CN110473215A - A kind of image partition method for overhead distribution monitoring scene - Google Patents
A kind of image partition method for overhead distribution monitoring scene Download PDFInfo
- Publication number
- CN110473215A CN110473215A CN201910770167.9A CN201910770167A CN110473215A CN 110473215 A CN110473215 A CN 110473215A CN 201910770167 A CN201910770167 A CN 201910770167A CN 110473215 A CN110473215 A CN 110473215A
- Authority
- CN
- China
- Prior art keywords
- image
- gray
- threshold value
- initial values
- threshold
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000012544 monitoring process Methods 0.000 title claims abstract description 28
- 238000005192 partition Methods 0.000 title claims abstract description 14
- 238000003709 image segmentation Methods 0.000 claims abstract description 22
- 238000011946 reduction process Methods 0.000 claims abstract description 6
- 230000005540 biological transmission Effects 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000005764 inhibitory process Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 abstract description 7
- 230000000694 effects Effects 0.000 description 18
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003292 diminished effect Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
Classifications
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- 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
Abstract
The invention discloses a kind of image partition methods for overhead distribution monitoring scene, it includes: that (1) obtains original image and carries out gray processing processing;(2) image enhancement processing is carried out to gray level image;(3) noise reduction process is carried out to gray level image;(4) dimension-reduction treatment is carried out to two-dimentional Otsu algorithm;(5) the threshold value scope of initial values of the one-dimensional Otsu algorithm after dimensionality reduction is set separately;(6) in threshold value scope of initial values, optimal threshold S and T are solved respectively using recurrence formula to two one-dimensional Otsu algorithms;(7) image segmentation is carried out to original image according to the optimal threshold of step (6);The method for solving segmentation of the prior art for overhead distribution monitoring image, the complexity of calculating and calculating time are greatly improved, it is difficult to meet the technical problems such as application demand.
Description
Technical field
The invention belongs to image processing techniques more particularly to a kind of image segmentations for overhead distribution monitoring scene
Method.
Background technique
With the rise of image monitoring technology and the application of transmission line of electricity image monitor, overhead distribution is real-time
Image monitoring is also gradually taken seriously.However, overhead distribution is generally in regional in the majority, the point being related to such as small towns, rural area
More and wide, common image monitoring mode needs a large amount of manpower to go voluntarily to analyze identification, it is difficult to inerrably find in time different
It often occurs as playing the effect checked and prevented accident early.Therefore, overhead distribution monitoring image Weigh sensor technology is
The feasible way to solve the above problems, and the image segmentation problem for solving route is then crucial.
However, from the point of view of the image effect actually taken, prison is presented in monitoring image in overhead distribution monitoring scene
Survey target and background ratio great disparity and the relatively low feature of signal-to-noise ratio, and due to the requirement of power network monitoring reliability and real-time,
The method of image segmentation will also reach higher standard on effect precision and time efficiency, but, currently used several classes
Image segmentation algorithm, the method such as based on threshold value, the method based on region, the method based on edge, region are combined with edge
Method and multi-scale division method etc. can not meet the image segmentation suitable for overhead distribution monitoring scene well
The requirement of method.Preferably method is that image segmentation is carried out using two dimension Otsu algorithm, and two-dimentional Otsu algorithm was both utilized
The pixel gray level information of image, while the domain level constraints relevant information of pixel, for low contrast, low noise being utilized again
The target of ratio, segmentation effect still with higher;But, the complexity and calculating time calculated is greatly improved, it is difficult to full
Sufficient application demand.
Summary of the invention
The technical problem to be solved by the present invention is providing a kind of image segmentation side for overhead distribution monitoring scene
Method, the method to solve segmentation of the prior art for overhead distribution monitoring image, the complexity of calculating and calculating time
It is greatly improved, it is difficult to meet the technical problems such as application demand.
Technical solution of the present invention:
A kind of image partition method for overhead distribution monitoring scene, it includes:
(1) it obtains original image and carries out gray processing processing;
(2) image enhancement processing is carried out to gray level image;
(3) noise reduction process is carried out to gray level image;
(4) dimension-reduction treatment is carried out to two-dimentional Otsu algorithm;
(5) the threshold value scope of initial values of the one-dimensional Otsu algorithm after dimensionality reduction is set separately;
(6) in threshold value scope of initial values, gray value is solved respectively most using recurrence formula to two one-dimensional Otsu algorithms
Good threshold value S and field gray value optimal threshold T;
(7) image point is carried out to original image according to the gray value optimal threshold S of step (6) and field gray value optimal threshold T
It cuts.
The image enhancement processing carries out the overhead transmission line image information for extracting target in the picture for image segmentation
It highlights and emphasizes, background image is subjected to reduction inhibition.
The image noise reduction is handled using wavelet adaptive threshold method.
Carrying out dimension-reduction treatment to two-dimentional Otsu algorithm is that conventional two-dimensional Otsu algorithm dimensionality reduction is formed to two one-dimensional Otsu to calculate
Method obtains a gray value optimal threshold S according to the pixel gray value f (x, y) of image, according to the pixel neighborhood of a point ash
It spends mean value g (x, y) and obtains a field gray value optimal threshold T.
The method of the threshold value scope of initial values of one-dimensional Otsu algorithm after dimensionality reduction is set separately are as follows:
The threshold value scope of initial values of image when (5a) goes to position determining low resolution by the way of reducing image resolution ratio;
Original image is saved as low-resolution image by (5b), and the threshold of original image is calculated by the threshold value scope of initial values of low-resolution image
It is worth scope of initial values.
By low-resolution image threshold value scope of initial values calculate original image threshold value scope of initial values be first original image is saved as it is low
The picture of resolution ratio, the ratio between original image resolution ratio and low resolution are k, k > 1, in the threshold value scope of initial values for determining low resolution figure
Afterwards, using the threshold value scope of initial values of original image threshold value scope of initial values and low resolution figure, there are k times of relationships to determine original image threshold value initial value
Range.
The recurrence formula specifically:
If s ∈ [0, L], piIt is the histogram pixel number of the i-stage after normalization:
(1) formula is utilized, initial value p is obtainedb(0)、μb(0)、pf(0)、μf(0);
(2) when s is incremented by, then have:
pb(s+1)=pb(s)+ps+1
pf(s+1)=pf(s)+ps+1
(3) until s=L-1, after circulation terminates, according to formulaObtain gray scale
It is worth optimal threshold S;
Wherein, in original image background and target inter-class variance σ2 B(s) are as follows:
pbIt (s) is background probability of occurrence, μbIt (s) is background probability of occurrence mean value, pfIt (s) is target probability of occurrence, μf(s)
For target probability of occurrence mean value.
When performing image segmentation:
Assuming that certain original image has L gray level, image total pixel number is N, and grey scale pixel value i, gray scale is pixel i's
Number is ni, then available each gray level occurs probability are as follows: pi=ni/N;In image segmentation, used according to image gray levels
Gray scale is divided into background C by gray threshold s0=(0,1,2 ..., s) and target C1(s+1, s+2 ..., L-1) two classes, the two occur
Probability pb(s) and pf(s) it is respectively as follows:
Therefore, C0And C1Mean μb(s)、μf(s) it is respectively as follows:
The average gray of original image are as follows:
Beneficial effects of the present invention
The present invention improves two-dimentional Otsu, computationally carries out quick locating threshold on the basis of two Otsu algorithms
Range and dimension-reduction treatment had both remained two-dimentional algorithm originally for the target of low contrast, low signal-to-noise ratio, still with higher point
The advantages of cutting effect, and two-dimentional Otsu algorithm is compensated for, calculate complicated, the defect of processing time length.
The present invention has fully considered that the characteristics of image having under overhead line distribution line monitoring scene, the algorithm proposed exist
There is good application effect under this scene.
The method for solving segmentation of the prior art for overhead distribution monitoring image, the complexity and calculating of calculating
Time is greatly improved, it is difficult to meet the technical problems such as application demand.
Detailed description of the invention
Fig. 1 is the implementation flow chart of image partition method provided in an embodiment of the present invention;
Fig. 2 is the overhead distribution figure used in the present invention for having converted grayscale image;
Fig. 3 is that the present invention carries out image enhancement effects figure to the overhead distribution figure of the processing of gray processing;
Fig. 4 is that Fig. 3 carries out the effect picture divided after noise reduction process using conventional two-dimensional Otsu algorithm;
Fig. 5 present invention is used for the effect picture of overhead distribution.
Specific embodiment
It is shown in Figure 1: a kind of image partition method for overhead distribution monitoring scene, which is characterized in that packet
Include following steps:
(1) it obtains original image and gray processing processing is carried out to it;
(2) image enhancement processing is carried out to gray level image;
(3) noise reduction process is carried out to gray level image;
(4) dimension-reduction treatment is carried out to traditional two-dimentional Otsu algorithm;
(5) the threshold value scope of initial values of the one-dimensional Otsu algorithm after dimensionality reduction is set separately;
The threshold value initial value of image when (5a) first quickly goes to position determining low resolution by the way of reducing image resolution ratio
Range;
Original image is saved as low-resolution image by (5b), calculates estimation original image by the threshold value scope of initial values of low-resolution image
Threshold value scope of initial values;
(6) more accurate in last determination on the basis of the threshold value scope of initial values estimated and there are certain nargin
In threshold value scope of initial values, gray value optimal threshold S and field are solved respectively using recurrence formula to two one-dimensional Otsu algorithms
Gray value optimal threshold T;
(7) image segmentation is carried out to original image according to the optimal threshold of step (6).
The image enhancement processing by accounting very little in the picture but for image segmentation extract target overhead transmission line
Image information highlight emphasizing, uninterested background image is carried out reduction inhibition, improving image quality is reinforced image and sentenced
It reads and identifies, the effect of segmentation.As shown in figure 3, Fig. 3 is to carry out image enhancement on the basis of the processing of Fig. 2 gray processing, it is prominent to increase
Overhead line in strong figure, weakens other backgrounds.
The described image noise reduction processing using can reserved line distal end details very well wavelet adaptive threshold method.
Fig. 4 is the effect picture for carrying out noise reduction process on the basis of Fig. 3 and carrying out image segmentation using two Otsu algorithms of tradition, can be seen
Out in the figure after noise reduction and dividing processing, substantially by out and original overhead line at the segmentation extraction of the figure of overhead line
Noise spot around road is also more to be eliminated, it is seen that noise reduction effect is more effective.
Described is that conventional two-dimensional Otsu algorithm dimensionality reduction is formed two to traditional two-dimentional Otsu algorithm progress dimension-reduction treatment
A one-dimensional Otsu algorithm obtains a gray value optimal threshold S according to the pixel gray value f (x, y) of image, according to the picture
Plain neighborhood of a point gray average g (x, y) obtains field gray average field gray value optimal threshold T.
When performing image segmentation, it is assumed that certain original image has L gray level, and image total pixel number is N, grey scale pixel value
For i, the number that gray scale is pixel i is ni, then available each gray level occurs probability are as follows: pi=ni/N;In image segmentation
In, gray scale is divided into background C with gray value optimal threshold S according to image gray levels0=(0,1,2 ..., s) and target C1(s+1,
S+2 ..., L-1) two classes, the Probability p that the two occursb(s) and pf(s) it is respectively as follows:
Therefore, C0And C1Mean μb(s)、μf(s) it is respectively as follows:
The average gray of original image are as follows:
The recurrence formula specifically:
If s ∈ [0, L], piIt is the histogram pixel number of the i-stage after normalization:
(1) formula is utilized, initial value p is obtainedb(0)、μb(0)、pf(0)、μf(0);
(2) when s is incremented by, then have:
pb(s+1)=pb(s)+ps+1
pf(s+1)=pf(s)+ps+1
(3) until s=L-1, after circulation terminates, according to formulaObtain gray scale
It is worth optimal threshold S.
Wherein, in original image background and target inter-class variance σ2 B(s) are as follows:
Background probability of occurrence pb(s), background goes out
Existing mathematical expectation of probability μb(s), target probability of occurrence pf(s), target probability of occurrence mean μf(s).It is recycled every time in one-dimensional Otsu algorithm
When be required to this 4 parameters and calculated: for example, when s is incremented by, background area will increase a gray level, and target area
Reduce by a gray level, only s value is the gray level newly increased in this process, and the preceding result once calculated, which can be used in, to be worked as
In secondary calculating.
Similarly, field gray value optimal threshold T can be acquired.
The threshold value scope of initial values by low-resolution image calculates that the threshold value scope of initial values of estimation original image is first by original image
The picture of low resolution is saved as, the ratio between original image resolution ratio and low resolution are k (k > 1), in quick determining low resolution figure
After threshold value scope of initial values, using the threshold value scope of initial values of original image threshold value scope of initial values and low resolution figure, there is also k times of relationship is true
Determine original image threshold value scope of initial values.
The resolution ratio of image is bigger, and the pixel number for including is more, takes time when calculating threshold value scope of initial values more
It is more, therefore the threshold value scope of initial values for using the image of low resolution quickly to calculate low-resolution image is such as [a, b], then estimates roughly
The threshold value scope of initial values for calculating original image is [ka, kb], so improves the speed for calculating threshold value scope of initial values, also more accurate contracting
Subtract threshold range, reduces the recursion number using recurrence formula, further improve the efficiency of image segmentation.
It is described on the basis of the threshold value scope of initial values estimated and there are certain nargin, such as [ka+ α, kb+ β], margin value
α, β can be manually set according to the shooting environmental of monitoring image, actual segmentation effect.Such as according to last point in some scene
Cutting may need from the point of view of effect α value to be that 10 effects are best, and may be needed in the scene at another place α value be 25 effects most
It is good.
Fig. 5 carries out the effect picture of image segmentation, comparison diagram 5 and Fig. 4 with method proposed by the invention, it can be seen that from
Dividing method proposed by the invention and two dimension Otsu algorithm are essentially the same in effect, but from the processing time,
Method proposed by the invention carries out dimensionality reduction to two-dimentional Otsu algorithm, decomposes the one-dimensional Otsu in position two and solves, the complicated journey of calculating
Degree is greatly diminished, and the processing time is wanted, it will be apparent that due to traditional two-dimentional Otsu algorithm.
Claims (8)
1. a kind of image partition method for overhead distribution monitoring scene, it includes:
(1) it obtains original image and carries out gray processing processing;
(2) image enhancement processing is carried out to gray level image;
(3) noise reduction process is carried out to gray level image;
(4) dimension-reduction treatment is carried out to two-dimentional Otsu algorithm;
(5) the threshold value scope of initial values of the one-dimensional Otsu algorithm after dimensionality reduction is set separately;
(6) in threshold value scope of initial values, the best threshold of gray value is solved respectively using recurrence formula to two one-dimensional Otsu algorithms
Value S and field gray value optimal threshold T;
(7) image segmentation is carried out to original image according to the gray value optimal threshold S of step (6) and field gray value optimal threshold T.
2. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In: the overhead transmission line image information for extracting target in the picture for image segmentation highlight by force by the image enhancement processing
It adjusts, background image is subjected to reduction inhibition.
3. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In: the image noise reduction is handled using wavelet adaptive threshold method.
4. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In: carrying out dimension-reduction treatment to two-dimentional Otsu algorithm is that conventional two-dimensional Otsu algorithm dimensionality reduction is formed two one-dimensional Otsu algorithms, i.e.,
A gray value optimal threshold S is obtained according to the pixel gray value f (x, y) of image, it is equal according to the pixel neighborhood of a point gray scale
Value g (x, y) obtains a field gray value optimal threshold T.
5. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In: the method for the threshold value scope of initial values of the one-dimensional Otsu algorithm after dimensionality reduction is set separately are as follows:
The threshold value scope of initial values of image when (5a) goes to position determining low resolution by the way of reducing image resolution ratio;
Original image is saved as low-resolution image by (5b), at the beginning of the threshold value for calculate original image as the threshold value scope of initial values of low-resolution image
It is worth range.
6. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In: the threshold value scope of initial values by low-resolution image calculates that the threshold value scope of initial values of original image is that original image is first saved as low point
The picture of resolution, the ratio between original image resolution ratio and low resolution be k, k > 1, after determining the threshold value scope of initial values of low resolution figure,
Using the threshold value scope of initial values of original image threshold value scope of initial values and low resolution figure, there are k times of relationships to determine original image threshold value initial value model
It encloses.
7. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In:
The recurrence formula specifically:
If s ∈ [0, L], piIt is the histogram pixel number of the i-stage after normalization:
(1) formula is utilized, initial value p is obtainedb(0)、μb(0)、pf(0)、μf(0);
(2) when s is incremented by, then have:
pb(s+1)=pb(s)+ps+1
pf(s+1)=pf(s)+ps+1
(3) until s=L-1, after circulation terminates, according to formulaObtain gray value most
Good threshold value S;
Wherein, in original image background and target inter-class variance σ2 B(s) are as follows:
pbIt (s) is background probability of occurrence, μbIt (s) is background probability of occurrence mean value, pfIt (s) is target probability of occurrence, μfIt (s) is target
Probability of occurrence mean value.
8. a kind of image partition method for overhead distribution monitoring scene according to claim 1, feature exist
In: when performing image segmentation:
Assuming that certain original image has L gray level, image total pixel number is N, and grey scale pixel value i, gray scale is the number of pixel i
For ni, then available each gray level occurs probability are as follows: pi=ni/N;In image segmentation, according to image gray levels gray scale
Gray scale is divided into background C by threshold value s0=(0,1,2 ..., s) and target C1(s+1, s+2 ..., L-1) two classes, the two occur general
Rate pb(s) and pf(s) it is respectively as follows:
Therefore, C0And C1Mean μb(s)、μf(s) it is respectively as follows:
The average gray of original image are as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910770167.9A CN110473215A (en) | 2019-08-20 | 2019-08-20 | A kind of image partition method for overhead distribution monitoring scene |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910770167.9A CN110473215A (en) | 2019-08-20 | 2019-08-20 | A kind of image partition method for overhead distribution monitoring scene |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110473215A true CN110473215A (en) | 2019-11-19 |
Family
ID=68512934
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910770167.9A Pending CN110473215A (en) | 2019-08-20 | 2019-08-20 | A kind of image partition method for overhead distribution monitoring scene |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110473215A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111311884A (en) * | 2020-01-13 | 2020-06-19 | 徐琳 | Signal identification system and method based on content detection |
CN111932563A (en) * | 2020-09-23 | 2020-11-13 | 平安科技(深圳)有限公司 | Picture region segmentation method and device, electronic equipment and storage medium |
CN113538499A (en) * | 2021-08-20 | 2021-10-22 | 深圳市爱协生科技有限公司 | Image threshold segmentation method and device, computer equipment and storage medium |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101894372A (en) * | 2010-08-03 | 2010-11-24 | 新疆大学 | New noise-containing remote sensing image segmentation method |
US8116522B1 (en) * | 2008-08-25 | 2012-02-14 | The United States Of America As Represented By The Secretary Of The Navy | Ship detection system and method from overhead images |
US20170061256A1 (en) * | 2014-05-14 | 2017-03-02 | Cellavision Ab | Segmentation based image transform |
CN106960438A (en) * | 2017-03-25 | 2017-07-18 | 安徽继远软件有限公司 | Method for recognizing impurities to transmission line of electricity is converted based on Hough straight line |
CN107590501A (en) * | 2017-07-30 | 2018-01-16 | 华南理工大学 | A kind of improved Canny adaptive edge extracting methods based on Otsu |
CN108335294A (en) * | 2018-02-05 | 2018-07-27 | 贵州电网有限责任公司 | The power distribution room abnormality image-recognizing method of complex condition |
CN108921868A (en) * | 2018-07-02 | 2018-11-30 | 中央民族大学 | A kind of improved Otsu threshold segmentation method |
CN109461148A (en) * | 2018-10-30 | 2019-03-12 | 兰州交通大学 | Steel rail defect based on two-dimentional Otsu divides adaptive fast algorithm |
-
2019
- 2019-08-20 CN CN201910770167.9A patent/CN110473215A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8116522B1 (en) * | 2008-08-25 | 2012-02-14 | The United States Of America As Represented By The Secretary Of The Navy | Ship detection system and method from overhead images |
CN101894372A (en) * | 2010-08-03 | 2010-11-24 | 新疆大学 | New noise-containing remote sensing image segmentation method |
US20170061256A1 (en) * | 2014-05-14 | 2017-03-02 | Cellavision Ab | Segmentation based image transform |
CN106960438A (en) * | 2017-03-25 | 2017-07-18 | 安徽继远软件有限公司 | Method for recognizing impurities to transmission line of electricity is converted based on Hough straight line |
CN107590501A (en) * | 2017-07-30 | 2018-01-16 | 华南理工大学 | A kind of improved Canny adaptive edge extracting methods based on Otsu |
CN108335294A (en) * | 2018-02-05 | 2018-07-27 | 贵州电网有限责任公司 | The power distribution room abnormality image-recognizing method of complex condition |
CN108921868A (en) * | 2018-07-02 | 2018-11-30 | 中央民族大学 | A kind of improved Otsu threshold segmentation method |
CN109461148A (en) * | 2018-10-30 | 2019-03-12 | 兰州交通大学 | Steel rail defect based on two-dimentional Otsu divides adaptive fast algorithm |
Non-Patent Citations (3)
Title |
---|
FU-BIN WANG等: "Spot image ablated by femtosecond laser segmentation and feature clustering after dimension reduction reconstruction", 《OPTIK》 * |
左攀攀: "基于FPGA的车牌号识别技术的研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
田岩等: "基于局部方差的多分辨率图像分割方法", 《系统工程与电子技术》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111311884A (en) * | 2020-01-13 | 2020-06-19 | 徐琳 | Signal identification system and method based on content detection |
CN111932563A (en) * | 2020-09-23 | 2020-11-13 | 平安科技(深圳)有限公司 | Picture region segmentation method and device, electronic equipment and storage medium |
CN111932563B (en) * | 2020-09-23 | 2021-07-06 | 平安科技(深圳)有限公司 | Picture region segmentation method and device, electronic equipment and storage medium |
CN113538499A (en) * | 2021-08-20 | 2021-10-22 | 深圳市爱协生科技有限公司 | Image threshold segmentation method and device, computer equipment and storage medium |
CN113538499B (en) * | 2021-08-20 | 2024-01-19 | 深圳市爱协生科技股份有限公司 | Image threshold segmentation method, device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110428433B (en) | Canny edge detection algorithm based on local threshold | |
CN110473215A (en) | A kind of image partition method for overhead distribution monitoring scene | |
CN110033446B (en) | Enhanced image quality evaluation method based on twin network | |
CN109461148A (en) | Steel rail defect based on two-dimentional Otsu divides adaptive fast algorithm | |
CN110415208B (en) | Self-adaptive target detection method and device, equipment and storage medium thereof | |
CN105761260B (en) | A kind of skin image affected part dividing method | |
CN104463870A (en) | Image salient region detection method | |
CN111402261B (en) | Improved skull segmentation method based on Hessian matrix enhancement | |
CN115908154B (en) | Video later-stage particle noise removing method based on image processing | |
CN105719275A (en) | Parallel combination image defect segmentation method | |
CN103020953A (en) | Segmenting method of fingerprint image | |
CN112330561B (en) | Medical image segmentation method based on interactive foreground extraction and information entropy watershed | |
CN110599422B (en) | Weighted average salt and pepper noise reduction method based on edge protection | |
CN109784216B (en) | Vehicle-mounted thermal imaging pedestrian detection Rois extraction method based on probability map | |
CN111161300A (en) | Nilback image segmentation method based on improved Otsu method | |
CN103207997A (en) | Kernel density estimation-based license plate character segmentation method | |
CN110211150B (en) | Real-time visual target identification method with scale coordination mechanism | |
CN108460383B (en) | Image significance refinement method based on neural network and image segmentation | |
CN114998186B (en) | Method and system for detecting surface scab defect of copper starting sheet based on image processing | |
CN111145197A (en) | Accurate turnout switch machine notch edge positioning method based on histogram and local gradient | |
CN112102350A (en) | Secondary image segmentation method based on Otsu and Tsallis entropy | |
CN111507933B (en) | DIBR synthetic image quality evaluation method based on cavity and contour amplification | |
CN111060887B (en) | Gm-APD laser radar low signal-to-noise ratio echo data signal extraction method based on concave-convex search | |
CN110288604B (en) | Image segmentation method and device based on K-means | |
CN108961290B (en) | Adaptive image edge detection method based on Ratio operator of Otsu |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191119 |
|
RJ01 | Rejection of invention patent application after publication |