331869B (en) | The image smoothing method that gradient is combined with curvature | |
CN107067382A (en) | A kind of improved method for detecting image edge | |
CN108765419A (en) | Structure light vision weld image information self-adapting extracting method | |
CN107645621A (en) | A kind of method and apparatus of Video processing | |
AU2013206597A1 (en) | Depth constrained superpixel-based depth map refinement | |
CN107622480B (en) | Kinect depth image enhancement method | |
CN104376319B (en) | A kind of method based on anisotropic Gaussian core extraction closed edge image outline | |
CN109003233B (en) | Image denoising method based on self-adaptive weight total variation model | |
CN106023204A (en) | Method and system for removing mosquito noise based on edge detection algorithm | |
CN106780526A (en) | A kind of ferrite wafer alligatoring recognition methods | |
CN105225243B (en) | One kind can antimierophonic method for detecting image edge | |
CN108921813A (en) | Unmanned aerial vehicle detection bridge structure crack identification method based on machine vision | |
CN103942756B (en) | A kind of method of depth map post processing and filtering | |
CN111223059A (en) | Robust depth map structure reconstruction and denoising method based on guide filter | |
CN114332081B (en) | Textile surface abnormity determination method based on image processing | |
CN109345609A (en) | Mural painting image denoising is carried out based on convolutional neural networks and line drawing draws the method generated | |
CN107230200A (en) | A kind of method for extracting rotor coil contour feature | |
CN102118547A (en) | Image weighted filtering method | |
CN106846271A (en) | A kind of method of reticulate pattern in removal identity card picture | |
CN101504769B (en) | Self-adaptive noise intensity estimation method based on encoder frame work | |
CN110033437A (en) | A kind of cut deal impression intelligent identification Method based on convolutional neural networks |
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 | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180928 |
|
WW01 | Invention patent application withdrawn after publication |