CN107220988A - Based on the parts image edge extraction method for improving canny operators - Google Patents
Based on the parts image edge extraction method for improving canny operators Download PDFInfo
- Publication number
- CN107220988A CN107220988A CN201710300398.4A CN201710300398A CN107220988A CN 107220988 A CN107220988 A CN 107220988A CN 201710300398 A CN201710300398 A CN 201710300398A CN 107220988 A CN107220988 A CN 107220988A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- gray
- mfrac
- image
- 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
- 238000000605 extraction Methods 0.000 title claims abstract description 12
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000001914 filtration Methods 0.000 claims abstract description 9
- 238000001228 spectrum Methods 0.000 claims description 8
- 230000006872 improvement Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 description 6
- 238000003708 edge detection Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000003706 image smoothing Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000009394 selective breeding Methods 0.000 description 1
- 230000001629 suppression Effects 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
Abstract
The present invention discloses a kind of based on the parts image edge extraction method for improving canny operators, first acquisition parts coloured image, and be converted into gray-scale map;Then frequency domain filtering is carried out using Fourier transform pairs gray-scale map;Convolution sharpening is carried out to filtered gray-scale map followed by the prewitt operators template in 4 directions, the greatest gradient figure of image is obtained;The grey level histogram of greatest gradient figure is calculated again, threshold value T is drawn using the lowest point method, and gray-scale map is divided into 2 regions using threshold value T, maximum variance between clusters are used to 2 regions respectively, high-low threshold value is calculated;Finally according to high-low threshold value by gray-scale map binaryzation, basic as edge using the binary map that high threshold is generated, the binary map produced using Low threshold carries out the connection on edge basis, obtains complete edge binary map.The inventive method improves edge extracting precision, and adaptively produces high-low threshold value, improves efficiency.
Description
Technical field
The invention belongs to image procossing and field of machine vision, and in particular to a kind of based on improving zero of canny operators
Part image edge extraction method.
Background technology
Image border is one of most basic feature of image, contains the most information of image.In image procossing and meter
In calculation machine visual field, Image Edge-Detection is most basic technology, has important effect in subsequent treatment.In parts
In image, the accurate detection at parts edge is to examining parts acceptable level to serve important function.
Image Edge-Detection can be divided into following 3 steps:
1. image denoising:When carrying out Image Edge-Detection using the single order and second dervative of gradation of image, the derivative of image
Calculate affected by noise very big, therefore first to image denoising, to improve the performance of rim detection.
2. edge strengthens:Obvious point can be changed by each vertex neighborhood of image or local strength by calculating the amplitude of image gradient
Highlight.
3. detection:The gradient magnitude of usual image border point is than larger, but the point of image gradient amplitude greatly has in practice
A lot, and these points are not necessarily all edge sometimes, therefore to use suitable method to determine which is marginal point.Rim detection
Method is generally divided into the operator detection method of first derivative and second dervative.
Edge is that the derivative value at the relatively more significant place of gradation of image value changes, gradation of image mutation is also larger, therefore
Typically grey scale change is represented with the size of gray scale derivative.By picture point ux(x, y) and point uyThe gradient of (x, y) is defined as following
Vector:
The amplitude of gradient is given by:
According to the size and the difference of weights of convolution mask, many gradient operators are occurred in that.
However, the denoising and positioning precision due to image border are always difficult to what is ensured simultaneously, in smooth noise, do not wish
Prestige smooths out marginal information, equally, when sharpening, and only wants to sharpen edge and be not intended to receive the interference of noise.How to balance
Relation between the two is the key problem in Edge extraction.Meanwhile, in boundary operator, most of template is only to level
There is good response with vertical edge, the edge extracting ability to other angles is poor.
Canny operators are the Optimizing operators for being combined image smoothing, edge sharping and detecting.It is first in rim detection
Carry out convolution to image, except denoising, each pixel then to be calculated using first-order difference template first with Gaussian function
Direction and gradient magnitude, then using non-maxima suppression principle, the point that each vertex neighborhood intensity level of image is had significant change
Highlight, to reach the effect at enhancing edge, finally will finally give edge graph using two threshold tests and connection edge
Picture.
But traditional first-order difference template only has preferable response to edge vertically and horizontally, to the side in other directions
Edge Detection results are poor.It is difficult to choose most using the method for artificial selection simultaneously in the selection of two final threshold values
Excellent threshold value, is connected so as to influence last edge to choose with edge.Inappropriate threshold value selection can cause image is produced a lot
Pseudo-edge and noise, the interference later stage is to the further processing of image with being applicable.
The content of the invention
It is an object of the invention to provide based on it is a kind of based on improve Canny operators parts image edge extraction method,
Automatically generate comprising the sharp-edged single pixel binary map of target part, the convenient follow-up analysis to each parameter of part.
The technical solution for realizing the object of the invention is:A kind of parts image border based on improvement canny operators
Extracting method, comprises the following steps:
Step 1, acquisition parts coloured image, and it is converted into gray-scale map;
Step 2, to gray-scale map carry out frequency domain filtering;
Step 3, using 4 directions prewitt operators template to filtered gray-scale map carry out convolution sharpening, obtain figure
The greatest gradient figure of picture;
Step 4, the grey level histogram for calculating greatest gradient figure, threshold value T is drawn using the lowest point method, using threshold value T by gray scale
Figure is divided into 2 regions, uses maximum variance between clusters to 2 regions respectively, calculates high-low threshold value;
Step 5, according to high-low threshold value by gray-scale map binaryzation, the binary map generated using high threshold is utilized as edge basis
The binary map that Low threshold is produced carries out the connection on edge basis, obtains complete edge binary map.
The step 2 carries out concretely comprising the following steps for frequency filtering:
Step 2.1, using fast Fourier algorithm F (u, v) is obtained to gray-scale map, and F (u, v) zero-frequency point is moved to
The center of spectrogram;
Step 2.2, calculate Gaussian function filter function H (u, v) and F (u, v) product G (u, v), by frequency spectrum G (u,
V) zero-frequency point is moved back into the upper left position of spectrogram, wherein
Gaussian filter function representation is:
Wherein, u, v are frequency domain variable, and M, N are respectively the level of image, vertical pixel number.σ is Gaussian function
Sigma parameters;
Frequency spectrum designation is:
G (u, v)=H (u, v) × F (u, v)
Step 2.3, to G (u, v) carry out inverse discrete fourier transform obtain g (x, y), take g (x, y) real part as filtering
Result images afterwards.
The step 3 asks concretely comprising the following steps for greatest gradient figure:
Step 3.1,45 ° using prewitt, 135 °, level, the template in vertical 4 directions and the filtered ash of completion
Degree figure carries out convolution respectively;
Step 3.2, take value maximum in each 4 convolution results of pixel as Grad, the corresponding direction of maximum is
Greatest gradient direction, produces greatest gradient image.
The specific method that the step 4 calculates Low threshold is:
If greatest gradient figure is Q (x, y), Q (x, y) tonal range is [0, L-1], show that threshold value is T using the lowest point method,
Then Q (x, y) is divided into 2 scopes [0, T] and [T, L-1] by threshold value T, if the pixel count that gray scale is i in image is ni, in gray scale
Scope is that [0, T] interior total pixel number is:
The probability that each gray scale occurs is:
In [0, T], threshold value T is used1It is divided into 2 class C0And C1, C0By [0, T1- 1] constitute, C1By [T1, T] and composition, then
Region C0And C1Probability be respectively:
P1=1-P0
C0And C1Average gray be respectively:
Wherein, μ is the average gray of [0, T]:
μ=P0μ0+P1μ1
The population variance in two regions is:
σB 2=P0(μ0-μ)2+P1(μ1-μ)2=P0P1(μ0-μ1)2
T is allowed [0, T1- 1] interior value successively, makes σB 2Maximum T values are the optimal selections of Low threshold.
Similarly, in [T1, L-1] and middle repetition above step, you can obtain the optimal selection of high threshold.
Compared with existing invention, the present invention has following advantage:1) present invention is located in advance by gaussian filtering to image
Reason, good filter out has been carried out to noise.2) in operator is sharpened, the prewitt operators of four direction have been used as gradient
Template, adds the template of 45 ° and 135 ° both directions, to the side in 8 directions on the basis of vertically and horizontally two templates
Edge has all accomplished good response, improves edge extracting precision.3) in the high-low threshold value selection of binaryzation, using histogram
The lowest point method is combined with maximum variance between clusters, adaptively produces high-low threshold value, and precision is improved while improving efficiency.
Brief description of the drawings
Fig. 1 is Edge extraction process overall schematic.
Fig. 2 is frequency filtering process schematic.
Fig. 3 is edge sharpening process schematic.
Fig. 4 is adaptive threshold selection and final edge binaryzation schematic diagram.
Fig. 5 is the edge binary map that the inventive method is produced.
Fig. 6 is the edge binary map produced with conventional method.
Fig. 7 is 4 Prototype drawings of Prewitt operators.
Embodiment
The invention will be further described below in conjunction with the accompanying drawings.
Based on the parts image edge extraction method for improving canny operators, industrial camera is opened and connected first
To PC, the part being taken is fixed, is brought into focus, a width is shot after clearly picture, installation it is determined that can collect
Clearly picture is by the incoming computer processing system of Ethernet, and processing platform is visual studio 2010+opencv
2.4.10.Next image processing step is entered:
Step 1, the RGB coloured pictures image collected is subjected to gradation conversion, generates gray-scale map, can be with by the threshold value of gray-scale map
Regard gray-scale map as f (x, a y) function.
Step 2, to gray-scale map carry out frequency filtering, concretely comprise the following steps:
Step 2.1, the discrete Fourier transform for calculating f (x, y), due to DFT to be directly realized by efficiency low, the application is used
A kind of fast Fourier algorithm of decimation in time (DIT-FFT) obtains F (u, v), and F (u, v) zero-frequency point is moved into frequency
The center of spectrogram;
Step 2.2, the product G (u, v) for calculating filter function H (u, v) and F (u, v), as frequency spectrum, the application are used
One Gaussian function:
Wherein, u, v are frequency domain variable, and M, N is respectively the level of image, vertical pixel number.σ is Gaussian function
Sigma parameters.
Frequency spectrum is filter function H (u, v) and F (u, v) product:
G (u, v)=H (u, v) × F (u, v)
Then frequency spectrum G (u, v) zero-frequency point is moved back into the upper left position of spectrogram;
Step 2.3, in step 2.2 G (u, v) carry out inverse discrete fourier transform (IDFT) calculating, obtain g (x,
Y), g (x, y) real part is taken as filtered result images.
Step 3, using 4 directions prewitt operators template to filtered gray-scale map carry out convolution sharpening, obtain figure
The greatest gradient figure of picture, is concretely comprised the following steps:
Step 3.1,45 ° using prewitt, 135 °, level, the template in vertical 4 directions and the filtered ash of completion
Degree figure g (x, y) carries out convolution respectively, and 4 templates of Prewitt operators are as shown in Figure 7:
Step 3.2, take value maximum in each 4 convolution results of pixel as Grad, the corresponding direction of maximum is
Greatest gradient direction, can obtain greatest gradient image Q (x, y) after so having handled.
Step 4:Greatest gradient image is subjected to thresholding by high-low threshold value, traditional threshold value selection is needed by being accomplished manually
Carry out test of many times and there may be the situation of pseudo-edge.Used here as histogram the lowest point and maximum between-cluster variance combined techniques
Adaptively to choose high-low threshold value, concretely comprise the following steps:
Step 4.1, the grey level histogram for calculating greatest gradient figure Q (x, y), draw threshold value T, and utilize threshold using the lowest point method
Gray-scale map is divided into 2 regions by value T, if Q (x, y) tonal range is that [0, L-1] then Q (x, y) is divided into 2 scopes by threshold value T
[0, T] and [T, L-1];
Step 4.2, maximum variance between clusters are used to 2 regions respectively, calculate high-low threshold value, specific method is:
If the pixel count that gray scale is i in image is ni, it is that [0, T] interior total pixel number is in tonal range:
The probability that each gray scale occurs is:
In [0, T], threshold value T is used1It is divided into 2 class C0And C1, C0By [0, T1- 1] constitute, C1By [T1, T] and composition, then
Region C0And C1Probability be respectively:
P1=1-P0
C0And C1Average gray be:
Wherein, μ is the average gray of [0, T]:
μ=P0μ0+P1μ1
The population variance in two regions is:
σB 2=P0(μ0-μ)2+P1(μ1-μ)2=P0P1(μ0-μ1)2
T is allowed [0, T1- 1] interior value successively, makes σB 2Maximum T values are the optimal selections of Low threshold;
Similarly, in [T1, L-1] and middle repetition above step, you can the optimal selection of high threshold is obtained, specific method is no longer
Repeat.
Step 5, using the high-low threshold value in step 4, obtain two threshold skirt image N1[i, j], N2[i, j], N1[i,
J] obtained by Low threshold, N2[i, j] is obtained by high threshold.Due to N2[i, j] is obtained by high threshold, and its edge is essentially true edge,
But there is fracture, so finding N in 8 fields of breaking part1Edge connection in [i, j], until by N2[i, j] is connected
Untill coming, so, a complete edge binary map has just been obtained.
As can be seen that the edge binary map generated with traditional prewitt methods is present largely from accompanying drawing 5 and accompanying drawing 6
Noise, and pseudo-edge is more, it is impossible to obtain accurate Single pixel edge.
The edge binary map generated with the inventive method, has effective filtered out most of noise first, in terms of edge, subtracts
The small generation of pseudo-edge, generates accurate Single pixel edge.
Claims (4)
1. based on the parts image edge extraction method for improving canny operators, it is characterised in that comprise the following steps:
Step 1, acquisition parts coloured image, and it is converted into gray-scale map;
Step 2, to gray-scale map carry out frequency domain filtering;
Step 3, using 4 directions prewitt operators template to filtered gray-scale map carry out convolution sharpening, obtain image
Greatest gradient figure;
Step 4, the grey level histogram for calculating greatest gradient figure, threshold value T is drawn using the lowest point method, is divided gray-scale map using threshold value T
For 2 regions, maximum variance between clusters are used to 2 regions respectively, high-low threshold value is calculated;
Step 5, according to high-low threshold value by gray-scale map binaryzation, the binary map generated using high threshold utilizes low threshold as edge basis
The binary map that value is produced carries out the connection on edge basis, obtains complete edge binary map.
2. the image edge extraction method based on improvement Canny operators according to described by claim 1, it is characterised in that institute
State step 2 and carry out concretely comprising the following steps for frequency filtering:
Step 2.1, using fast Fourier algorithm F (u, v) is obtained to gray-scale map, and F (u, v) zero-frequency point is moved to frequency spectrum
The center of figure;
Step 2.2, the product G (u, v) for calculating Gaussian function filter function H (u, v) and F (u, v), by the zero of frequency spectrum G (u, v)
Frequency is moved back into the upper left position of spectrogram, wherein
Gaussian filter function representation is:
<mrow>
<mi>H</mi>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>,</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mn>1</mn>
<mo>-</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mo>&lsqb;</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>-</mo>
<mfrac>
<mi>M</mi>
<mn>2</mn>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>v</mi>
<mo>-</mo>
<mfrac>
<mi>N</mi>
<mn>2</mn>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>&rsqb;</mo>
<mo>/</mo>
<mn>2</mn>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</msup>
</mrow>
Wherein, u, v are frequency domain variable, and M, N are respectively the level of image, vertical pixel number, and σ is the sigma of Gaussian function
Parameter;
Frequency spectrum designation is:
G (u, v)=H (u, v) × F (u, v)
Step 2.3, to G (u, v) carry out inverse discrete fourier transform obtain g (x, y), take g (x, y) real part as filtered
Result images.
3. the image edge extraction method based on improvement Canny operators according to described by claim 1, it is characterised in that institute
State step 3 and ask concretely comprising the following steps for greatest gradient figure:
Step 3.1,45 ° using prewitt, 135 °, level, the template in vertical 4 directions and the filtered gray-scale map of completion
Convolution is carried out respectively;
Step 3.2, value maximum in each 4 convolution results of pixel is taken as Grad, the corresponding direction of maximum is maximum
Gradient direction, produces greatest gradient image.
4. the image edge extraction method based on improvement Canny operators according to described by claim 1, it is characterised in that institute
State step 4 calculate Low threshold specific method be:
If greatest gradient figure is Q (x, y), Q (x, y) tonal range is [0, L-1], show that threshold value is T using the lowest point method, then threshold
Q (x, y) is divided into 2 scopes [0, T] and [T, L-1] by value T, if the pixel count that gray scale is i in image is ni, it is in tonal range
[0, T] interior total pixel number is:
<mrow>
<msub>
<mi>N</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>T</mi>
</munderover>
<msub>
<mi>n</mi>
<mi>i</mi>
</msub>
</mrow>
The probability that each gray scale occurs is:
<mrow>
<msub>
<mi>p</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>n</mi>
<mi>i</mi>
</msub>
<mi>N</mi>
</mfrac>
</mrow>
In [0, T], threshold value T is used1It is divided into 2 class C0And C1, C0By [0, T1- 1] constitute, C1By [T1, T] and composition, then region C0
And C1Probability be respectively:
<mrow>
<msub>
<mi>P</mi>
<mn>0</mn>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<msub>
<mi>T</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>p</mi>
<mi>i</mi>
</msub>
</mrow>
P1=1-P0
C0And C1Average gray be respectively:
<mrow>
<msub>
<mi>&mu;</mi>
<mn>0</mn>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msub>
<mi>P</mi>
<mn>0</mn>
</msub>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>T</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>ip</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>&mu;</mi>
<mrow>
<mo>(</mo>
<mi>T</mi>
<mo>)</mo>
</mrow>
</mrow>
<msub>
<mi>P</mi>
<mn>0</mn>
</msub>
</mfrac>
</mrow>
<mrow>
<msub>
<mi>&mu;</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msub>
<mi>P</mi>
<mn>0</mn>
</msub>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<msub>
<mi>T</mi>
<mi>i</mi>
</msub>
</mrow>
<mi>T</mi>
</munderover>
<msub>
<mi>ip</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>&mu;</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mrow>
<mo>(</mo>
<mi>T</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mn>1</mn>
<mo>-</mo>
<msub>
<mi>P</mi>
<mn>0</mn>
</msub>
</mrow>
</mfrac>
</mrow>
Wherein, μ is the average gray of [0, T]:
μ=P0μ0+P1μ1
The population variance in two regions is:
σB 2=P0(μ0-μ)2+P1(μ1-μ)2=P0P1(μ0-μ1)2
T is allowed [0, T1- 1] interior value successively, makes σB 2Maximum T values are the optimal selections of Low threshold.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710300398.4A CN107220988B (en) | 2017-04-30 | 2017-04-30 | Part image edge extraction method based on improved canny operator |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710300398.4A CN107220988B (en) | 2017-04-30 | 2017-04-30 | Part image edge extraction method based on improved canny operator |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107220988A true CN107220988A (en) | 2017-09-29 |
CN107220988B CN107220988B (en) | 2020-09-18 |
Family
ID=59943716
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710300398.4A Expired - Fee Related CN107220988B (en) | 2017-04-30 | 2017-04-30 | Part image edge extraction method based on improved canny operator |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107220988B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108109155A (en) * | 2017-11-28 | 2018-06-01 | 东北林业大学 | A kind of automatic threshold edge detection method based on improvement Canny |
CN109101976A (en) * | 2018-07-10 | 2018-12-28 | 温州大学 | A kind of detection method of arc extinguishing grid pieces surface defect |
CN110378866A (en) * | 2019-05-22 | 2019-10-25 | 中国水利水电科学研究院 | A kind of canal lining breakage image recognition methods based on unmanned plane inspection |
CN111027567A (en) * | 2019-10-30 | 2020-04-17 | 四川轻化工大学 | Edge extraction method based on algorithm learning |
CN111062317A (en) * | 2019-12-16 | 2020-04-24 | 中国计量大学上虞高等研究院有限公司 | Method and system for cutting edges of scanned document |
CN111127498A (en) * | 2019-12-12 | 2020-05-08 | 重庆邮电大学 | Canny edge detection method based on edge self-growth |
CN111681256A (en) * | 2020-05-07 | 2020-09-18 | 浙江大华技术股份有限公司 | Image edge detection method and device, computer equipment and readable storage medium |
CN112381850A (en) * | 2020-12-04 | 2021-02-19 | 亿嘉和科技股份有限公司 | Cabinet surface circular target automatic segmentation method, system, device and storage medium |
CN112488940A (en) * | 2020-11-30 | 2021-03-12 | 哈尔滨市科佳通用机电股份有限公司 | Method for enhancing image edge of railway locomotive component |
CN112508795A (en) * | 2020-11-26 | 2021-03-16 | 珠海格力电器股份有限公司 | Image contour detection method and device |
CN112669265A (en) * | 2020-12-17 | 2021-04-16 | 华中科技大学 | Method for realizing surface defect detection based on Fourier transform and image gradient characteristics |
CN113850809A (en) * | 2021-12-01 | 2021-12-28 | 武汉飞恩微电子有限公司 | Method for detecting appearance defects of protective film of chip resistor |
CN114331923A (en) * | 2022-03-11 | 2022-04-12 | 中国空气动力研究与发展中心低速空气动力研究所 | Improved Canny algorithm-based bubble contour extraction method in ice structure |
CN115131387A (en) * | 2022-08-25 | 2022-09-30 | 山东鼎泰新能源有限公司 | Gasoline engine spray wall collision parameter automatic extraction method and system based on image processing |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110007982A1 (en) * | 2009-07-13 | 2011-01-13 | Yeping Su | Methods and Systems for Reducing Compression Artifacts |
CN102521836A (en) * | 2011-12-15 | 2012-06-27 | 江苏大学 | Edge detection method based on gray-scale image of specific class |
CN104700421A (en) * | 2015-03-27 | 2015-06-10 | 中国科学院光电技术研究所 | Adaptive threshold edge detection algorithm based on canny |
CN105354815A (en) * | 2015-09-12 | 2016-02-24 | 沈阳理工大学 | Flat micro-part based accurate identification and positioning method |
-
2017
- 2017-04-30 CN CN201710300398.4A patent/CN107220988B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110007982A1 (en) * | 2009-07-13 | 2011-01-13 | Yeping Su | Methods and Systems for Reducing Compression Artifacts |
CN102521836A (en) * | 2011-12-15 | 2012-06-27 | 江苏大学 | Edge detection method based on gray-scale image of specific class |
CN104700421A (en) * | 2015-03-27 | 2015-06-10 | 中国科学院光电技术研究所 | Adaptive threshold edge detection algorithm based on canny |
CN105354815A (en) * | 2015-09-12 | 2016-02-24 | 沈阳理工大学 | Flat micro-part based accurate identification and positioning method |
Non-Patent Citations (1)
Title |
---|
朱晓林 等: ""基于Canny算子的机械零件图像分块自适应边缘检测"", 《江南大学学报(自然科学版)》 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108109155A (en) * | 2017-11-28 | 2018-06-01 | 东北林业大学 | A kind of automatic threshold edge detection method based on improvement Canny |
CN109101976B (en) * | 2018-07-10 | 2021-11-30 | 温州大学 | Method for detecting surface defects of arc-extinguishing grid plate |
CN109101976A (en) * | 2018-07-10 | 2018-12-28 | 温州大学 | A kind of detection method of arc extinguishing grid pieces surface defect |
CN110378866A (en) * | 2019-05-22 | 2019-10-25 | 中国水利水电科学研究院 | A kind of canal lining breakage image recognition methods based on unmanned plane inspection |
CN111027567A (en) * | 2019-10-30 | 2020-04-17 | 四川轻化工大学 | Edge extraction method based on algorithm learning |
CN111127498A (en) * | 2019-12-12 | 2020-05-08 | 重庆邮电大学 | Canny edge detection method based on edge self-growth |
CN111062317A (en) * | 2019-12-16 | 2020-04-24 | 中国计量大学上虞高等研究院有限公司 | Method and system for cutting edges of scanned document |
CN111681256A (en) * | 2020-05-07 | 2020-09-18 | 浙江大华技术股份有限公司 | Image edge detection method and device, computer equipment and readable storage medium |
CN111681256B (en) * | 2020-05-07 | 2023-08-18 | 浙江大华技术股份有限公司 | Image edge detection method, image edge detection device, computer equipment and readable storage medium |
CN112508795A (en) * | 2020-11-26 | 2021-03-16 | 珠海格力电器股份有限公司 | Image contour detection method and device |
CN112488940A (en) * | 2020-11-30 | 2021-03-12 | 哈尔滨市科佳通用机电股份有限公司 | Method for enhancing image edge of railway locomotive component |
CN112381850A (en) * | 2020-12-04 | 2021-02-19 | 亿嘉和科技股份有限公司 | Cabinet surface circular target automatic segmentation method, system, device and storage medium |
CN112669265A (en) * | 2020-12-17 | 2021-04-16 | 华中科技大学 | Method for realizing surface defect detection based on Fourier transform and image gradient characteristics |
CN113850809A (en) * | 2021-12-01 | 2021-12-28 | 武汉飞恩微电子有限公司 | Method for detecting appearance defects of protective film of chip resistor |
CN114331923A (en) * | 2022-03-11 | 2022-04-12 | 中国空气动力研究与发展中心低速空气动力研究所 | Improved Canny algorithm-based bubble contour extraction method in ice structure |
CN114331923B (en) * | 2022-03-11 | 2022-05-13 | 中国空气动力研究与发展中心低速空气动力研究所 | Improved Canny algorithm-based bubble profile extraction method in ice structure |
CN115131387A (en) * | 2022-08-25 | 2022-09-30 | 山东鼎泰新能源有限公司 | Gasoline engine spray wall collision parameter automatic extraction method and system based on image processing |
CN115131387B (en) * | 2022-08-25 | 2023-01-24 | 山东鼎泰新能源有限公司 | Gasoline engine spray wall collision parameter automatic extraction method and system based on image processing |
Also Published As
Publication number | Publication date |
---|---|
CN107220988B (en) | 2020-09-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107220988A (en) | Based on the parts image edge extraction method for improving canny operators | |
CN107808378B (en) | Method for detecting potential defects of complex-structure casting based on vertical longitudinal and transverse line profile features | |
CN106886977B (en) | Multi-image automatic registration and fusion splicing method | |
CN103984961B (en) | A kind of image detecting method for being used to detect underbody foreign matter | |
CN103955922B (en) | Method for detecting flaws of printed fabric based on Gabor filter | |
CN108022233A (en) | A kind of edge of work extracting method based on modified Canny operators | |
CN113034452B (en) | Weldment contour detection method | |
CN112614062B (en) | Colony counting method, colony counting device and computer storage medium | |
CN105139391B (en) | A kind of haze weather traffic image edge detection method | |
CN103279957A (en) | Method for extracting remote sensing image interesting area based on multi-scale feature fusion | |
CN101079149A (en) | Noise-possessing movement fuzzy image restoration method based on radial basis nerve network | |
CN103679656B (en) | A kind of Automated sharpening of images method | |
CN103778611B (en) | Utilize the switch weight vectors median filter method of rim detection | |
CN110414308B (en) | Target identification method for dynamic foreign matters on power transmission line | |
CN109118442B (en) | Image enhancement method based on Sobel operator filtering | |
CN109472788A (en) | A kind of scar detection method on airplane riveting surface | |
CN105678245A (en) | Target position identification method based on Haar features | |
CN116664457B (en) | Image processing method for enhancing denoising | |
Hou et al. | Welding image edge detection and identification research based on canny operator | |
CN107545549B (en) | Method for estimating scattered focus point spread function based on one-dimensional spectrum curve | |
CN114332081B (en) | Textile surface abnormity determination method based on image processing | |
CN112801141B (en) | Heterogeneous image matching method based on template matching and twin neural network optimization | |
CN107451608B (en) | SAR image non-reference quality evaluation method based on multi-view amplitude statistical characteristics | |
CN113781413A (en) | Electrolytic capacitor positioning method based on Hough gradient method | |
CN115063679B (en) | Pavement quality assessment method based on deep learning |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200918 |