CN104794495B - A kind of large format remote sensing image territorial classification method based on straight line statistical nature - Google Patents
A kind of large format remote sensing image territorial classification method based on straight line statistical nature Download PDFInfo
- Publication number
- CN104794495B CN104794495B CN201510221008.5A CN201510221008A CN104794495B CN 104794495 B CN104794495 B CN 104794495B CN 201510221008 A CN201510221008 A CN 201510221008A CN 104794495 B CN104794495 B CN 104794495B
- Authority
- CN
- China
- Prior art keywords
- mrow
- straight line
- mfrac
- remote sensing
- sensing image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of large format remote sensing image territorial classification method based on straight line statistical nature.Comprise the following steps：Step 1, uniform grid divides；Step 2, straight line is extracted；Step 3, straight line support region is extracted；Step 4, ellipse fitting；Step 5, straight line is redefined；Step 6, the length attribute of straight line is extracted；Step 7, the statistical nature of straight line is calculated；Step 8, classify.Make full use of different types of remote sensing image region that there is different straight line statistical natures, the types such as the wasteland included in large format remote sensing image, rural area, outskirts of a town and High-Density Urban Area can be distinguished exactly, solve the problems, such as blindly to search for building in remote sensing image, be advantageous to improve the accuracy and efficiency of building analyte detection, reach the effect of full automation.
Description
Technical field
The present invention relates to a kind of remote sensing image process field, specifically a kind of large format based on straight line statistical nature is distant
Feel imagery zone sorting technique.
Background technology
Traditional remote sensing image territorial classification method based on pixel is only located at " image procossing " stage in Image Engineering,
The requirement of modern Remote Sensing Technical development can not have been met.For high-resolution remote sensing image, single pixel can not typically reflect
Real geographic object, the topological relation between pixel are also very limited.Due to the processing method based on pixel have ignored texture,
The space characteristics such as context and shape, introduced wherein, only even if soft grader, sub-pixed mapping sorting technique and spectrum solution are mixed into technology
Also difficulty greatly improves classification results based on image spectrum information.When classifying to high-resolution remote sensing image, it is based on
The method of pixel can cause even more serious " spiced salt effect ", so as to influence the accuracy of classification results.Although tradition is visually sentenced
The remote sensing information extracting method precision of reading is higher, but needs to put into substantial amounts of manpower and time, and oneself can not meet that mass data is believed
The processing requirement of breath.Currently, the means of extracting geographic information are also relatively backward from high-resolution remote sensing image, and more prominent asks
Topic is still " data magnanimity, information deficiency, knowledge are hard to find ".In addition, the people being in information-intensive society are to the ageing of information
Increasingly pay attention to, this effectively promotes Remote Sensing Data Processing method and developed towards semi-automatic, automation direction.Therefore,
Efficiently, intelligentized remote sensing image territorial classification certainly will turn into the current or even study hotspot of remote sensing information process from now on.
The content of the invention
The invention provides a kind of large format remote sensing image territorial classification method based on straight line statistical nature, mesh can be overcome
The problem of building extraction accuracy rate is low in preceding remote sensing image, makes full use of different types of remote sensing image region to have difference
Straight line statistical nature, the wasteland included in large format remote sensing image, rural area, outskirts of a town and urban compact can be distinguished exactly
The types such as area are high without manual intervention, automaticity.
Technical scheme is used by realize the target of the present invention：Method comprises the following steps：
Step 1：Input remote sensing image image1 is divided by uniform grid, with the subgraph I after division1、I2、…、
InAs minimal processing unit, n is the sum of the subgraph after division；
Step 2：To subgraph I1、I2、…、InRespectively with Straight Line Extraction extraction straight line；
Step 3：Respectively to subgraph I1、I2、…、InThe middle straight line for meeting syntople merges into a straight line support region；
Step 4：Straight line support region is fitted to ellipse with Fourier descriptor；
Step 5：Oval major axis is extracted, and is newly defined as straight line；
Step 6：Extract the length attribute of straight line；
Step 7：The statistical nature of straight line is calculated, includes the average and entropy of straight length；
Step 8：The statistical nature of the straight line in step 7 is classified using grader.
Described Straight Line Extraction uses the method based on gradient, the gradient G in x and y directionsx(x, y) and Gy(x, y) point
Li Yong not following two formula calculating：
Gx(x, y)={ e (x) h (y) * f (x, y) }
Gy(x, y)={ e (y) h (x) * f (x, y) }
The gradient calculation formula of pixel (x, y) is：
Wherein, e (k) is boundary filter, and h (k) is mapped filter, and k value is x or y, and corresponding formula is as follows：
H (k)=e-αkcos(αβk+π/2)
Wherein α is scale parameter, and β is image resolution parameter.
The determination method of described syntople is：Choose a threshold value T (threshold value T is set as 5), if two straight lines it
Between beeline be less than T, then it is directly adjacent to judge this two, otherwise to be non-adjacent.
It is described straight line support region is fitted to ellipse with Fourier descriptor method be：Utilize Fourier expansion
And Euler's formula, multiple periodic function is changed to obtain following Fourier coefficient：
Wherein x (k) and y (k) represents the real and imaginary parts of the borderline point of straight line support region respectively.
By extracting three system numbers：(α-1,β-1),(α0,β0) and (α1,β1), three oval parameters can be obtained：
Center：cxy=(α0,β0)
Length：
Direction：
The computational methods of the entropy of the statistical nature cathetus length of described straight line are：Structure one is by the straight of 50 bin
Fang Tu, each bin width are 4 pixels, are calculated using below equation：
Described grader uses the gloomy window grader of Paar.
The beneficial effects of the invention are as follows：Solve the problems, such as blindly to search for building in remote sensing image, be advantageous to carry
The accuracy and efficiency of high constructure detection, reach the effect of full automation.
Brief description of the drawings
Fig. 1 is the overall process flow figure of the present invention.
Embodiment
The embodiment of the present invention is described in detail below in conjunction with the accompanying drawings.
In step 101, the pending remote sensing image of input is Quick bird high spatial resolution remote sense image image1,
Size is 8000 × 8000, and has carried out the pretreatment such as radiant correction and geometric correction.
In step 102, remote sensing image image1 is divided into by uniform grid to the subgraph of 400 400 × 400.
In step 103, the subgraph to each 400 × 400, gradient, x and y are calculated using the method based on gradient respectively
The gradient G in directionx(x, y) and Gy(x, y) is utilized respectively following two formula and calculated：
Gx(x, y)={ e (x) h (y) * f (x, y) }
Gy(x, y)={ e (y) h (x) * f (x, y) }
The gradient calculation formula of pixel (x, y) is：
Wherein, e (k) is boundary filter, and h (k) is mapped filter, and k value is x or y, and corresponding formula is as follows：
H (k)=e-αkcos(αβk+π/2)
Wherein α is scale parameter, and it is image resolution parameter to take 1.8, β, takes 70.
Finally, the pixel that gradient magnitude is less than 10 is excluded, for suppressing noise.
In step 104, the straight line extracted in step 103 for each 400 × 400 subgraph, selected threshold T=5,
Calculate the beeline L between two straight linesminIf LminLess than T, then it is directly adjacent to judge this two, otherwise to be non-
Adjacent, the straight line of all of its neighbor is merged into a straight line support region.
In step 105, the straight line support region that step 104 obtains is fitted to ellipse with Fourier descriptor：Using in Fu
Leaf series expansion and Euler's formula, multiple periodic function is changed to obtain following Fourier coefficient：
Wherein x (k) and y (k) represents the real and imaginary parts of the borderline point of straight line support region respectively.
By extracting three system numbers：(α-1,β-1),(α0,β0) and (α1,β1), three oval parameters can be obtained：
Center：cxy=(α0,β0)
Length：
Direction：
In step 106, the ellipse that step 105 is obtained is represented with its major axis, i.e., ellipse fitting is in line.
In step 107, the statistical nature of straight line is calculated, includes the average and entropy of straight length.
The calculation formula of the average of straight length is：
Wherein, N is the quantity of last fitting a straight line in subgraph, LiFor the length of i-th straight line
The computational methods of the entropy of straight length are：A histogram by 50 bin is built, each bin width is 4
Pixel, calculated using below equation：
In step 108, classified using the gloomy window grader of Paar.
In step 109, output category result：In remote sensing image image1 is inputted, classification results are carried out with pseudo-colours
Superposition, wherein, wasteland is represented with yellow, green represents rural area, and blueness represents outskirts of a town, and red represents High-Density Urban Area.
Claims (5)
- A kind of 1. large format remote sensing image territorial classification method based on straight line statistical nature, it is characterised in that including following step Suddenly：Step 1：Input remote sensing image image1 is divided by uniform grid, with the subgraph I after division1、I2、…、InMake For minimal processing unit, n is the sum of the subgraph after division；Step 2：To subgraph I1、I2、…、InRespectively with Straight Line Extraction extraction straight line；Step 3：Respectively to subgraph I1、I2、…、InThe middle straight line for meeting syntople merges into a straight line support region；Step 4：Straight line support region is fitted to ellipse with Fourier descriptor；Step 5：Oval major axis is extracted, and is newly defined as straight line；Step 6：Extract the length attribute of straight line；Step 7：The statistical nature of straight line is calculated, includes the average and entropy of straight length；Step 8：The statistical nature of the straight line in step 7 is classified using grader；Straight Line Extraction uses the method based on gradient, the gradient G in x and y directionsx(x, y) and Gy(x, y) is utilized respectively following Two formula calculate：Gx(x, y)={ e (x) h (y) * f (x, y) }Gy(x, y)={ e (y) h (x) * f (x, y) }The gradient calculation formula of pixel (x, y) is：<mrow> <mi>G</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>arctan</mi> <mfrac> <mrow> <msub> <mi>G</mi> <mi>x</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> </mrow> <mrow> <msub> <mi>G</mi> <mi>y</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> </mrow> </mfrac> </mrow>Wherein, e (k) is boundary filter, and h (k) is mapped filter, and k value is x or y, and corresponding formula is as follows：<mrow> <mi>e</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <msup> <mi>e</mi> <mrow> <mo>-</mo> <mi>&alpha;</mi> <mi>k</mi> </mrow> </msup> <mi>&alpha;</mi> </mfrac> <mo>&lsqb;</mo> <mo>-</mo> <mfrac> <mrow> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mi>&beta;</mi> <mi>k</mi> <mo>+</mo> <mi>&pi;</mi> <mo>/</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>-</mo> <mi>&beta;</mi> <mi>sin</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mi>&beta;</mi> <mi>k</mi> <mo>+</mo> <mi>&pi;</mi> <mo>/</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> <mrow> <mn>1</mn> <mo>+</mo> <msup> <mi>&beta;</mi> <mn>2</mn> </msup> </mrow> </mfrac> <mo>&rsqb;</mo> </mrow>H (k)=e-αkcos(αβk+π/2)Wherein α is scale parameter, and β is image resolution parameter.
- 2. a kind of large format remote sensing image territorial classification method based on straight line statistical nature according to claim 1, its The determination method for being characterised by syntople is：A threshold value T is chosen, threshold value T is set as 5, if between two straight lines most Short distance is less than T, then judges that this two straight lines are adjacent, otherwise to be non-adjacent.
- 3. a kind of large format remote sensing image territorial classification method based on straight line statistical nature according to claim 1, its It is characterised by that straight line support region is fitted into oval method with Fourier descriptor is：Utilize Fourier expansion and Euler Formula, multiple periodic function is changed to obtain following Fourier coefficient：<mrow> <msub> <mi>a</mi> <mi>n</mi> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <mi>T</mi> </mfrac> <mo>&lsqb;</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>0</mn> </mrow> <mrow> <mi>T</mi> <mo>-</mo> <mn>1</mn> </mrow> </munderover> <mi>x</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mn>2</mn> <mi>&pi;</mi> <mi>n</mi> <mi>k</mi> </mrow> <mi>T</mi> </mfrac> <mo>)</mo> </mrow> <mo>+</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>0</mn> </mrow> <mrow> <mi>T</mi> <mo>-</mo> <mn>1</mn> </mrow> </munderover> <mi>y</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mi>sin</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mn>2</mn> <mi>&pi;</mi> <mi>n</mi> <mi>k</mi> </mrow> <mi>T</mi> </mfrac> <mo>)</mo> </mrow> <mo>&rsqb;</mo> </mrow><mrow> <msub> <mi>&beta;</mi> <mi>n</mi> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <mi>T</mi> </mfrac> <mo>&lsqb;</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>0</mn> </mrow> <mrow> <mi>T</mi> <mo>-</mo> <mn>1</mn> </mrow> </munderover> <mi>y</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mn>2</mn> <mi>&pi;</mi> <mi>n</mi> <mi>k</mi> </mrow> <mi>T</mi> </mfrac> <mo>)</mo> </mrow> <mo>-</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>0</mn> </mrow> <mrow> <mi>T</mi> <mo>-</mo> <mn>1</mn> </mrow> </munderover> <mi>x</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mi>sin</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <mn>2</mn> <mi>&pi;</mi> <mi>n</mi> <mi>k</mi> </mrow> <mi>T</mi> </mfrac> <mo>)</mo> </mrow> <mo>&rsqb;</mo> </mrow>Wherein x (k) and y (k) represents the real and imaginary parts of the borderline point of straight line support region respectively；By extracting three system numbers：(α-1,β-1),(α0,β0) and (α1,β1), three oval parameters can be obtained：Center：cxy=(α0,β0)Length：Direction：
- 4. a kind of large format remote sensing image territorial classification method based on straight line statistical nature according to claim 1, its The computational methods for being characterised by the entropy of the statistical nature cathetus length of straight line are：A histogram by 50 bin is built, often Individual bin width is 4 pixels, is calculated using below equation：<mrow> <mi>E</mi> <mo>=</mo> <mo>-</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mn>50</mn> </munderover> <mi>h</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> <msub> <mi>log</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>h</mi> <mo>(</mo> <mi>i</mi> <mo>)</mo> <mo>)</mo> </mrow> <mo>.</mo> </mrow>
- 5. a kind of large format remote sensing image territorial classification method based on straight line statistical nature according to claim 1, its It is characterised by that grader uses the gloomy window grader of Paar.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510221008.5A CN104794495B (en) | 2015-05-04 | 2015-05-04 | A kind of large format remote sensing image territorial classification method based on straight line statistical nature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510221008.5A CN104794495B (en) | 2015-05-04 | 2015-05-04 | A kind of large format remote sensing image territorial classification method based on straight line statistical nature |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104794495A CN104794495A (en) | 2015-07-22 |
| CN104794495B true CN104794495B (en) | 2018-02-06 |
Family
ID=53559282
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510221008.5A Expired - Fee Related CN104794495B (en) | 2015-05-04 | 2015-05-04 | A kind of large format remote sensing image territorial classification method based on straight line statistical nature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104794495B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778774B (en) * | 2016-11-25 | 2020-04-03 | 福建师范大学 | High-resolution remote sensing image artificial ground feature contour detection method |
| CN107390152B (en) * | 2017-07-14 | 2020-02-18 | 歌尔科技有限公司 | Calibration method and device for magnetometer and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043958A (en) * | 2010-11-26 | 2011-05-04 | 华中科技大学 | High-definition remote sensing image multi-class target detection and identification method |
| CN102750703A (en) * | 2012-06-27 | 2012-10-24 | 北京航空航天大学 | Remote sensing image airport automatic detecting method based on linear cutting |
| CN104408463A (en) * | 2014-10-15 | 2015-03-11 | 中国土地勘测规划院 | High-resolution construction land graph spot identification method based on PanTex and linear characteristic |
-
2015
- 2015-05-04 CN CN201510221008.5A patent/CN104794495B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043958A (en) * | 2010-11-26 | 2011-05-04 | 华中科技大学 | High-definition remote sensing image multi-class target detection and identification method |
| CN102750703A (en) * | 2012-06-27 | 2012-10-24 | 北京航空航天大学 | Remote sensing image airport automatic detecting method based on linear cutting |
| CN104408463A (en) * | 2014-10-15 | 2015-03-11 | 中国土地勘测规划院 | High-resolution construction land graph spot identification method based on PanTex and linear characteristic |
Non-Patent Citations (2)
| Title |
|---|
| 利用梯度信息快速提取直线边缘特征;夏军营等;《中国图象图形学报》;20120831;第17卷(第8期);全文 * |
| 基于边缘连接的快速直线段检测算法;董晶等;《光学学报》;20130331;第33卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104794495A (en) | 2015-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104766084B (en) | A kind of nearly copy image detection method of multiple target matching | |
| CN102855487B (en) | Method for automatically extracting newly added construction land change image spot of high-resolution remote sensing image | |
| CN102930518B (en) | Improved sparse representation based image super-resolution method | |
| CN109800698A (en) | Icon detection method based on depth network | |
| CN103955583B (en) | Method for determining threshold value of urban built-up area extracted through nighttime light data | |
| CN110992366B (en) | Image semantic segmentation method, device and storage medium | |
| CN103699532A (en) | Image color retrieval method and system | |
| CN106649663A (en) | Video copy detection method based on compact video representation | |
| CN104766341B (en) | A kind of remote sensing image building change detecting method based on DSM and Density Estimator | |
| CN101872415A (en) | Video copying detection method being suitable for IPTV | |
| CN104616026A (en) | Monitor scene type identification method for intelligent video monitor | |
| Zhu et al. | Change detection based on the combination of improved SegNet neural network and morphology | |
| CN103955909A (en) | Method and system for manufacturing thematic map by fusing images based on MapGISK9 | |
| CN104794495B (en) | A kind of large format remote sensing image territorial classification method based on straight line statistical nature | |
| Zhou et al. | Deep learning-based instance segmentation for indoor fire load recognition | |
| Wang et al. | Global contextual guided residual attention network for salient object detection | |
| CN101635851B (en) | Method for extracting video fingerprints | |
| CN103714544B (en) | A kind of optimization method based on SIFT feature Point matching | |
| CN103871089B (en) | Image superpixel meshing method based on fusion | |
| CN113139511A (en) | Water and soil conservation production and construction project disturbance pattern spot extraction method, medium and equipment | |
| Shen et al. | HQA‐Trans: An end‐to‐end high‐quality‐awareness image translation framework for unsupervised cross‐domain pedestrian detection | |
| CN107464272A (en) | The interpolation method of central diffusion type meteorological causes isopleth based on key point | |
| CN104850623A (en) | Dynamic extension method and system for multidimensional data analysis model | |
| CN113159044A (en) | Deep learning-based road material identification method for convolutional neural network | |
| WO2021042549A1 (en) | Image mask filter method, device, system, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180206 |