CN108537812A - In conjunction with minimum spanning tree dividing method, system and the device of Ostu threshold methods - Google Patents
In conjunction with minimum spanning tree dividing method, system and the device of Ostu threshold methods Download PDFInfo
- Publication number
- CN108537812A CN108537812A CN201810304954.XA CN201810304954A CN108537812A CN 108537812 A CN108537812 A CN 108537812A CN 201810304954 A CN201810304954 A CN 201810304954A CN 108537812 A CN108537812 A CN 108537812A
- Authority
- CN
- China
- Prior art keywords
- minimum spanning
- spanning tree
- image
- ostu threshold
- ostu
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of minimum spanning tree dividing method, system and the devices of combination Ostu threshold methods, wherein in conjunction with the minimum spanning tree dividing method of Ostu threshold methods, including:Convert the RGB color of image to hsv color space;H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain final Ostu threshold values;Described image is mapped as the figure in graph theory, constructs minimum spanning tree;The minimum spanning tree is merged, to obtain the image of segmentation.Realize the advantage for reducing that accidentally segmentation rate is high.
Description
Technical field
The present invention relates to image processing fields, and in particular, to a kind of minimum spanning tree segmentation of combination Ostu threshold methods
Method, system and device.
Background technology
Image segmentation refers to the process of dividing an image into different zones according to certain similarity criterion, is that computer regards
One of Basic Problems in fields such as feel, image procossing are that image classification, scene parsing, object detection, image 3D reconstruct etc. are appointed
The pretreatment of business.Traditional image partition method includes mainly threshold method, Edge Detection Using, field method etc., the reality of these methods
Existing principle is different, but is substantially and utilizes the letters such as the rudimentary semanteme, including the color of image pixel, texture and shape of image
Breath, it is ideal not to the utmost to encounter practical segmentation effect when complex scene.The theory and method of graph theory introduce image segmentation problem.Its principle
It is that image to be split is mapped as weighted-graph, according to the information structuring cost function on the vertex of figure and side and is subject to excellent
Change, the problem of dividing the image into is converted to the vertex mark problem of figure, and the pixel corresponding to the identical vertex of label belongs to same
Image block.
Minimum spanning tree is a concept in graph theory, refers to the spanning tree of the weights sum minimum on side.Minimum is generated
Tree algorithm is applied to image segmentation problem, can obtain the global characteristics of image, and segmentation effect is good, and algorithm structure is simple, meter
It is fast to calculate speed.
There is the defect that accidentally segmentation rate is high to the separation of image in the prior art.
Invention content
It is an object of the present invention in view of the above-mentioned problems, propose a kind of minimum spanning tree segmentation of combination Ostu threshold methods
Method, system and device reduce the advantage that accidentally segmentation rate is high to realize.
To achieve the above object, on the one hand technical solution of the present invention, provides a kind of most your pupil of combination Ostu threshold methods
At tree dividing method, including:
Convert the RGB color of image to hsv color space;
H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain most
Whole Ostu threshold values;
Described image is mapped as the figure in graph theory, constructs minimum spanning tree;
The minimum spanning tree is merged, to obtain the image of segmentation.
Preferably, described that described image is mapped as the figure in graph theory, it constructs in minimum spanning tree:
Use Kruskal algorithm construction minimum spanning trees.
Preferably, described that described image is mapped as the figure in graph theory, it constructs in minimum spanning tree:
Using the Euclidean distance of weighting as the weight of Kruskal algorithms, to construct minimum spanning tree.
Preferably, described that the minimum spanning tree is merged, to show that the merging condition of the image of segmentation is:
It is unsatisfactory for the final Ostu threshold values.
A kind of minimum spanning tree segmenting system of combination Ostu threshold methods is also disclosed in technical solution of the present invention, including,
Conversion module:Convert the RGB color of image to hsv color space;
Computing module:H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are added
Power, obtains final Ostu threshold values;
Constructing module:Described image is mapped as the figure in graph theory, constructs minimum spanning tree;
Merging module:The minimum spanning tree is merged, to obtain the image of segmentation.
Preferably, in the constructing module:
Use Kruskal algorithm construction minimum spanning trees.
Preferably, in the constructing module:
Using the Euclidean distance of weighting as the weight of Kruskal algorithms, to construct minimum spanning tree.
Preferably, the minimum spanning tree is merged in the merging module, to obtain segmentation image conjunction
And condition is:
It is unsatisfactory for the final Ostu threshold values.
A kind of minimum spanning tree segmenting device of combination Ostu threshold methods is also disclosed in technical solution of the present invention, and feature exists
In, including memory:Storage executes program code;And
Processor:It is configured as running the execution program code, so as to:
Convert the RGB color of image to hsv color space;
H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain most
Whole Ostu threshold values;
Described image is mapped as the figure in graph theory, constructs minimum spanning tree;
The minimum spanning tree is merged, to obtain the image of segmentation.
Technical scheme of the present invention has the advantages that:
Minimal spanning tree algorithm is applied to image segmentation problem, can obtain the complete of image by technical solution of the present invention
Office's feature, segmentation effect is good, and algorithm structure is simple, and calculating speed is fast.In conjunction with the segmentation side of the minimum spanning tree of Ostu threshold methods
Method can preferably utilize minimum spanning tree and cluster the inner link between structure, effectively be partitioned into target and background, weight
The target and background for using Ostu threshold values further clearly to divide again, improve the accuracy rate of segmentation.To reduce accidentally segmentation rate.
And obtain the visual effect of better image segmentation.
The image segmentation algorithm of the technical program is a kind of method of the optimization of minimum spanning tree, and region merging technique is adjudicated item
Part depends on the Ostu threshold values in two regions mutually merged, it is also contemplated that smaller region is included in larger target area
Or in background area, reuses Ostu threshold values and carry out region merging technique.The over-segmentation that image can effectively be weakened divides with deficient
Cut ratio.
Hsv color space itself can be very good, by different color segmentations, to be weighted them, can preferably adjust
The weights between different colours are saved, to carry out careful segmentation.
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
Description of the drawings
Fig. 1 is the flow of the minimum spanning tree dividing method of the combination Ostu threshold methods described in the embodiment of the present invention;
Fig. 2 a to Fig. 2 c are to be divided using the minimum spanning tree dividing method of combination Ostu threshold methods of the present invention
Every lab diagram;
Fig. 3 a to Fig. 3 c are to be divided using the minimum spanning tree dividing method of combination Ostu threshold methods of the present invention
Every another lab diagram.
Specific implementation mode
It will be appreciated that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.
As shown in Figure 1, a kind of minimum spanning tree dividing method of combination Ostu threshold methods, including:
Step S1:Input picture converts the RGB color of image to hsv color space;
Step S2:H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted,
Obtain final Ostu threshold values;
Step S3:Image is mapped as the figure in graph theory, constructs minimum spanning tree;
Step S4:Minimum spanning tree is merged, to obtain the image of segmentation.
Preferably, image is mapped as the figure in graph theory, constructed in minimum spanning tree:
Use Kruskal algorithm construction minimum spanning trees.
Preferably, image is mapped as the figure in graph theory, constructed in minimum spanning tree:
Using the Euclidean distance of weighting as the weight of Kruskal algorithms, to construct minimum spanning tree.
Preferably, minimum spanning tree is merged, to show that the merging condition of the image of segmentation is:
It is unsatisfactory for final Ostu threshold values.
A kind of minimum spanning tree segmenting system of combination Ostu threshold methods is also disclosed in technical solution of the present invention, including,
Conversion module:Convert the RGB color of image to hsv color space;
Computing module:H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are added
Power, obtains final Ostu threshold values;
Constructing module:Image is mapped as the figure in graph theory, constructs minimum spanning tree;
Merging module:Minimum spanning tree is merged, to obtain the image of segmentation.
Preferably, in constructing module:
Use Kruskal algorithm construction minimum spanning trees.
Preferably, in constructing module:
Using the Euclidean distance of weighting as the weight of Kruskal algorithms, to construct minimum spanning tree.
Preferably, minimum spanning tree is merged in merging module, to show that the merging condition of the image of segmentation is:
It is unsatisfactory for final Ostu threshold values.
A kind of minimum spanning tree segmenting device of combination Ostu threshold methods is also disclosed in technical solution of the present invention, and feature exists
In, including memory:Storage executes program code;And
Processor:It is configured as operation and executes program code, so as to:
Convert the RGB color of image to hsv color space;
H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain most
Whole Ostu threshold values;
Image is mapped as the figure in graph theory, constructs minimum spanning tree;
Minimum spanning tree is merged, to obtain the image of segmentation.
Wherein, Fig. 2 a and Fig. 3 a are the image of input, and Fig. 2 b and Fig. 3 b are transformed hsv color spatial image, Fig. 2 c
It is the image after automatic Segmentation using the present invention with Fig. 3 c.
In conclusion of the invention
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, all answer by the change or replacement that can be readily occurred in
It is included within the scope of the present invention.Therefore, protection scope of the present invention should be subject to the protection scope in claims.
Claims (9)
1. a kind of minimum spanning tree dividing method of combination Ostu threshold methods, which is characterized in that including:
Convert the RGB color of image to hsv color space;
H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain final
Ostu threshold values;
Described image is mapped as the figure in graph theory, constructs minimum spanning tree;
The minimum spanning tree is merged, to obtain the image of segmentation.
2. the minimum spanning tree dividing method of combination Ostu threshold methods according to claim 1, which is characterized in that described to incite somebody to action
Described image is mapped as the figure in graph theory, constructs in minimum spanning tree:
Use Kruskal algorithm construction minimum spanning trees.
3. the minimum spanning tree dividing method of combination Ostu threshold methods according to claim 2, which is characterized in that described to incite somebody to action
Described image is mapped as the figure in graph theory, constructs in minimum spanning tree:
Using the Euclidean distance of weighting as the weight of Kruskal algorithms, to construct minimum spanning tree.
4. the minimum spanning tree dividing method of combination Ostu threshold methods according to claim 1, which is characterized in that described right
The minimum spanning tree merges, to show that the merging condition of the image of segmentation is:
It is unsatisfactory for the final Ostu threshold values.
5. a kind of minimum spanning tree segmenting system of combination Ostu threshold methods, which is characterized in that including,
Conversion module:Convert the RGB color of image to hsv color space;
Computing module:H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain
Go out final Ostu threshold values;
Constructing module:Described image is mapped as the figure in graph theory, constructs minimum spanning tree;
Merging module:The minimum spanning tree is merged, to obtain the image of segmentation.
6. the minimum spanning tree segmenting system of combination Ostu threshold methods according to claim 5, which is characterized in that the structure
In modeling block:
Use Kruskal algorithm construction minimum spanning trees.
7. the minimum spanning tree segmenting system of combination Ostu threshold methods according to claim 6, which is characterized in that the structure
In modeling block:
Using the Euclidean distance of weighting as the weight of Kruskal algorithms, to construct minimum spanning tree.
8. the minimum spanning tree segmenting system of combination Ostu threshold methods according to claim 5, which is characterized in that the conjunction
And the minimum spanning tree is merged in module, to show that the merging condition of the image of segmentation is:
It is unsatisfactory for the final Ostu threshold values.
9. a kind of minimum spanning tree segmenting device of combination Ostu threshold methods, which is characterized in that including memory:Storage executes journey
Sequence code;And
Processor:It is configured as running the execution program code, so as to:
Convert the RGB color of image to hsv color space;
H in hsv color space is calculated separately, then the Ostu threshold values of tri- Color Channels of S, V are weighted, obtain final
Ostu threshold values;
Described image is mapped as the figure in graph theory, constructs minimum spanning tree;
The minimum spanning tree is merged, to obtain the image of segmentation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810304954.XA CN108537812A (en) | 2018-04-08 | 2018-04-08 | In conjunction with minimum spanning tree dividing method, system and the device of Ostu threshold methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810304954.XA CN108537812A (en) | 2018-04-08 | 2018-04-08 | In conjunction with minimum spanning tree dividing method, system and the device of Ostu threshold methods |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108537812A true CN108537812A (en) | 2018-09-14 |
Family
ID=63482861
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810304954.XA Pending CN108537812A (en) | 2018-04-08 | 2018-04-08 | In conjunction with minimum spanning tree dividing method, system and the device of Ostu threshold methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108537812A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102136145A (en) * | 2011-04-27 | 2011-07-27 | 中国科学院遥感应用研究所 | Region merging method for threshold-restrained minimum spanning tree algorithm |
CN102147921A (en) * | 2011-04-08 | 2011-08-10 | 浙江理工大学 | Graph theory-based Chinese medicinal tongue nature and tongue coat separation algorithm |
CN106097313A (en) * | 2016-06-02 | 2016-11-09 | 甘肃读者动漫科技有限公司 | Image partition method and device |
-
2018
- 2018-04-08 CN CN201810304954.XA patent/CN108537812A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102147921A (en) * | 2011-04-08 | 2011-08-10 | 浙江理工大学 | Graph theory-based Chinese medicinal tongue nature and tongue coat separation algorithm |
CN102136145A (en) * | 2011-04-27 | 2011-07-27 | 中国科学院遥感应用研究所 | Region merging method for threshold-restrained minimum spanning tree algorithm |
CN106097313A (en) * | 2016-06-02 | 2016-11-09 | 甘肃读者动漫科技有限公司 | Image partition method and device |
Non-Patent Citations (2)
Title |
---|
冯芳 等: "复杂背景下的***细胞图像分割方法", 《武汉大学学报(理学版)》 * |
王海罗 等: "基于超像素融合算法的显著区域检测", 《北京理工大学学报》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110543858A (en) | Multi-mode self-adaptive fusion three-dimensional target detection method | |
Hurtado et al. | Semantic scene segmentation for robotics | |
CN107506792B (en) | Semi-supervised salient object detection method | |
CN111241963B (en) | First person view video interactive behavior identification method based on interactive modeling | |
Xu et al. | Weakly supervised deep semantic segmentation using CNN and ELM with semantic candidate regions | |
Anthwal et al. | An overview of optical flow-based approaches for motion segmentation | |
CN111462149A (en) | Example human body analysis method based on visual saliency | |
CN113762201A (en) | Mask detection method based on yolov4 | |
CN112651423A (en) | Intelligent vision system | |
CN111652181B (en) | Target tracking method and device and electronic equipment | |
Yuan et al. | A lightweight network for smoke semantic segmentation | |
Li et al. | Deep learning based monocular depth prediction: Datasets, methods and applications | |
Khan et al. | Lrdnet: lightweight lidar aided cascaded feature pools for free road space detection | |
Jiang et al. | Salient regions detection for indoor robots using RGB-D data | |
Tang et al. | SDRNet: An end-to-end shadow detection and removal network | |
Xiao et al. | 3D-VDNet: Exploiting the vertical distribution characteristics of point clouds for 3D object detection and augmentation | |
CN104504715A (en) | Image segmentation method based on local quaternion-moment characteristic | |
Sabater et al. | Event Transformer+. A multi-purpose solution for efficient event data processing | |
Cambuim et al. | An efficient static gesture recognizer embedded system based on ELM pattern recognition algorithm | |
Chen et al. | Real-time lane detection model based on non bottleneck skip residual connections and attention pyramids | |
CN110458203B (en) | Advertisement image material detection method | |
CN108537812A (en) | In conjunction with minimum spanning tree dividing method, system and the device of Ostu threshold methods | |
Garcia-Ugarriza et al. | Automatic color image segmentation by dynamic region growth and multimodal merging of color and texture information | |
CN106296740B (en) | A kind of target fine definition tracking based on low-rank sparse expression | |
CN102184524A (en) | Neighbourhood learning meme image segmentation method based on standard cut |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180914 |