CN108921025A - A kind of object level classification samples automatic selecting method of collaborative variation detection - Google Patents
A kind of object level classification samples automatic selecting method of collaborative variation detection Download PDFInfo
- Publication number
- CN108921025A CN108921025A CN201810556010.1A CN201810556010A CN108921025A CN 108921025 A CN108921025 A CN 108921025A CN 201810556010 A CN201810556010 A CN 201810556010A CN 108921025 A CN108921025 A CN 108921025A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- sample
- sensing image
- image
- pixel
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a kind of remote sensing image object level classification samples automatic selecting methods of collaborative variation detection, under the premise of obtaining two phase of the same area remote sensing image, multi-scale division is carried out in the way of uniformly drifting about to the image of new phase, obtain the object bounds information of atural object, detection is changed to two phase images simultaneously, obtains constant pixel；" constant " information MAP relationship between two phase images is further established on the basis of " constant " pixel position, to carry out migration of the original sample information on correspondence " constant " position；Then, it is constraint with the vector boundary that new phase Image Segmentation obtains, extracts " constant " object and its sample class label information；Finally, carrying out Sample purification using object association attributes, the object that part has wrong class label information is rejected, finally establishes the object level sample database of new image, the classification for new phase remote sensing image.
Description
Technical field
The present invention relates to remote sensing technology fields, and in particular to a kind of remote sensing image object level classification sample of collaborative variation detection
This automatic selecting method.
Background technique
Remote sensing can be with the acquisition surface data of quick-speed large-scale.Remote sensing image application in, classification be still it is most basic most
The problem of core, although there are many more mature sorting algorithm, nicety of grading and speed issues still not to have
To very good solution.The artificial visual interpretation mode classification of early stage has the shortcomings that expend a large amount of manpowers and time.With calculating
The advantage of the development of machine technology, machine sort mode gradually shows, and wherein the Pixel-level mode classification of early stage divides between low-to-medium altitude
It is widely used in resolution remote sensing image, has pushed the application of Classification in Remote Sensing Image technology.But in recent years, remotely-sensed data
Spatial resolution it is higher and higher, while its data volume is huge, and background information is complicated, and noise information serious interference, " jljl is different
Spectrum " and " foreign matter is with spectrum " phenomenon are obvious, and traditional Pixel-level classification method precision is difficult to meet practical application, and object level is classified
Increasingly prominent its advantage out of method.In object level assorting process, no longer using pixel as basic unit, but Image Segmentation is utilized
The object of acquisition is as minimum classification unit.Be effectively applied one of this mode is on condition that have a large amount of object level sample
This needs to expend a large amount of manpower and object but if the repeated acquisition for carrying out sample when classifying to each issue of image is chosen
Power cost, this is to carry out the bottleneck problem encountered when the application of big region long period using object level classification at present.(pertinent literature：
1.Schowengerdt R A.Techniques for image processing and classifications in
remote sensing[M].Academic Press,2012.2.MENNISJ.,GUO D.Spatial Data Mining
and Geographic Knowledge Discovery:An Introduction[J].Computers,Environment
and Urban Systems,2009,33(6):403-408.3.XIA Liegang.Study on Automatic
Classificati on Method for Remotely Sensed Imagery by Incorporating Spatial-
Spectral Features[D].Hangzhou:Zhejiang University of Technology,2011.)
Summary of the invention
For the problem that the shortcomings of the prior art, i.e. object level sample is difficult to reuse, purport of the present invention
Iing is proposed a kind of sample automatic selecting method for the classification of remote sensing image object level for being able to carry out collaborative variation detection.
The present invention provides a kind of remote sensing image object level classification samples automatic selecting methods of collaborative variation detection, including
Following step：
1) at least two phase remote sensing images in same survey region, i.e., newest remote sensing image and old remote sensing image are obtained；
2) to newest remote sensing image carry out multi-scale division, obtain wherein the object level vector boundary of atural object, texture and
Spectral information；
3) the variation detection of pixel grade is carried out to the two phases remote sensing image, extracts the position of wherein not changed pixel
Distribution, and not changed picture dot is established as sample label；
It 4) is constraint with vector boundary described in step 2), the sample label in conjunction with described in step 3) is established and do not become
The object level sample of change, at the same according to the position distribution of the not changed picture dot set up newest remote sensing image with it is old distant
Feel the variation corresponding relationship between image；
5) change corresponding relationship according to step 4), the sample class label in old remote sensing image is migrated to newest
In remote sensing image, realize in newest remote sensing image to the automatic label class label of constant pixel；
6) threshold value, row information of going forward side by side purifying are arranged to object level sample described in step 4)；
7) according to the sample label marked in step 5) automatically, by the object level sample established in step 4) migrate to
It in newest remote sensing image, is verified using texture described in step 2) and spectral information, rejects wherein error sample, building is most
The object level sample database of new remote sensing image, the training for subsequent classification.
Preferably, at least two phase remote sensing images come from the same sensor in the step 1), have similar spectrum
And space rate respectively.
Preferably, multi-scale division is carried out to newest remote sensing image using mean shift process in the step 2), with
It realizes the extraction of homogeneity primitive object, and calculates vector boundary, texture and the spectral information of extracted object.
Preferably, it is extracted in the step 3) using the binarization segmentation that big law is realized described not changed
The position distribution of pixel.
Preferably, in the step 6) foundation of setting threshold value be the size of object, mark the pixel scale of label with
And similar label pixel proportion.
Preferably, the threshold value is used for the screening of object level sample, to reject wherein area or mark label pixel
The lesser object of ratio.
Preferably, the object level sample is marked as the highest sample label of wherein proportion, to it is newest
The sample label marked automatically in remote sensing image mutually identifies matching.
The beneficial effects of the invention are as follows：The present invention relates to the remote sensing image object level classification samples of collaborative variation detection are automatic
Selection method is suitable for solving the problems, such as the updated sample collection of remote sensing image.Sample automatic selecting method through the invention,
It is adapted to the high spatial resolution remote sense image classificating requirement of time-intensive, compared to the artificial side of acquisition of traditional sample
Formula, this method can obtain promotion in terms of efficiency and speed；The base that this method is expressed in remote sensing multi-scale division and feature quantification
On plinth, couple variations test problems carry out automatically selecting for object level classification samples, and sample is overcome to need manually to acquire
Problem；Meanwhile this method combines current remote sensing image data and priori sample data, detects and goes through by variation
The automatic collection to new image classification sample is completed in the guidance of history auxiliary data, and then current new image can be automated
Object level classification.
Detailed description of the invention
Fig. 1 is the remote sensing image object level classification samples automatic selecting method stream of collaborative variation according to the present invention detection
Journey schematic diagram；
Fig. 2 is mean shift segmentation schematic illustration；
Wherein, circle is the computer capacity drifted about every time, radius r, central point An；
Fig. 3 is that Da-Jin algorithm changes detection schematic diagram；
Wherein, vertical line corresponding position is binary segmentation point；
Fig. 4 is the object level sample distribution schematic diagram automatically selected；
Fig. 5 is based on the result schematic diagram for automatically selecting sample and carrying out new phase classification of remote-sensing images.
Specific embodiment
Present invention will be described in further detail below with reference to the accompanying drawings, to enable those skilled in the art referring to specification text
Word can be implemented accordingly.
It should be appreciated that such as " having ", "comprising" and " comprising " term used herein are not discharged one or more
The presence or addition of a other elements or combinations thereof.
The present invention is further illustrated below by way of specific embodiment.But the detail of embodiment is only used for explaining this hair
It is bright, it should not be construed as limited overall technical solution.
Embodiment 1
Main idea is that by finding " constant " atural object on two phase remote sensing images, and based on this, it builds
The mapping relations in " constant " region between vertical two phase images, while being moved to the sample information of image early period using the mapping relations
On new image, sample Autonomic Migration Framework and reuse on different time are realized.
Premise of the invention is the two phase remote sensing images for obtaining same survey region, it is desirable that this two phases image has similar light
Spectrum and spatial resolution, i.e. optimal case are this two phases images from same sensor, and the auxiliary remote sensing image of early period has
Standby corresponding sample or sorting result information.Specific step is as follows：
1) multi-scale division is carried out by the way of average drifting to the image of new phase, realizes mentioning for homogeneity primitive object
It takes；
2) spectral signature, shape feature and the textural characteristics that object has been obtained in step 1) are calculated；
3) variation for carrying out pixel rank to two phase images detects, and extracts not changed picture in conjunction with Da-Jin algorithm binaryzation
Member；
4) recording step 3) " constant " Pixel domain position in testing result, i.e., the position point of not changed picture dot
Cloth, and the incidence relation of usable samples information between two phase images is established on this basis；
5) incidence relation established according to step 4), by the sample class label information of image early period at " constant " pixel
It moves in new image, realizes on new image the sample class mark of " constant " pixel；
6) the object bounds information and step 4) extracted using step 1) are changed " constant " pixel obtained after detection,
Intercept out " constant " object；
7) setting threshold value screens " constant " object that step 6) generates, and area is smaller for rejecting, obtains label in object
The lesser object of area ratio, is purified on this basis, obtains more pure object, and accounting example highest in object
Class label assign object, to obtain the sample of object level；
8) result of step 7) is examined using the spectrum and texture information of step 2) statistics again, it is wrong further rejects part
Accidentally sample；
9) sample database established using step 8) can carry out the automatic classification of object level to new phase remote sensing image.
Wherein, step 1) carries out image multi-scale division by the way of average drifting, extracts the imaged object of homogeneity；
Step 3) utilizes the variation testing result of pixel grade, counts " constant " pixel, and the variation mapping established between image is closed
System；
Step 5) carries out the historical sample information transfer in " constant " region using the mapping relations established；
Step 7) carries out sample rejecting by the way that multiple threshold values are arranged, and obtains pure reliable object by Sample purification technology
Grade sample.
Embodiment 2
Fig. 1 illustrates main realization approach of the invention.Crucial step of the invention includes mutual between remote sensing image
Match, including space geometry matching and spectral radiance matching；New phase remote sensing image is carried out in the way of average drifting multiple dimensioned
Earth object is extracted in segmentation；Pixel grade between different phase images changes detection, utilizes the spatial position between " constant " pixel
Relationship establishes the mapping of sample information between image；The migration of sample class label is carried out using the mapping relations of foundation；In new image
It is upper that " constant " object and its sample label are constructed based on the sample information migrated in multi-scale division result and " constant " pixel；It is logical
It crosses given threshold to purify the object samples of new image, rejects error sample, it is final to obtain more reliable new samples number
According to training process for subsequent classification.
Fig. 2 illustrates mean shift segmentation principle.Average drifting basic thought be by setting multiple initial centers at random,
And certain radius is set up according to these points, the vector sum in radius is calculated, is further in new with the position that vector is directed toward
The heart, continuous iteration, until meeting some requirements, as cluster centre.Specific step is as follows for mean shift segmentation：
1) starting point of 1 point A1 as starting segmentation is randomly selected, and sets radius r；
2) point centered on set point calculates all elements vector sum in radius r, obtains ‖ shift1 ‖；
3) set the position of new starting point A1 ' as A1+shift1, i.e. A1 along the direction shift1 move ‖ shift1 ‖ away from
From；
4) iterative process 3), until A1 ' restrains；
5) repeat step 1)~step 4), until whole picture image be traversed and each central point restrain；
6) given threshold then carries out center merging if it is less than the threshold value of setting；
7) remaining each point is divided to the access frequency of each point in the central point ergodic process that foundation obtains, realizes image
Segmentation.
Fig. 3 illustrates the principle that Da-Jin algorithm is changed testing result binaryzation, and main purpose is entire by traversing
Image calculates maximum between-cluster variance, and then determines segmentation threshold, if 0.53 is the segmentation threshold in this experiment in figure, that is, is less than
0.53 pixel is divided into one kind, and assignment 0, and the pixel greater than 0.53 is divided into one kind and is assigned a value of 1.It is specific as follows：
For piece image, foreground pixel proportion is ω0, average gray μ0, background pixel proportion is ω1,
Average gray is μ1, the average gray of image is μ, and inter-class variance g, image size is M × N, and T is the threshold value of object brightness,
ω0And ω1It is determined by T, i.e., the ratio that the pixel greater than T accounts for is ω0, the pixel proportion less than T is ω1.Thus
Known to：
ω0+ω1=1；
μ=ω0*μ0+ω1*μ1；
G=ω0*(μ0-μ)2+ω1*(μ1-μ)2；
G can be calculated after traversing whole image, can determine T on this basis, realize the two-value point of image
It cuts, extracts interested " constant " pixel.
Embodiment 3
It is test image with two phase SPOT5 of Dongguan City, has carried out automatically selecting for sample, effect by the method for the present invention
As shown in Figure 4.It can be seen that the landscape ground, meadow, construction land, waters, arable land and wasteland in image establish it is new
Object level sample.The automatic classification of model training and atural object can be carried out on the basis of sample is established, as a result as shown in Figure 5.
By verifying, user's precision is as follows：Arable land is 89.6%, landscape ground 92.6%, meadow 68.91%, and construction land is
59.1%, waters 69.2%, wasteland 53.3%, overall accuracy 80.6%, kappa coefficient is 0.7379.In this sample
In this autoselect process, artificial visual acquisition, and the subsequent classification result precision base based on auto selecting swatches are not relied on
Originally it may conform to require, therefore the remote sensing image object level classification samples automatic selecting method of this collaborative variation detection of the invention
With preferable application effect, can be promoted.
Although the embodiments of the present invention have been disclosed as above, but its is not only in the description and the implementation listed
With.It can be applied to various suitable the field of the invention completely.It for those skilled in the art, can be easily
Realize other modification.Therefore without departing from the general concept defined in the claims and the equivalent scope, the present invention is simultaneously unlimited
In specific details and embodiment shown here.
Claims (7)
1. a kind of object level classification samples automatic selecting method of collaborative variation detection, which is characterized in that include the following steps：
1) at least two phase remote sensing images in same survey region, i.e., newest remote sensing image and old remote sensing image are obtained；
2) multi-scale division is carried out to newest remote sensing image, obtains object level vector boundary, texture and the spectrum of wherein atural object
Information；
3) the variation detection of pixel grade is carried out to the two phases remote sensing image, extracts the position point of wherein not changed pixel
Cloth, and not changed picture dot is established as sample label；
It 4) is constraint with vector boundary described in step 2), the sample label in conjunction with described in step 3) is established not changed
Object level sample, while newest remote sensing image and old remote sensing shadow are set up according to the position distribution of the not changed picture dot
Variation corresponding relationship as between；
5) change corresponding relationship according to step 4), the sample label in old remote sensing image is migrated to newest remote sensing image
In, it realizes in newest remote sensing image to the automatic label class label of constant pixel；
6) threshold value, row information of going forward side by side purifying are arranged to object level sample described in step 4)；
7) according to the sample label marked automatically in step 5), the object level sample established in step 4) is migrated to newest
It in remote sensing image, is verified using texture described in step 2) and spectral information, rejects wherein error sample, constructed newest distant
Feel the object level sample database of image, the training for subsequent classification.
2. automatic selecting method according to claim 1, which is characterized in that at least two phase remote sensing images in the step 1)
From the same sensor, there is similar spectrum and space rate respectively.
3. automatic selecting method according to claim 1, which is characterized in that use mean shift process in the step 2)
Multi-scale division is carried out to newest remote sensing image, to realize the extraction of homogeneity primitive object, and calculates the vector of extracted object
Boundary, texture and spectral information.
4. automatic selecting method according to claim 1, which is characterized in that realized in the step 3) using big law
Binarization segmentation extract the position distribution of the not changed pixel.
5. automatic selecting method according to claim 1, which is characterized in that the foundation of setting threshold value is in the step 6)
The size of object, the pixel scale for marking label and similar label pixel proportion.
6. automatic selecting method according to claim 5, which is characterized in that the threshold value is used for the sieve of object level sample
Choosing, to reject wherein area or the mark lesser object of label pixel scale.
7. automatic selecting method according to claim 6, which is characterized in that the object level sample is marked as wherein institute
The highest class label of accounting example, matches mutually to identify with the sample label marked automatically in newest remote sensing image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810556010.1A CN108921025A (en) | 2018-06-01 | 2018-06-01 | A kind of object level classification samples automatic selecting method of collaborative variation detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810556010.1A CN108921025A (en) | 2018-06-01 | 2018-06-01 | A kind of object level classification samples automatic selecting method of collaborative variation detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108921025A true CN108921025A (en) | 2018-11-30 |
Family
ID=64418661
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810556010.1A Pending CN108921025A (en) | 2018-06-01 | 2018-06-01 | A kind of object level classification samples automatic selecting method of collaborative variation detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108921025A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144487A (en) * | 2019-12-27 | 2020-05-12 | 二十一世纪空间技术应用股份有限公司 | Method for establishing and updating remote sensing image sample library |
| CN111339800A (en) * | 2018-12-18 | 2020-06-26 | 中科星图股份有限公司 | Method and device for producing remote sensing sample data and computer readable storage medium |
| CN111597861A (en) * | 2019-02-21 | 2020-08-28 | 中科星图股份有限公司 | System and method for automatically interpreting ground object of remote sensing image |
| CN112308024A (en) * | 2020-11-23 | 2021-02-02 | 中国水利水电科学研究院 | Water body information extraction method |
| CN112884791A (en) * | 2021-02-02 | 2021-06-01 | 重庆市地理信息和遥感应用中心 | Method for constructing large-scale remote sensing image semantic segmentation model training sample set |
| CN113128565A (en) * | 2021-03-25 | 2021-07-16 | 之江实验室 | Automatic image annotation system and device oriented to agnostic pre-training annotation data |
| CN116681873A (en) * | 2023-07-31 | 2023-09-01 | 山东省国土测绘院 | Image orthorectification method and system based on rapid updating of digital elevation model |
| CN117994062A (en) * | 2024-04-01 | 2024-05-07 | 中国科学院东北地理与农业生态研究所 | Crop dynamic monitoring method and device |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899897A (en) * | 2015-05-27 | 2015-09-09 | 中国科学院地理科学与资源研究所 | High-resolution remote-sensing image land cover change detection method based on history data mining |
-
2018
- 2018-06-01 CN CN201810556010.1A patent/CN108921025A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899897A (en) * | 2015-05-27 | 2015-09-09 | 中国科学院地理科学与资源研究所 | High-resolution remote-sensing image land cover change detection method based on history data mining |
Non-Patent Citations (1)
| Title |
|---|
| 吴田军 等: "迁移学习支持下的遥感影像对象级分类样本自动选择方法", 《测绘学报》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111339800A (en) * | 2018-12-18 | 2020-06-26 | 中科星图股份有限公司 | Method and device for producing remote sensing sample data and computer readable storage medium |
| CN111597861A (en) * | 2019-02-21 | 2020-08-28 | 中科星图股份有限公司 | System and method for automatically interpreting ground object of remote sensing image |
| CN111144487A (en) * | 2019-12-27 | 2020-05-12 | 二十一世纪空间技术应用股份有限公司 | Method for establishing and updating remote sensing image sample library |
| CN111144487B (en) * | 2019-12-27 | 2023-09-26 | 二十一世纪空间技术应用股份有限公司 | Method for establishing and updating remote sensing image sample library |
| CN112308024A (en) * | 2020-11-23 | 2021-02-02 | 中国水利水电科学研究院 | Water body information extraction method |
| CN112884791A (en) * | 2021-02-02 | 2021-06-01 | 重庆市地理信息和遥感应用中心 | Method for constructing large-scale remote sensing image semantic segmentation model training sample set |
| CN113128565A (en) * | 2021-03-25 | 2021-07-16 | 之江实验室 | Automatic image annotation system and device oriented to agnostic pre-training annotation data |
| CN116681873A (en) * | 2023-07-31 | 2023-09-01 | 山东省国土测绘院 | Image orthorectification method and system based on rapid updating of digital elevation model |
| CN116681873B (en) * | 2023-07-31 | 2023-10-27 | 山东省国土测绘院 | Image orthorectification method and system based on rapid updating of digital elevation model |
| CN117994062A (en) * | 2024-04-01 | 2024-05-07 | 中国科学院东北地理与农业生态研究所 | Crop dynamic monitoring method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108921025A (en) | A kind of object level classification samples automatic selecting method of collaborative variation detection | |
| Kavzoglu et al. | Parameter-based performance analysis of object-based image analysis using aerial and Quikbird-2 images | |
| CN101840581B (en) | Method for extracting profile of building from satellite remote sensing image | |
| CN107784657A (en) | A kind of unmanned aerial vehicle remote sensing image partition method based on color space classification | |
| CN103632363B (en) | Object level high-resolution remote sensing image change detecting method based on Multiscale Fusion | |
| CN104182985B (en) | Remote sensing image change detection method | |
| CN100595782C (en) | Classification method for syncretizing optical spectrum information and multi-point simulation space information | |
| CN107563413A (en) | The accurate extracting method of unmanned plane image farmland block object | |
| CN111191628B (en) | Remote sensing image earthquake damage building identification method based on decision tree and feature optimization | |
| CN107330422A (en) | A kind of method for carrying out mima type microrelief classification to semiarid zone based on high accuracy number elevation model | |
| CN104182763A (en) | Plant type identification system based on flower characteristics | |
| CN104282026A (en) | Distribution uniformity assessment method based on watershed algorithm and minimum spanning tree | |
| CN106875481A (en) | A kind of preparation method of three-dimensional visualization remote sensing image Surface classification model | |
| CN108268527A (en) | A kind of method for detecting land use pattern variation | |
| CN108537286A (en) | A kind of accurate recognition methods of complex target based on key area detection | |
| CN106228136A (en) | Panorama streetscape method for secret protection based on converging channels feature | |
| CN107657207A (en) | A kind of forest land sorting technique based on remote sensing image | |
| Qin et al. | OpenGF: An ultra-large-scale ground filtering dataset built upon open ALS point clouds around the world | |
| CN109146876A (en) | A kind of mine environment change detecting method based on high score remote sensing image | |
| CN106384074A (en) | Detection apparatus of pavement defects and method thereof, and image processing equipment | |
| CN109993042A (en) | A kind of face identification method and its device | |
| Zhao et al. | Ground surface recognition at voxel scale from mobile laser scanning data in urban environment | |
| CN103984758A (en) | Remote sensing image time sequence clustering method for conducting separated treatment according to cloud pixel number | |
| CN110147780A (en) | The landform recognition methods of real-time field robot and system based on level landform | |
| Baluyan et al. | Novel approach for rooftop detection using support vector machine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181130 |
|
| RJ01 | Rejection of invention patent application after publication |