CN105869175A - Image segmentation method and system - Google Patents
Image segmentation method and system Download PDFInfo
- Publication number
- CN105869175A CN105869175A CN201610250328.8A CN201610250328A CN105869175A CN 105869175 A CN105869175 A CN 105869175A CN 201610250328 A CN201610250328 A CN 201610250328A CN 105869175 A CN105869175 A CN 105869175A
- Authority
- CN
- China
- Prior art keywords
- pixel
- super
- target image
- image
- segmentation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The embodiment of the invention provides an image segmentation method and system. The method comprises the steps of acquiring a target image to be processed; calculating the grey level histogram of the target image; calculating the occurring probability of each pixel value covered by pixel points in the target image based on the grey level histogram; calculating the complexity of the target image based on the number of the pixel points of the target image and the occurring probability of each pixel value by means of a preset complexity calculation formula; calculating the superpixel segmentation number based on the complexity of the target image by means of a preset superpixel segmentation number calculation formula; segmenting the target image based on the superpixel segmentation number. By the adoption of the method and system, the superpixel number required by image segmentation can be determined automatically according to the specific condition of the image, so that image processing complexity is reduced and insufficient segmentation is avoided.
Description
Technical field
The present invention relates to technical field of image processing, particularly relate to a kind of image partition method and system.
Background technology
At present, can by mobile terminal by camera collection to image be applied to multiple fields, such as:
Mobile terminal can be applied to location technology by camera collection to image, will collect by mobile terminal
Image mate with database images, and then obtain position data.But, due to existing mobile whole
Resolution ratio of camera head in end has generally exceeded 12,000,000 pixels, and the image data amount i.e. collected is relatively big,
I.e. need data volume to be processed relatively big, and relatively limited at computing capability and the memory source of mobile terminal,
Make widely mobile terminal can not to be collected image at present and be applied to location technology.
And usually through super-pixel segmentation Turbopixels algorithm, image is processed at present, i.e. image is entered
Row super-pixel is split, it is possible to is effectively reduced and needs image data amount to be processed, but the super-pixel number of segmentation
Amount needs artificial appointment, when super-pixel quantity is arranged excessive time, meeting there will be when image is split
The phenomenon of less divided, when super-pixel quantity is arranged too small time, can increase need image data amount to be processed.
Therefore, a kind of new image splitting scheme of offer is needed badly, to automatically determine figure according to image concrete condition
As the super-pixel quantity required for segmentation, thus avoid the occurrence of while reducing image procossing complexity and owe to divide
The problem cut.
Summary of the invention
The purpose of the embodiment of the present invention is to provide a kind of image partition method and system, with concrete according to image
Situation automatically determines the super-pixel quantity of needs and splits image, is reducing the same of image procossing complexity
Time avoid the occurrence of less divided.Concrete technical scheme is as follows:
First aspect, embodiments provides a kind of image partition method, and described method may include that
Obtain pending target image；
Calculate the grey level histogram of described target image；
Based on described grey level histogram, calculate each pixel value that pixel in described target image contained and go out
Existing probability；
The probability that the number of the pixel comprised based on described target image and each pixel value described occur,
Utilize the complicated dynamic behaviour formula preset, calculate the complexity of described target image；
Complexity based on described target image, utilizes the super-pixel segmentation number computing formula preset, calculates super
Pixel segmentation number；
Split number based on described super-pixel, described target image is split.
Preferably, described default complicated dynamic behaviour formula is:
Wherein, described ε represents that complexity, described Η represent the picture that in described target image, pixel is contained
The number of element value, described piRepresent the probability that pixel value i occurs.
Preferably, described default super-pixel segmentation number computing formula is:
Wherein, described N represents that super-pixel splits number, and in target image described in described Η, pixel is contained
The number of pixel value, ε represents complexity, describedFor preset constant.
Preferably, described based on described super-pixel split number, described target image is split, including:
Split number based on described super-pixel, utilize super-pixel segmentation Turbopixels algorithm to described target image
Split.
Preferably, split number described based on described super-pixel, after described target image is split, also
Including:
Obtain the target super-pixel region of the target image after segmentation, and calculate the picture in this target super-pixel region
Element variance；
Based on described variance and variance correction formula, it is judged that described target super-pixel region is examined the need of edge
Survey；
If judging, described target super-pixel region needs rim detection, then utilize Sobel operator to surpass described target
Pixel region carries out rim detection.
Preferably, variance correction formula is:
Wherein, described σiRepresenting the pixel variance of target super-pixel region i, described Thr represents default and judges threshold
Value.
Second aspect, embodiments provides a kind of image segmentation system, and described system may include that
Acquisition module, the first computing module, the second computing module, the 3rd computing module, the 4th computing module and figure
As segmentation module；
Described acquisition module, for obtaining pending target image；
Described first computing module, for calculating the grey level histogram of described target image；
Described second computing module, for based on described grey level histogram, calculates pixel in described target image
The probability that each pixel value that point is contained occurs；
Described 3rd computing module, for the number of pixel comprised based on described target image and described
The probability that each pixel value occurs, utilizes the complicated dynamic behaviour formula preset, calculates answering of described target image
Miscellaneous degree；
Described 4th computing module, for complexity based on described target image, utilizes the super-pixel preset
Segmentation number computing formula, calculates super-pixel segmentation number；
Described image segmentation module, for splitting number based on described super-pixel, is carried out described target image point
Cut, with the target image after being split.
Preferably, the described default complicated dynamic behaviour formula that described 3rd computing module is utilized is:
Wherein, described ε represents that complexity, described Η represent the picture that in described target image, pixel is contained
The number of element value, described piRepresent the probability that pixel value i occurs.
Preferably, the super-pixel the preset segmentation number computing formula that described 4th computing module is utilized is:
Wherein, described N represents that super-pixel splits number, and in target image described in described Η, pixel is contained
The number of pixel value, ε represents complexity, describedFor preset constant.
Preferably, described image segmentation module includes: image segmentation submodule；
Described image segmentation submodule, for splitting number based on described super-pixel, utilizes super-pixel to split
Target image described in Turbopixels algorithm is split.
Preferably, described image segmentation system, also include:
4th computing module, the target super-pixel region of the target image after obtaining segmentation, and calculating should
The pixel variance in target super-pixel region；
Judge module, for based on described variance and variance correction formula, it is judged that described target super-pixel region
The need of rim detection；
Detection module, for when judging that described target super-pixel region needs rim detection, utilizes Sobel
Operator carries out rim detection to described target super-pixel region；
Wherein, the variance correction formula that described judge module is utilized is:
Wherein, described σiRepresenting the pixel variance of target super-pixel region i, described Thr represents default and judges threshold
Value.
The image partition method of embodiment of the present invention offer and system, first obtain target image to be split, and
Obtain the grey level histogram of this target image；According to the grey level histogram acquired, calculate in target image
The probability that each pixel value that pixel is contained occurs；Number based on the pixel that target image is comprised
The probability occurred with each pixel value, utilizes the complexity formula preset, and calculates the complexity of target image；
According to this complexity, utilizing the super-pixel segmentation number computing formula preset, the super-pixel calculating target image is divided
Cut number so that super-pixel segmentation number can be calculated according to the image complexity of target image automatically, it is to avoid
Artificially determine less divided that super-pixel segmentation number causes or increase the problem needing image data amount to be processed.
Certainly, arbitrary product or the method for implementing the present invention it is not absolutely required to reach all the above excellent simultaneously
Point.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to enforcement
In example or description of the prior art, the required accompanying drawing used is briefly described, it should be apparent that, describe below
In accompanying drawing be only some embodiments of the present invention, for those of ordinary skill in the art, do not paying
On the premise of going out creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.
A kind of image partition method that Fig. 1 provides for the embodiment of the present invention；
A kind of image segmentation system that Fig. 2 provides for the embodiment of the present invention.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clearly
Chu, be fully described by, it is clear that described embodiment be only a part of embodiment of the present invention rather than
Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not making creation
The every other embodiment obtained under property work premise, broadly falls into the scope of protection of the invention.
In order to solve prior art problem, embodiments provide a kind of image partition method and system.
First, a kind of image partition method is embodiments provided, as it is shown in figure 1, the method is permissible
Including:
S101: obtain pending target image；
Wherein, target image to be split is first obtained, it is to be understood that this target image can be from network
In directly obtain, it is also possible to be to image capture device send image capture instruction, image acquisition device perform
The image that this image capture instruction collects.
It is understood that the image partition method that the embodiment of the present invention is provided can apply to framing
In, such as: in application based on framing, need to first pass through the camera collection entrained by terminal and work as
The image of front user present position or the image of user current location can be embodied, and the figure that this is collected
As carrying out gray proces, obtain the gray-scale map of correspondence, and using this gray-scale map as first object image, wherein,
This first object image can be one or more；
Furthermore it is possible to by TC-OFDM (Time&Code Division-Orthogonal Frequency
Division Multiplexing, time-division code division orthogonal frequency division multiplexi) technology determines current user to be positioned
Be likely to be at region, and from pre-set image storehouse, the figure meeting this area condition is filtered out based on this region
Picture, and using the image that meets this area condition as the second target image, to reduce to enter with first object image
The amount of images of row coupling, and then reduce amount of calculation, wherein, this second target image can be one or more；
Wherein, the first image and the second image can be as pending target images；
And then first object image and the second target image are carried out image segmentation, and the image after segmentation is entered
Row feature extraction, and based on extracting feature, first object image and the second target image are carried out image
Join, and then obtain the address date corresponding to the second target image successful with first object images match, enter
And realize location.
Certainly, the image partition method that the embodiment of the present invention is provided can be not only applicable to framing
In, other any application scenarios that there is super-pixel segmentation demand all can utilize the embodiment of the present invention to be provided
Method.
S102: calculate the grey level histogram of this target image；
Wherein, after obtaining target image, calculate the grey level histogram of this target image, to obtain each picture
The number of the pixel corresponding to element value.If it is understood that by total for image pixel intensity (gray level
) do not regard a stochastic variable as, then its distribution situation just reflects the statistical property of image, and this can use
Probability density function (PDF) portrays and describes, and shows as grey level histogram.So-called
Grey level histogram be the function about grey level distribution, be to the statistics of grey level distribution in image.Specifically
, grey level histogram is by all pixels in digital picture, according to the size of gray value, adds up it and occurs
Frequency.
In the present embodiment, the specific implementation calculating grey level histogram can use of the prior art arbitrary
Plant implementation, do not limit at this.
S103: based on this grey level histogram, calculate each pixel value that pixel in this target image is contained
The probability occurred；
Wherein, after obtaining the grey level histogram of target image, the pixel value contained due to target image is deposited
At at least one, therefore, in order to calculated for subsequent image splits the complexity being based on, can be based on this gray scale
Rectangular histogram, calculates the probability of each pixel value appearance that pixel in this target image is contained.
It is emphasized that to ensure higher precision, all pixels in this target image can be calculated
The probability that each pixel value that point is contained occurs；And in order to ensure higher processing speed, can be calculated this
The probability that each pixel value that in target image, part pixel is contained occurs, this partial pixel point can root
Predefine according to practical situation, such as, the pixel of central area, the pixel etc. of foreground area, this
It is all rational.In actual applications, target figure can be calculated according to the demand to precision or speed
The probability that the pixel value that in Xiang, part or all of pixel is contained occurs, this is not limited by the embodiment of the present invention
Fixed.
It is general that S104: the number of the pixel comprised based on this target image and this each pixel value occur
Rate, utilizes the complicated dynamic behaviour formula preset, calculates the complexity of this target image；
Wherein, may determine that while target image and obtain pixel that this target image comprised obtaining
Number；After obtaining the probability that each pixel value occurs, in conjunction with the number of the pixel that this target image is comprised
Mesh, utilizes the complicated dynamic behaviour formula preset, is calculated the complexity of target image, in order to this target
Image performs next step operation.
Wherein, the complicated dynamic behaviour formula preset exists multiple, in order to layout is clear, follow-up carry out citing and is situated between
Continue.
S105: complexity based on this target image, utilizes the super-pixel segmentation number computing formula preset, meter
Calculate super-pixel segmentation number；
Wherein, the value corresponding to the complexity of target image is substituted into the super-pixel separation calculation formula preset,
Make can complexity based on target image, be calculated super-pixel most suitable for this target image segmentation
Number so that super-pixel segmentation number can be calculated according to the image complexity of target image automatically.
Wherein, in order to layout is clear, follow-up default super-pixel segmentation number computing formula is introduced.
S106: split number based on this super-pixel, this target image is split, with the mesh after being split
Logo image.
Wherein, split number according to calculated super-pixel, target image is split, after being split
Target image, it is to avoid artificially determine that super-pixel segmentation number causes problem or the increase of target image less divided
The problem needing image data amount to be processed.
Wherein, after determining this super-pixel segmentation number, prior art can be used to realize based on this super picture
Element segmentation number, splits this target image, with the target image after being split.
It is emphasized that above-mentioned image partition method can be used for what any one needed to split image
In device.
In embodiments of the present invention, first obtain target image to be split, and obtain the gray scale of this target image
Rectangular histogram；According to the grey level histogram acquired, calculate each picture that in target image, pixel is contained
The probability that element value occurs；It is general that the number of the pixel comprised based on target image and each pixel value occur
Rate, utilizes the complexity formula preset, and calculates the complexity of target image；According to this complexity, utilize pre-
If super-pixel segmentation number computing formula, calculate target image super-pixel segmentation number so that can be automatic
Image complexity according to target image calculates super-pixel segmentation number, it is to avoid artificially determine that super-pixel is split
Count the less divided caused or increase the problem needing image data amount to be processed.
Preferably, this complicated dynamic behaviour formula preset is:
Wherein, this ε represents complexity, and this Η represents the pixel value that in this target image, pixel is contained
Number, this piRepresent the probability that pixel value i occurs.
It is understood that in this kind of implementation, utilize this complicated dynamic behaviour formula preset to calculate mesh
The complexity of logo image, this kind of complicated dynamic behaviour mode calculates simplicity, greatly reduces amount of calculation so that this
The image partition method energy wider application that embodiment is provided is in the limited terminal of computing capability, wherein,
This terminal includes but are not limited to: mobile phone, panel computer and intelligent watch.
Preferably, the super-pixel segmentation number computing formula that this is preset is:
Wherein, this N represents that super-pixel splits number, the pixel value that in this target image of this Η, pixel is contained
Number, ε represents complexity, shouldFor preset constant.
It is understood that in this kind of implementation, this super-pixel segmentation number computing formula is utilized to calculate the
The segmentation number of two images so that segmentation number can be determined accurately by more simple and fast so that image segmentation speed is more
Hurry up, and the situation of less divided is sent out to avoid super-pixel edge to occur while decreasing the treating capacity of data
Raw.Wherein, less divided represents that image splits halfway situation.
Preferably, number should be split based on this super-pixel, this target image was split, including:
Split number based on this super-pixel, utilize super-pixel segmentation Turbopixels algorithm that this target image is carried out
Segmentation.
It is understood that in this kind of implementation, utilize calculated super-pixel segmentation number to determine super
The super-pixel segmentation number manually determined needed in pixel segmentation Turbopixels algorithm, and split by super-pixel
Turbopixels algorithmic preliminaries builds super-pixel edge and splits target image.
Preferably, split number at this based on this super-pixel, after this target image is split, also include:
Obtain the target super-pixel region of the target image after segmentation, and calculate the picture in this target super-pixel region
Element variance；
Based on this variance and variance correction formula, it is judged that this target super-pixel region is the need of rim detection；
If judging, this target super-pixel region needs rim detection, then utilize Sobel operator to this target super-pixel
Region carries out rim detection.
It is understood that in this kind of implementation, obtain the target super-pixel of the target image after segmentation
Region, and calculate the pixel variance in this target super-pixel region, according to this variance and variance correction formula,
Judge whether to need this target super-pixel region is carried out rim detection.Make according to calculated super picture
Element segmentation several target image is split after, can further determine whether need to segmentation after target image
Detect, i.e. when judging to need to carry out rim detection, utilize Sobel (Sobeloperator, Sobel
Operator) operator according to pixel up and down and the weighted difference of left and right adjoint point gray value (pixel value), in super-pixel
Edge reaches this feature of extreme value and the target image after segmentation is carried out rim detection, further avoid super
Pixel edge occurs that the situation of less divided occurs, thus improves the precision at super-pixel edge.
Wherein, the algorithm corresponding to Sobel operator calculates simple, calculates speed fast, it is possible to quickly complete super
Pixel is split；Wherein, super-pixel edge refers at the sudden change of the information such as gray scale or structure, and this spy available
Levy and image is split.
Preferably, variance correction formula is:
Wherein, this σiRepresenting the pixel variance of target super-pixel region i, this Thr represents default judgment threshold.
It is understood that in this kind of implementation, when the pixel variance at target super-pixel region i is big
When 1.2 times of default judgment thresholds, then needing to detect super-pixel edge, otherwise, then it is right to need not
Super-pixel edge detects.Wherein, Thr value can be carried out according to real needs by those skilled in the art
Determine, be not detailed at this.
Second aspect, embodiments provides a kind of image segmentation system, and this image segmentation system is permissible
Including: acquisition module the 201, first computing module the 202, second computing module the 203, the 3rd computing module
204, the 4th computing module 205 and image segmentation module 206；
This acquisition module 201, for obtaining pending target image；
This first computing module 202, for calculating the grey level histogram of this target image；
This second computing module 203, for based on this grey level histogram, calculates pixel in this target image
The probability that each pixel value contained occurs；
3rd computing module 204, the number of the pixel for being comprised based on this target image is each with this
The probability that individual pixel value occurs, utilizes the complicated dynamic behaviour formula preset, calculates the complexity of this target image；
4th computing module 205, for complexity based on this target image, utilizes the super-pixel preset
Segmentation number computing formula, calculates super-pixel segmentation number；
This image segmentation module 206, for splitting number based on this super-pixel, splits this target image,
With the target image after being split.
In embodiments of the present invention, first obtain target image to be split, and obtain the gray scale of this target image
Rectangular histogram；According to the grey level histogram acquired, calculate each picture that in target image, pixel is contained
The probability that element value occurs；It is general that the number of the pixel comprised based on target image and each pixel value occur
Rate, utilizes the complexity formula preset, and calculates the complexity of target image；According to this complexity, utilize pre-
If super-pixel segmentation number computing formula, calculate target image super-pixel segmentation number so that can be automatic
Image complexity according to target image calculates super-pixel segmentation number, it is to avoid artificially determine that super-pixel is split
Count the less divided caused or increase the problem needing image data amount to be processed.
Preferably, this complicated dynamic behaviour formula preset that the 3rd computing module 204 is utilized is:
Wherein, this ε represents complexity, and this Η represents the pixel value that in this target image, pixel is contained
Number, this piRepresent the probability that pixel value i occurs.
Preferably, the super-pixel the preset segmentation number computing formula that the 4th computing module 205 is utilized is:
Wherein, this N represents that super-pixel splits number, the pixel value that in this target image of this Η, pixel is contained
Number, ε represents complexity, shouldFor preset constant.
Preferably, this image segmentation module 206 includes: image segmentation submodule；
This image segmentation submodule, for splitting number based on this super-pixel, utilizes super-pixel to split Turbopixels
This target image is split by algorithm.
Preferably, image segmentation system, also include:
4th computing module, the target super-pixel region of the target image after obtaining segmentation, and calculating should
The pixel variance in target super-pixel region；
Judge module, for based on this variance and variance correction formula, it is judged that whether this target super-pixel region
Need rim detection；
Detection module, for when judging that this target super-pixel region needs rim detection, utilizes Sobel operator
This target super-pixel region is carried out rim detection；
Wherein, the variance correction formula that this judge module is utilized is:
Wherein, this σiRepresenting the pixel variance of target super-pixel region i, this Thr represents default judgment threshold.
For device embodiment, owing to it is substantially similar to embodiment of the method, so describing the simplest
Single, relevant part sees the part of embodiment of the method and illustrates.
It should be noted that in this article, the relational terms of such as first and second or the like be used merely to by
One entity or operation separate with another entity or operating space, and not necessarily require or imply these
Relation or the order of any this reality is there is between entity or operation.And, term " includes ", " bag
Contain " or its any other variant be intended to comprising of nonexcludability, so that include a series of key element
Process, method, article or equipment not only include those key elements, but also include being not expressly set out
Other key elements, or also include the key element intrinsic for this process, method, article or equipment.?
In the case of there is no more restriction, statement " including ... " key element limited, it is not excluded that at bag
Include and the process of described key element, method, article or equipment there is also other identical element.
Each embodiment in this specification all uses relevant mode to describe, phase homophase between each embodiment
As part see mutually, what each embodiment stressed is different from other embodiments it
Place.For system embodiment, owing to it is substantially similar to embodiment of the method, so describe
Fairly simple, relevant part sees the part of embodiment of the method and illustrates.
The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the protection model of the present invention
Enclose.All any modification, equivalent substitution and improvement etc. made within the spirit and principles in the present invention, all wrap
Containing within the scope of the present invention.
Claims (10)
1. an image partition method, it is characterised in that described method includes:
Obtain pending target image；
Calculate the grey level histogram of described target image；
Based on described grey level histogram, calculate each pixel value that pixel in described target image contained and go out
Existing probability；
The probability that the number of the pixel comprised based on described target image and each pixel value described occur,
Utilize the complicated dynamic behaviour formula preset, calculate the complexity of described target image；
Complexity based on described target image, utilizes the super-pixel segmentation number computing formula preset, calculates super
Pixel segmentation number；
Split number based on described super-pixel, described target image is split, with the target after being split
Image.
Method the most according to claim 1, it is characterised in that described default complicated dynamic behaviour formula
For:
Wherein, described ε represents that complexity, described Η represent the picture that in described target image, pixel is contained
The number of element value, described piRepresent the probability that pixel value i occurs.
Method the most according to claim 2, it is characterised in that described default super-pixel segmentation number meter
Calculation formula is:
Wherein, described N represents that super-pixel splits number, and described Η represents that in described target image, pixel is contained
The number of the pixel value of lid, ε represents complexity, describedFor preset constant.
The most according to the method in any one of claims 1 to 3, it is characterised in that described based on described
Super-pixel segmentation number, splits described target image, including:
Split number based on described super-pixel, utilize super-pixel segmentation Turbopixels algorithm to described target image
Split.
The most according to the method in any one of claims 1 to 3, it is characterised in that described based on institute
State super-pixel segmentation number, after described target image is split, also include:
Obtain the target super-pixel region of the target image after segmentation, and calculate the picture in this target super-pixel region
Element variance；
Based on described variance and variance correction formula, it is judged that described target super-pixel region is examined the need of edge
Survey；
When judging that described target super-pixel region needs rim detection, utilize Sobel Operator Sobel operator pair
Described target super-pixel region carries out rim detection.
Method the most according to claim 5, it is characterised in that described variance correction formula is:
Wherein, described σiRepresenting the pixel variance of target super-pixel region i, described Thr represents default and judges threshold
Value.
7. an image segmentation system, it is characterised in that including:
Acquisition module, for obtaining pending target image；
First computing module, for calculating the grey level histogram of described target image；
Second computing module, for based on described grey level histogram, calculates pixel institute in described target image
The probability that each pixel value contained occurs；
3rd computing module, for the number of pixel comprised based on described target image and described each
The probability that pixel value occurs, utilizes the complicated dynamic behaviour formula preset, calculates the complexity of described target image；
4th computing module, for complexity based on described target image, utilizes the super-pixel segmentation preset
Number computing formula, calculates super-pixel segmentation number；
Image segmentation module, for splitting number based on described super-pixel, splits described target image,
With the target image after being split.
System the most according to claim 7, it is characterised in that described 3rd computing module is utilized
Described default complicated dynamic behaviour formula is:
Wherein, described ε represents that complexity, described Η represent the picture that in described target image, pixel is contained
The number of element value, described piRepresent the probability that pixel value i occurs.
System the most according to claim 8, it is characterised in that described default super-pixel segmentation number meter
Calculation formula is:
Wherein, described N represents that super-pixel splits number, and in target image described in described Η, pixel is contained
The number of pixel value, ε represents complexity, describedFor preset constant.
10. according to the system according to any one of claim 7 to 9, it is characterised in that also include:
4th computing module, the target super-pixel region of the target image after obtaining segmentation, and calculating should
The pixel variance in target super-pixel region；
Judge module, for based on described variance and variance correction formula, it is judged that described target super-pixel region
The need of rim detection；
Detection module, for when judging that described target super-pixel region needs rim detection, utilizes Sobel
Operator carries out rim detection to described target super-pixel region；
Wherein, the variance correction formula that described judge module is utilized is:
Wherein, described σiRepresenting the pixel variance of target super-pixel region i, described Thr represents default and judges threshold
Value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610250328.8A CN105869175A (en) | 2016-04-21 | 2016-04-21 | Image segmentation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610250328.8A CN105869175A (en) | 2016-04-21 | 2016-04-21 | Image segmentation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105869175A true CN105869175A (en) | 2016-08-17 |
Family
ID=56633144
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610250328.8A Pending CN105869175A (en) | 2016-04-21 | 2016-04-21 | Image segmentation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105869175A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106296715A (en) * | 2016-08-23 | 2017-01-04 | 湖南文理学院 | Thresholding Method for Grey Image Segmentation based on symmetrical Beta divergence |
| CN106340023A (en) * | 2016-08-22 | 2017-01-18 | 腾讯科技（深圳）有限公司 | Image segmentation method and image segmentation device |
| CN106651841A (en) * | 2016-12-02 | 2017-05-10 | 北京航星机器制造有限公司 | Analysis method for security inspection of image complexity |
| CN106919401A (en) * | 2017-03-10 | 2017-07-04 | 广东欧珀移动通信有限公司 | A kind of method of dynamic regulation resolution ratio, device and mobile terminal |
| CN107767383A (en) * | 2017-11-01 | 2018-03-06 | 太原理工大学 | A kind of Road image segmentation method based on super-pixel |
| CN109035220A (en) * | 2018-07-10 | 2018-12-18 | 珠海格力智能装备有限公司 | The image detecting method and device of target object |
| CN109753955A (en) * | 2018-12-31 | 2019-05-14 | 中国北方车辆研究所 | A kind of multidirectional imaging system |
| CN110827309A (en) * | 2019-11-12 | 2020-02-21 | 太原理工大学 | Polaroid appearance defect segmentation method based on super-pixels |
| CN114693816A (en) * | 2022-04-02 | 2022-07-01 | 河南睿焓电子科技有限公司 | Intelligent image big data storage method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504662A (en) * | 2014-12-26 | 2015-04-08 | 北京慧眼智行科技有限公司 | Homomorphic filtering based image processing method and system |
| CN104966054A (en) * | 2015-06-11 | 2015-10-07 | 西安电子科技大学 | Weak and small object detection method in visible image of unmanned plane |
-
2016
- 2016-04-21 CN CN201610250328.8A patent/CN105869175A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504662A (en) * | 2014-12-26 | 2015-04-08 | 北京慧眼智行科技有限公司 | Homomorphic filtering based image processing method and system |
| CN104966054A (en) * | 2015-06-11 | 2015-10-07 | 西安电子科技大学 | Weak and small object detection method in visible image of unmanned plane |
Non-Patent Citations (4)
| Title |
|---|
| DAVID ENGEL 等: "Medial Features for Superpixel Segmentation", 《ELEVENTH IAPR CONFERENCE ON MACHINE VISION APPLICATIONS》 * |
| JICHAO JIAO 等: "Individual Building Rooftop and Tree Crown Segmentation from High-Resolution Urban Aerial Optical Images", 《JOURNAL OF SENSORS》 * |
| 宋熙煜 等: "图像分割中的超像素方法研究概述", 《中国图象图形学报》 * |
| 王玉顺: "《MATLAB实践教程》", 31 August 2012, 西安电子科技大学出版社 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106340023B (en) * | 2016-08-22 | 2019-03-05 | 腾讯科技（深圳）有限公司 | The method and apparatus of image segmentation |
| CN106340023A (en) * | 2016-08-22 | 2017-01-18 | 腾讯科技（深圳）有限公司 | Image segmentation method and image segmentation device |
| CN106296715A (en) * | 2016-08-23 | 2017-01-04 | 湖南文理学院 | Thresholding Method for Grey Image Segmentation based on symmetrical Beta divergence |
| CN106651841B (en) * | 2016-12-02 | 2020-10-16 | 北京航星机器制造有限公司 | Analysis method for security inspection image complexity |
| CN106651841A (en) * | 2016-12-02 | 2017-05-10 | 北京航星机器制造有限公司 | Analysis method for security inspection of image complexity |
| CN106919401A (en) * | 2017-03-10 | 2017-07-04 | 广东欧珀移动通信有限公司 | A kind of method of dynamic regulation resolution ratio, device and mobile terminal |
| CN106919401B (en) * | 2017-03-10 | 2021-04-09 | Oppo广东移动通信有限公司 | Method and device for dynamically adjusting resolution and mobile terminal |
| CN107767383A (en) * | 2017-11-01 | 2018-03-06 | 太原理工大学 | A kind of Road image segmentation method based on super-pixel |
| CN107767383B (en) * | 2017-11-01 | 2021-05-11 | 太原理工大学 | Road image segmentation method based on superpixels |
| CN109035220A (en) * | 2018-07-10 | 2018-12-18 | 珠海格力智能装备有限公司 | The image detecting method and device of target object |
| CN109753955A (en) * | 2018-12-31 | 2019-05-14 | 中国北方车辆研究所 | A kind of multidirectional imaging system |
| CN110827309A (en) * | 2019-11-12 | 2020-02-21 | 太原理工大学 | Polaroid appearance defect segmentation method based on super-pixels |
| CN114693816A (en) * | 2022-04-02 | 2022-07-01 | 河南睿焓电子科技有限公司 | Intelligent image big data storage method |
| CN114693816B (en) * | 2022-04-02 | 2023-04-28 | 陕西合友网络科技有限公司 | Intelligent image big data storage method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105869175A (en) | Image segmentation method and system | |
| CN102129693B (en) | Image vision significance calculation method based on color histogram and global contrast | |
| CN110781756A (en) | Urban road extraction method and device based on remote sensing image | |
| CN104408429A (en) | Method and device for extracting representative frame of video | |
| CN110909640A (en) | Method and device for determining water level line, storage medium and electronic device | |
| CA2326631A1 (en) | Representative color designating method using reliability | |
| CN109858441A (en) | A kind of monitoring abnormal state method and apparatus for construction site | |
| CN110866926B (en) | Infrared remote sensing image rapid and fine sea-land segmentation method | |
| CN107240092A (en) | A kind of image blur detection method and device | |
| CN105792131A (en) | Positioning method and system | |
| CN106228186A (en) | Classification hyperspectral imagery apparatus and method | |
| CN104243815A (en) | Focusing method and electronic equipment | |
| CN111310727A (en) | Object detection method and device, storage medium and electronic device | |
| CN111428653B (en) | Pedestrian congestion state judging method, device, server and storage medium | |
| CN110310301B (en) | Method and device for detecting target object | |
| CN104361357A (en) | Photo set classification system and method based on picture content analysis | |
| CN111783732A (en) | Group mist identification method and device, electronic equipment and storage medium | |
| CN110210774A (en) | Landslide risk evaluation method and system | |
| CN114155241A (en) | Foreign matter detection method and device and electronic equipment | |
| CN112597995B (en) | License plate detection model training method, device, equipment and medium | |
| CN105956595A (en) | Image feature extraction method and system | |
| CN117037127A (en) | Pallet distribution method based on luggage type | |
| CN112017163A (en) | Image blur degree detection method and device, electronic equipment and storage medium | |
| CN112969027B (en) | Focusing method and device of electric lens, storage medium and electronic equipment | |
| CN108769521A (en) | A kind of photographic method, mobile terminal and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160817 |