CN101826204A - Quick particle image segmentation method based on improved waterline algorithm - Google Patents

Quick particle image segmentation method based on improved waterline algorithm Download PDF

Info

Publication number
CN101826204A
CN101826204A CN200910118851A CN200910118851A CN101826204A CN 101826204 A CN101826204 A CN 101826204A CN 200910118851 A CN200910118851 A CN 200910118851A CN 200910118851 A CN200910118851 A CN 200910118851A CN 101826204 A CN101826204 A CN 101826204A
Authority
CN
China
Prior art keywords
image
value
point
cut
rule
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN200910118851A
Other languages
Chinese (zh)
Other versions
CN101826204B (en
Inventor
于军玲
肖凯涛
邱琳
武强
李庆伟
李勇
史红星
米浦春
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chinese People's Liberation Army No63976
Original Assignee
Chinese People's Liberation Army No63976
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chinese People's Liberation Army No63976 filed Critical Chinese People's Liberation Army No63976
Priority to CN200910118851A priority Critical patent/CN101826204B/en
Publication of CN101826204A publication Critical patent/CN101826204A/en
Application granted granted Critical
Publication of CN101826204B publication Critical patent/CN101826204B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention relates to a quick and effective particle image segmentation algorithm which is derived from the classic waterline segmentation algorithm and simultaneously avoids the over-segmentation problem of the traditional waterline segmentation algorithm. The method comprises three parts of image pretreatment, segmentation algorithm and segmentation after-treatment, wherein the image pretreatment eliminates the influence of different noises by adopting a variable metric filter, the segmentation algorithm is introduced with the concept of marking arrays and marks the correspondence of all pixel points when the least value of gradient rises to the maximum value of the gradient, and the after-treatment compares the similarity level of adjacent areas by one-step scanning and merge the two areas if the similarity level of areas at two sides of a waterline is smaller than a set threshold value. The method can quickly and accurately segment a target particle from images and greatly enhances the computation speed.

Description

Quick particle image segmentation method based on improved waterline algorithm
Technical field
The present invention relates to a kind of image partition method, especially relate to a kind of quick particle image segmentation method based on improved waterline partitioning algorithm.
Background technology
Numeral particle image analytical technology is widely used in occasions such as particle image analysis such as medical image analysis, airborne particulate graphical analysis, fluid graphical analysis, the grain size analysis of powder China ink, spar image, has a wide range of applications in fields such as material, chemical industry, geological analysis, biomedicine, agricultural breeding, textile material, industry manufacturings.Wherein, it is that computing machine carries out the first step that automatic particle image is analyzed that particle image is separated from the image that collects, and also is the important foundation of image analysis technology.The result that particle image is cut apart can directly have influence on robustness and the validity of later stage to particle image parameters analysis result.Therefore analysis has great importance and economical, societal benefits to particle image to propose simple and effective particle image segmentation method.
As the patented claim number is 200710178737.2 " a kind of method for detecting image edge based on Threshold Segmentation " introduced, this method is all calculated an average and variance to the territory of facing of each pixel, in this zone, calculate the probability density function of makeing mistakes then, and then the threshold value of cutting apart in definite zone.This method is owing to all calculate a segmentation threshold to each pixel, so calculated amount is very big.And in essence, this method is a kind of method of local segmentation, does not consider the effect of global information in cutting apart, so haves much room for improvement on the segmentation effect.
As the patented claim number is 200710120550.7 " a kind of watershed divide image segmentation disposal routes " introduced, adopt to pixel sort, the method for mark judges whether adjacent area.This method has been inherited the thought of waterline partitioning algorithm, but does not solve the problem of over-segmentation in the waterline partitioning algorithm, and owing to be based on the algorithm that pixel repeatedly scans, so counting yield has much room for improvement.
Summary of the invention
The present invention proposes a kind ofly fast based on the particle image partitioning algorithm of waterline algorithm, its thought comes from classical waterline partitioning algorithm.But avoided the over-segmentation problem of traditional waterline algorithm, can cut apart accurately fast particle image.The method that the present invention proposes has solved in the above-mentioned existing patented technology that splitting speed is slow, segmentation effect does not consider that global information and transition problem such as cut apart.
The present invention has adopted improved waterline partitioning algorithm, and this algorithm may further comprise the steps:
1) input gray level image to be split;
2) picture noise filtering is selected the wave filter of different scale according to different image, carries out noise in the convolution operation filtering image to importing image to be split;
3) image segmentation, at first calculate gradient information through the image to be split after the noise filtering processing, set up the mark array according to the minimum and maximum value of gradient information then, all numerical value in the gradient information are referred in the mark array, according to the mark array pixel in the image is classified at last, realize image segmentation;
4) Regional Integration, the image of traversal after cutting apart, the similarity degree of cut-off rule two side areas relatively judges whether that according to similarity degree needs integrate adjacent area in the segmentation effect figure after obtaining integrating.
Picture noise filtering process of the present invention comprises:
1) selective filter.According to the resolution difference of imageing sensor, the wave filter among the present invention is divided into 3 grades: 3 * 3,5 * 5 and 7 * 7; Image is divided into 3 grades: being low equiaccuracy chart picture smaller or equal to the image of 640 * 480 pixels, is the medium accuracy image greater than 640 * 480 but less than the image of 1024 * 768 pixels, is high precision image greater than the image of 1024 * 768 pixels; According to the accuracy selection wave filter of input picture, when input picture was the image of precision such as low, selecting scale was 3 * 3 wave filter; When input picture was the medium accuracy image, selecting scale was 5 * 5 wave filter; When input picture was high precision image, selecting scale was 7 * 7 wave filter;
2) behind the yardstick of selected wave filter, when filtering window slides in input picture, the pixel value that comprises in the wave filter is sorted, the value of size is as the new value of current filter center point pixel in the middle of obtaining.And the like finished by traversal up to the view picture input picture, obtain the image behind the filtering noise.
Image segmentation process concrete steps of the present invention comprise:
1) gradient information of image after the calculating noise filtering obtains the maximum and the minimum value of gradient information, thereby determines the span of gradient, is [minimum value, maximal value];
2) set up the mark array according to the scope of gradient, the span of mark array is consistent with the span of gradient, is all [minimum value, maximal value];
3) getting minimum point is current point, seeks adjacent inferior minimum point and carry out mark near current point, then the point of mark is put into the mark array;
4) with the point of mark as current point, continue to seek the point of current relatively point time minimum value, put into the mark array;
5) repeat said process and all be placed into the mark array up to all points, traversal is finished;
6) according to the mark of mark array to being had a few, with underlined identical point be communicated with processing, constitute cut-off rule, finish image segmentation.
Regional Integration of the present invention is based on the regional union operation of line scanning, and concrete steps comprise:
1) for obtaining cut-off rule image afterwards, it is carried out line scanning, when scanning cut-off rule, similarity measurement is carried out in the zone of cut-off rule both sides;
2) judge regional similarity degree according to metric function, cut-off rule both sides similarity degree is less than the same area that is judged to be of predetermined threshold;
3) carry out Regional Integration, the similarity of cut-off rule both sides is merged into a zone less than two adjacent areas of predetermined threshold;
4) cut-off rule both sides similarity degree keeps former segmentation result constant greater than two adjacent areas of predetermined threshold.
The metric function expression formula that adopts in the Regional Integration algorithm of the present invention is:
σ ( R i , R j ) = n i · n j n i + n j · ( m i - m j ) 2
Wherein, collection of pixels is R i, R iNumber of pixels be n i, gray values of pixel points is p I, j, i=1 .., M, j=1 .., n i, region R iInterior pixel average gray value is
Figure B2009101188515D0000032
The accurate function σ (R that surveys i, R j) more little, the similarity degree of cut-off rule both sides is big more; Predetermined threshold span in the Regional Integration algorithm
Figure B2009101188515D0000033
Wherein k is 0～1 elasticity coefficient.
The gray level image of the present invention's input is the standard grayscale image of tonal range between [0,255], when pending image is coloured image or other image, need be to handle behind the standard grayscale image pending image transitions again.
The present invention has adopted improved waterline partitioning algorithm to carry out image segmentation, utilizes the gradient sudden change information of target particles in the image to cut apart.Basic thought is: the maximum of points of gradient image can be considered as the potential marginal point of particle target, and the minimum point of gradient image can be considered the potential central point of particle target.As seed points, from the outwards constantly expansion of each seed points, the similar water surface is rising steadily with the minimum point in the gradient image, will converge to a time-out when the water surface of zones of different is elevated to, and sets up interphase at meet, finally finishes and cuts apart task.
The advantage of this algorithm mainly contains: 1. splitting speed is fast; 2. can produce the body outline of sealing; 3. locate more accurately, can produce good response more weak edge.This algorithm can be avoided the over-segmentation problem of traditional waterline partitioning algorithm, improves splitting speed, can obtain good segmentation performance with very little calculation cost.
Description of drawings
Fig. 1 is the quick particle image segmentation method process flow diagram based on improved waterline partitioning algorithm;
Fig. 2 is the original image that has salt-pepper noise of input;
Fig. 3 is the image after handling through the picture noise filtering;
The design sketch of Fig. 4 for tentatively cutting apart through image;
Fig. 5 is through the final effect figure after the Regional Integration.
Embodiment
Import original image to be split, as shown in Figure 2.The partitioning algorithm of implementing among the present invention needs following three steps:
The first step, noise filtering is removed the picture noise that produces in the image acquisition process, extracts the Global Information of particle target.
Fig. 2 belongs to low equiaccuracy chart picture, therefore selects the wave filter of 3*3 size to carry out convolutional filtering.Utilize the wave filter of 3*3 yardstick, implement following operation.
Select certain image window 3*3 matrix: [3 54; 123; 98 7]
The ordering of image window interior pixel is [1 2345678 9], and wherein the intermediate point pixel value is 5, and then the pixel value of 3*3 filter center point is set to 5, then wave filter is continued to scan on image window, finishes filtering operation up to full figure.
Through above noise filtering operation, the image behind the removal noise is seen Fig. 3.
Second step, image segmentation: adopt improved waterline algorithm that Fig. 3 is carried out image segmentation.
Idiographic flow is as follows:
1. the gradient information of calculating chart 3 obtains the maximal value (max=63) and the minimum value (min=0) of gradient;
2. set up the mark array, its label range is [0,63], the gradient minimum point that is interconnected is coupled together, and be marked as 0;
3. optional minimum point p1 (0,0) (mark s1), the some p2 (0 that it is adjacent, 1) (being labeled as s2), all is mark s1 as the point around the fruit dot p2, and then changing the some p2 that is labeled as s2 is s1, as a more than mark around the fruit dot p2, then p2 is carried out mark according to nearest neighbouring rule.The rest may be inferred, and near all points reason p1 point is finished;
4. seeking gray-scale value in unmarked point is the minimum point of s2, and the part that can be communicated with is connected to become the zone, and puts into the mark array;
5. returned for the 3rd step, continue untreated gradient minimum point in the marks for treatment array, up to all disposing.
Fig. 3 behind the filtering noise is through the processing of above-mentioned steps, and its effect is seen Fig. 4, has 3 zones and is split.Can see that the part that obvious border is arranged among the figure is successfully split.
The 3rd step, Regional Integration: adopt and integrate based on the region merging algorithm of line scanning.
From the result of Fig. 4 as can be seen, target image is successfully cut apart according to gradient information, and next step only needs adjacent similar area integrated becomes a significant cut zone.Among Fig. 2, the target segmentation effect should be 1 isolated area, and through after the processing of first two steps, also remaining 3 similar cut zone are labeled as regional A, area B, zone C from big to small according to area, as shown in Figure 4.At this moment image has been cut apart by transition.The various piece that transition is cut apart is integrated effectively, and concrete steps are as follows:
1. measuring similarity: regional A and area B are two adjacent cut zone, the similarity measurement function of zoning A and area B
σ ( R i , R j ) = n i · n j n i + n j · ( m i - m j ) 2
2. regional A respective pixel set is R i, R wherein iNumber of pixels be n i, wherein gray values of pixel points is p I, j, i=1 .., M, j=1 ..., n i, R iInterior pixel average gray value is In like manner, the set of area B respective pixel is R j, number of pixels is n j, R jInterior pixel average gray value is m j
3. the zone merges: get elasticity coefficient k=0.01, predetermined threshold is in view of the above:
0.01 × n i · n j n i + n j · 255 2
Calculate, therefore the similarity measurement functional value of regional A and area B, carries out the zone and merges less than described predetermined threshold;
4. in the same manner, regional A and zone C are also carried out regional union operation;
Finally, wait to integrate regional A for 3 that comprise in the image, area B, zone C is integrated into 1 effective coverage, and the segmentation effect figure after the integration sees Fig. 5, is our target split image.

Claims (6)

1. the quick particle image segmentation method of a waterline algorithm is characterized in that this method has adopted improved waterline algorithm, specifically may further comprise the steps:
1) input gray level image to be split;
2) picture noise filtering is selected the wave filter of different scale according to different image, carries out noise in the convolution operation filtering image to importing image to be split;
3) image segmentation, at first calculate gradient information through the image to be split after the noise filtering processing, set up the mark array according to the minimum and maximum value of gradient information then, all numerical value in the gradient information are referred in the mark array, according to the mark array pixel in the image is classified at last, realize image segmentation;
4) Regional Integration, the image of traversal after cutting apart, the similarity degree of cut-off rule two side areas relatively judges whether that according to similarity degree needs integrate adjacent area in the segmentation effect figure after obtaining integrating.
2. quick particle image segmentation method according to claim 1 is characterized in that described picture noise filtering process comprises:
1) selective filter.According to the resolution difference of imageing sensor, the wave filter among the present invention is divided into 3 grades: 3 * 3,5 * 5 and 7 * 7; Image is divided into 3 grades: being low equiaccuracy chart picture smaller or equal to the image of 640 * 480 pixels, is the medium accuracy image greater than 640 * 480 but less than the image of 1024 * 768 pixels, is high precision image greater than the image of 1024 * 768 pixels; According to the accuracy selection wave filter of input picture, when input picture was the image of precision such as low, selecting scale was 3 * 3 wave filter; When input picture was the medium accuracy image, selecting scale was 5 * 5 wave filter; When input picture was high precision image, selecting scale was 7 * 7 wave filter;
2) behind the yardstick of selected wave filter, when filtering window slides in input picture, the pixel value that comprises in the wave filter is sorted, the value of size is as the new value of current filter center point pixel in the middle of obtaining.And the like finished by traversal up to the view picture input picture, obtain the image behind the filtering noise.
3. quick particle image segmentation method according to claim 1 is characterized in that described image segmentation process concrete steps comprise:
1) gradient information of image after the calculating noise filtering obtains the maximum and the minimum value of gradient information, thereby determines the span of gradient, is [minimum value, maximal value];
2) set up the mark array according to the scope of gradient, the span of mark array is consistent with the span of gradient, is all [minimum value, maximal value];
3) getting minimum point is current point, seeks adjacent inferior minimum point and carry out mark near current point, then the point of mark is put into the mark array;
4) with the point of mark as current point, continue to seek the point of current relatively point time minimum value, put into the mark array;
5) repeat said process and all be placed into the mark array up to all points, traversal is finished;
6) according to the mark of mark array to being had a few, with underlined identical point be communicated with processing, constitute cut-off rule, finish image segmentation.
4. quick particle image segmentation method according to claim 1 is characterized in that described Regional Integration process concrete steps comprise:
1) for obtaining cut-off rule image afterwards, it is carried out line scanning, when scanning cut-off rule, similarity measurement is carried out in the zone of cut-off rule both sides;
2) judge regional similarity degree according to metric function, cut-off rule both sides similarity degree is less than the same area that is judged to be of predetermined threshold;
3) carry out Regional Integration, the similarity of cut-off rule both sides is merged into a zone less than two adjacent areas of predetermined threshold;
4) cut-off rule both sides similarity degree keeps former segmentation result constant greater than two adjacent areas of predetermined threshold.
5. quick particle image segmentation method according to claim 4 is characterized in that the metric function expression formula that adopts in the Regional Integration algorithm is:
σ ( R i , R j ) = n i · n j n i + n j · ( m i - m j ) 2
Wherein, collection of pixels is R i, R iNumber of pixels be n i, gray values of pixel points is p I, j, i=1 ..., M, j=1 .., n i, region R iInterior pixel average gray value is
Figure F2009101188515C0000022
The accurate function σ (R that surveys i, R j) more little, the similarity degree of cut-off rule both sides is big more; Predetermined threshold span in the Regional Integration algorithm
Figure F2009101188515C0000023
Wherein k is 0～1 elasticity coefficient.
6. quick particle image segmentation method according to claim 1, it is characterized in that the gray level image of importing is that tonal range is [0,255] the standard grayscale image between, when pending image is coloured image or other image, need be to handle again behind the standard grayscale image pending image transitions.
CN200910118851A 2009-03-04 2009-03-04 Quick particle image segmentation method based on improved waterline algorithm Expired - Fee Related CN101826204B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN200910118851A CN101826204B (en) 2009-03-04 2009-03-04 Quick particle image segmentation method based on improved waterline algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN200910118851A CN101826204B (en) 2009-03-04 2009-03-04 Quick particle image segmentation method based on improved waterline algorithm

Publications (2)

Publication Number Publication Date
CN101826204A true CN101826204A (en) 2010-09-08
CN101826204B CN101826204B (en) 2012-09-26

Family

ID=42690110

Family Applications (1)

Application Number Title Priority Date Filing Date
CN200910118851A Expired - Fee Related CN101826204B (en) 2009-03-04 2009-03-04 Quick particle image segmentation method based on improved waterline algorithm

Country Status (1)

Country Link
CN (1) CN101826204B (en)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102999924A (en) * 2011-09-09 2013-03-27 富士施乐株式会社 Image processing apparatus and image processing method
CN103413301A (en) * 2013-07-27 2013-11-27 信阳师范学院 Color image cutting method based on spherical particle calculation
CN103512883A (en) * 2013-09-29 2014-01-15 中国科学院半导体研究所 Digital image processing based method for detecting geometrical characteristics of impurity in polyolefin material
CN103632361A (en) * 2012-08-20 2014-03-12 阿里巴巴集团控股有限公司 An image segmentation method and a system
CN104915630A (en) * 2014-03-14 2015-09-16 欧姆龙株式会社 Image processing apparatus and segmentation method
CN105095846A (en) * 2014-09-28 2015-11-25 航天恒星科技有限公司 Method and system for extracting region growing seed points based on remote sensing images and sea-land segmentation
CN105279748A (en) * 2014-06-18 2016-01-27 江南大学 Rapid pixel size calculation method
CN109961451A (en) * 2019-03-22 2019-07-02 西北工业大学 A kind of material grains tissue segmentation methods based on marginal information
CN112053360A (en) * 2020-10-10 2020-12-08 腾讯科技（深圳）有限公司 Image segmentation method and device, computer equipment and storage medium
CN113436091A (en) * 2021-06-16 2021-09-24 中国电子科技集团公司第五十四研究所 Object-oriented remote sensing image multi-feature classification method
CN117562098A (en) * 2024-01-15 2024-02-20 河南科技学院 Pig small-lining separation method
CN117562098B (en) * 2024-01-15 2024-05-03 河南科技学院 Pig small-lining separation method

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110322445B (en) * 2019-06-12 2021-06-22 浙江大学 Semantic segmentation method based on maximum prediction and inter-label correlation loss function

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100505884C (en) * 2007-08-21 2009-06-24 北京大学软件与微电子学院 A shed image division processing method
CN100512374C (en) * 2007-12-05 2009-07-08 北京航空航天大学 A method for image edge detection based on threshold sectioning

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102999924A (en) * 2011-09-09 2013-03-27 富士施乐株式会社 Image processing apparatus and image processing method
CN103632361A (en) * 2012-08-20 2014-03-12 阿里巴巴集团控股有限公司 An image segmentation method and a system
CN103413301B (en) * 2013-07-27 2016-02-10 信阳师范学院 A kind of color image segmentation method based on spherical Granule Computing
CN103413301A (en) * 2013-07-27 2013-11-27 信阳师范学院 Color image cutting method based on spherical particle calculation
CN103512883A (en) * 2013-09-29 2014-01-15 中国科学院半导体研究所 Digital image processing based method for detecting geometrical characteristics of impurity in polyolefin material
CN104915630A (en) * 2014-03-14 2015-09-16 欧姆龙株式会社 Image processing apparatus and segmentation method
CN104915630B (en) * 2014-03-14 2018-06-29 欧姆龙株式会社 Image processing apparatus and region segmentation method
CN105279748B (en) * 2014-06-18 2017-11-17 江南大学 A kind of quickly pixel dimension computational methods
CN105279748A (en) * 2014-06-18 2016-01-27 江南大学 Rapid pixel size calculation method
CN105095846A (en) * 2014-09-28 2015-11-25 航天恒星科技有限公司 Method and system for extracting region growing seed points based on remote sensing images and sea-land segmentation
CN105095846B (en) * 2014-09-28 2019-07-30 航天恒星科技有限公司 Region growing seed point extracting method and system towards the segmentation of remote sensing images sea land
CN109961451A (en) * 2019-03-22 2019-07-02 西北工业大学 A kind of material grains tissue segmentation methods based on marginal information
CN112053360A (en) * 2020-10-10 2020-12-08 腾讯科技（深圳）有限公司 Image segmentation method and device, computer equipment and storage medium
CN112053360B (en) * 2020-10-10 2023-07-25 腾讯科技（深圳）有限公司 Image segmentation method, device, computer equipment and storage medium
CN113436091A (en) * 2021-06-16 2021-09-24 中国电子科技集团公司第五十四研究所 Object-oriented remote sensing image multi-feature classification method
CN117562098A (en) * 2024-01-15 2024-02-20 河南科技学院 Pig small-lining separation method
CN117562098B (en) * 2024-01-15 2024-05-03 河南科技学院 Pig small-lining separation method

Also Published As

Publication number Publication date
CN101826204B (en) 2012-09-26

Similar Documents

Publication Publication Date Title
CN101826204B (en) Quick particle image segmentation method based on improved waterline algorithm
CN101551456B (en) Method for detecting water area margin of SAR image based on improved shearlet transformation
CN102032875B (en) Image-processing-based cable sheath thickness measuring method
CN102013015B (en) Object-oriented remote sensing image coastline extraction method
CN101634705B (en) Method for detecting target changes of SAR images based on direction information measure
CN105825169B (en) A kind of pavement crack recognition methods based on road image
CN103761743B (en) A kind of solid wooden floor board detection method of surface flaw based on image co-registration segmentation
CN104899892B (en) A kind of quickly star map image asterism extracting method
CN102214290B (en) License plate positioning method and license plate positioning template training method
CN102508219A (en) Turbulent current target detection method of wind profiler radar
CN110378921B (en) Intelligent identification method for substrate layer boundary of channel based on floating mud rheological property and gray level co-occurrence matrix
CN104751474A (en) Cascade quick image defect segmentation method
CN112001218A (en) Three-dimensional particle category detection method and system based on convolutional neural network
CN105976376A (en) High resolution SAR image target detection method based on part model
CN103559724A (en) Method for synchronously tracking multiple cells in high-adhesion cell environment
CN102360503B (en) SAR (Specific Absorption Rate) image change detection method based on space approach degree and pixel similarity
CN108122221A (en) The dividing method and device of diffusion-weighted imaging image midbrain ischemic area
CN105787950A (en) Infrared image sea-sky-line detection algorithm based on line gradient accumulation
CN112415536A (en) Method for automatically acquiring abnormal area of vehicle-mounted laser point cloud driving track
CN110176020A (en) A kind of bird's nest impurity method for sorting merging 2D and 3D rendering
CN115049640A (en) Road crack detection method based on deep learning
CN104966091A (en) Strip mine road extraction method based on unmanned plane remote sensing images
CN202433521U (en) Wind profile radar turbulence target detection processing plate
CN103065296B (en) High-resolution remote sensing image residential area extraction method based on edge feature
CN101839980A (en) Unsupervised remote sensing image change detection method based on segmentation window

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20120926

Termination date: 20140304