CN106446875A - County remote-sensing scale-oriented crop planting area information extraction method and device - Google Patents
County remote-sensing scale-oriented crop planting area information extraction method and device Download PDFInfo
- Publication number
- CN106446875A CN106446875A CN201611006469.1A CN201611006469A CN106446875A CN 106446875 A CN106446875 A CN 106446875A CN 201611006469 A CN201611006469 A CN 201611006469A CN 106446875 A CN106446875 A CN 106446875A
- Authority
- CN
- China
- Prior art keywords
- segmentation
- factor
- image
- scale
- crop
- 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
- 238000000605 extraction Methods 0.000 title claims abstract description 14
- 230000011218 segmentation Effects 0.000 claims abstract description 73
- 238000000034 method Methods 0.000 claims abstract description 57
- 230000006870 function Effects 0.000 claims description 40
- 230000000694 effects Effects 0.000 claims description 24
- 230000003595 spectral effect Effects 0.000 claims description 19
- 238000012512 characterization method Methods 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 17
- 238000007635 classification algorithm Methods 0.000 claims description 13
- VMXUWOKSQNHOCA-UKTHLTGXSA-N ranitidine Chemical compound [O-][N+](=O)\C=C(/NC)NCCSCC1=CC=C(CN(C)C)O1 VMXUWOKSQNHOCA-UKTHLTGXSA-N 0.000 claims description 11
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000005520 cutting process Methods 0.000 claims description 7
- 238000001228 spectrum Methods 0.000 claims description 7
- 235000007164 Oryza sativa Nutrition 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 6
- 238000003709 image segmentation Methods 0.000 claims description 6
- 210000000582 semen Anatomy 0.000 claims description 6
- 238000000926 separation method Methods 0.000 claims description 5
- 241000209094 Oryza Species 0.000 claims description 3
- 240000007594 Oryza sativa Species 0.000 claims description 3
- 235000009566 rice Nutrition 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 3
- 238000004321 preservation Methods 0.000 claims description 2
- 239000010410 layer Substances 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 7
- 241000196324 Embryophyta Species 0.000 description 6
- 230000008901 benefit Effects 0.000 description 6
- 238000009313 farming Methods 0.000 description 5
- 235000013339 cereals Nutrition 0.000 description 4
- 230000004927 fusion Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 239000000654 additive Substances 0.000 description 2
- 230000000996 additive effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 239000011229 interlayer Substances 0.000 description 2
- 150000003839 salts Chemical class 0.000 description 2
- 241000894007 species Species 0.000 description 2
- 238000003892 spreading Methods 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000012271 agricultural production Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 235000021393 food security Nutrition 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
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
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Mining & Mineral Resources (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a county remote-sensing scale-oriented crop planting area information extraction method and device. The method comprises the following steps: preparing an orthophoto map subjected to ortho-rectification; setting segmentation scale parameters according to terrain texture and shape features, and selecting the optimal segmentation scale for performing scale segmentation, so that the segmented plot has the same terrain object attribute; selecting a plot with a typical example as a sample object; establishing a subordinating degree function comprising a vegetation index feature function, a texture feature function and a shape feature function at least; performing rule classification review; extracting the crop planting area information; saving classification results. According to the method and device disclosed by the invention, the crop information extraction accuracy and efficiency can be improved.
Description
Technical field
Quantitative remote sensing image information of the present invention extracts field, and in particular to a kind of farming species towards county domain scale in remote sensing
Plant area information extracting method and device.
Background technology
Grain security is the significant challenge that the mankind face 21 century.Population in the world to the year two thousand fifty will break through 9,000,000,000, how really
The food supply of 9,000,000,000 populations is protected, agricultural will be faced with unprecedented pressure in 40 years in future.
County domain counts the statistic unit of most benchmark as State Statistics Bureau crops, accurately grasps the number of cereal crops
Amount, especially its locus distributed intelligence, is to be related to national food security, can be agricultural production, yield estimation, grain valency
Lattice prediction and national food production distribution and planning offer science important evidence.
The crop acreage information retrieval of county domain scale in remote sensing is many based on high spatial resolution satellite image at present
(GF, SPOT, QuickBird, WorldView etc.) is extracted, common classification algorithm include supervised classification (method of maximum likelihood,
Neutral net, support vector machine, decision tree), object oriented classification etc., most cases based on single when relatively species do not carry out
Discriminatory analysiss.Two big class sorting algorithm of relative analyses, based on the supervised classification algorithm of pixel, classification results easily produce " spiced salt "
Effect；And the object oriented classification based on plot yardstick can eliminate " spiced salt " effect to a great extent, but receive single image
The impact do not enriched of spectral information, it is difficult to while Different Crop class area be separated.Therefore, how fast and accurately to recognize
The crop acreage information for going out county domain scale in remote sensing becomes the difficult point of Classification in Remote Sensing Image technology.
Content of the invention
Technical problem is how to improve the precision that county domain scale in remote sensing crop acreage is extracted.
In view of this, the embodiment of the present invention provides a kind of crop acreage information retrieval towards county domain scale in remote sensing
Method and device is used for solving technical problem.
The solution of problem
The invention provides a kind of crop acreage information extracting method towards county domain scale in remote sensing, including step
Suddenly：
Prepare the orthophotoquad after processing through ortho-rectification；
According to atural object texture, shape facility, segmentation scale parameter is set, select optimal segmentation yardstick to carry out yardstick and divide
Cut so that after segmentation, plot is same terrain object attribute；
The plot with Typical Representative is selected as sample object；
Membership function is set up, at least includes vegetation index characteristic function, textural characteristics function, shape facility function；
Rule classification preview；
Crop acreage information retrieval；
Preserve classification results.
Further, when the orthophotoquad includes phasor and three images when phasor, the second image during the first image
Phasor；
Crop acreage information extracting method circulation for during the first image when phasor, the second image phasor and
During three images, phasor carries out crop acreage information retrieval, specifically includes following steps：
Phasor during the first image of input, based on object oriented classification algorithm, carries out farmland information extraction for the first time, ploughs with non-
Ground data separation comes；
Phasor during second the second image of input, extracts result in conjunction with farmland information, based on object oriented classification algorithm, base
In spectral signature, textural characteristics and shape facility, nonirrigated farmland information retrieval is carried out, arable land layer is divided into rice terrace and nonirrigated farmland；
Phasor during third time three image of input, in conjunction with nonirrigated farmland information retrieval result, based on phasor difference during three images
The spectral signature of crop and vegetation index difference, nonirrigated farmland information is at least refined as the first dry crop and the second dry crop.
I.e. with object oriented classification algorithm, from based on spectral signature, textural characteristics and shape facility, the spectral signature of Different Crop and plant
By aspects such as index differential, remote sensing image data is processed, more rapidly and accurately identify the farming of county domain scale in remote sensing
Species plant area information.
Further, merge all of extraction result, obtain accurate crops and type of ground objects information, classification results
For crops include Oryza sativa L., Semen Maydiss and Semen sojae atricolor, and type of ground objects includes water body, building, bare area, forest land and road.
Further, the step of optimal segmentation yardstick of the selection carries out multi-scale segmentation includes process：
Segmentation scale parameter setting：
The digital picture of one width M N array is divided into several folded regions of mutually disjointing；
With multi-scale division algorithm to generate the Image Segmentation region of height homogeneity (or heterogeneous minimum), so as to suitable
In optimal separation and expression ground object target；
The major parameter that segmentation is arranged at least includes out to out number, the yardstick interval of segmentation, out to out, color, puts down
Slippery, invalid value are arranged；
Segmentation effect is checked, if effect fails to reach default standard, the step of repeat size is split.Many chis
Degree segmentation is the patented technology extracted by imaged object；Image fusion segmentation is calculated using heterogeneous minimum region merging technique
Method；
Segmentation effect is checked, it is possible to reduce the deviation extension of error in subsequent step, specifically, if deposit herein
In error, then after setting up the step process such as membership function, will be enlarged by of error, cause then information retrieval result
There is relatively large deviation.
Further, described include process the step of set up membership function：
Luminance factor, the average factor, variance of unit weight are provided；
The area features factor, boundary index characterization factor, boundary length characterization factor, degree of compacting characterization factor, length are provided
Characterization factor, the quant's sign factor, the aspect ratio features factor and shape index characterization factor；
The neighbouring relations factor, the syntopy factor, the filiation factor are provided；
The texture factor is provided；Support assistance data；
Self-defined editor's formula forms characterization factor.Include for the characteristics of objects of object oriented classification：Spectral signature, shape
Shape feature, textural characteristics.Execute after multi-scale division generates imaged object homogeneity map, feature list and characteristics of objects value list
It is activated.Feature list is a very powerful instrument, for finding to distinguish the feature of different images object class；
The step of rule classification preview, includes process：Inspection-classification effect, if effect fails to reach default mark
Standard, then return the step of setting up membership function.Wherein, check segmentation result in, the image after segmentation is checked, be for
Ensure the heterogeneous suitable degree of homogeneity inside the polygon object of Image Segmentation generation and adjacent polygons object.Its
In, the validity check of rule classification preview can use sight check, it is also possible to using additive method；And carry out classifying quality inspection
Test, reach a standard and could continue next step, the precision of information retrieval can be improved.
Further, in the preservation classification results, result type includes two kinds of forms of vector grid.
Present invention also offers a kind of present invention arbitrary disclosed crop acreage information extracting method of employing
Towards the crop acreage information extracting device of county domain scale in remote sensing, including：
Module is just being penetrated, for preparing the orthophotoquad after ortho-rectification is processed；
Multi-scale segmentation module, for according to atural object texture, shape facility, setting segmentation scale parameter, selects optimal dividing
Cutting yardstick carries out multi-scale segmentation so that after segmentation, plot is same terrain object attribute；
Module chosen by sample, for selecting the plot with Typical Representative as sample object；
Membership function sets up module, for setting up membership function, at least includes vegetation index characteristic function, texture spy
Levy function, shape facility function；
Classification previewing module, for rule classification preview；
Information extraction modules, for crop acreage information retrieval；
Memory module, for preserving classification results.
Beneficial effects of the present invention：
Using technique scheme, the present invention/invention can at least obtain following technique effects：
The present invention is due to having carried out multi-scale segmentation according to atural object texture, shape facility etc., so it is able to ensure that same plot
Height homogeneity (or heterogeneous minimum), and be classified according to membership function, improve Objects recognition precision；Such as
This, the present invention possesses advantage of the object oriented classification algorithm in high score image is processed, and simultaneously as orthography tool
Possess the abundant spectral information advantage of multi_temporal images for multidate, greatly improve county domain scale in remote sensing farming species
Plant area information extraction accuracy.
Description of the drawings
For the technical scheme being illustrated more clearly that in the embodiment of the present invention, below will be to institute in embodiment of the present invention description
The accompanying drawing for using is needed to be briefly described, it should be apparent that, drawings in the following description are only some enforcements of the present invention
Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also be implemented according to the present invention
The content of example and these accompanying drawings obtain other accompanying drawings.
Fig. 1 is the flow chart of the crop acreage information extracting method towards county domain scale in remote sensing；
Fig. 2 is multidate terrain classification flow chart；Fig. 3 is and neighborhood contrast figure；
Fig. 4 is boundary index figure；
Fig. 5 is boundary length figure；
Fig. 6 is shape index figure；
Fig. 7 is gradation of image co-occurrence matrix figure；
Fig. 8 is the schematic diagram of the crop acreage information extracting device towards county domain scale in remote sensing.
Through accompanying drawing, it should be noted that similar label is used for describing same or analogous element, feature and structure.
Specific embodiment
The disclosure for describing to help comprehensive understanding to be limited by claim and its equivalent referring to the drawings is provided
Various embodiments.Hereinafter describe including the various details for understanding are helped, but these details will be considered to be only example
Property.Therefore, it will be appreciated by those of ordinary skill in the art that without departing from the scope of the present disclosure and spirit in the case of, can be right
Various embodiments described herein makes various changes and modifications.In addition, in order to clear and succinct, the retouching of known function and construction
State and can be omitted.
Term and vocabulary used in following description and claims is not limited to document implication, but only by inventor
It is used for so that the disclosure clearly and is as one man understood.Therefore, to those skilled in the art it should be apparent that carrying
Description for following various embodiments of this disclosure merely to exemplary purpose, and unrestricted by claims and its
The purpose of the disclosure that equivalent is limited.
It should be understood that unless in addition context clearly indicates, otherwise singulative also includes plural.Thus, for example,
Quoting including quoting to one or more such surfaces to " assembly surface ".
Fig. 1 is a kind of flow process of the crop acreage information extracting method towards county domain scale in remote sensing of the present embodiment
Figure.With reference to Fig. 1, a kind of schematic diagram of the crop acreage information extracting method towards county domain scale in remote sensing：The method, bag
Include step：
S1：Prepare the orthophotoquad after processing through ortho-rectification；
S2：According to atural object texture, shape facility, segmentation scale parameter is set, select optimal segmentation yardstick to carry out yardstick
Segmentation so that after segmentation, plot is same terrain object attribute；
S3：The plot with Typical Representative is selected as sample object；
S4：Membership function is set up, at least includes vegetation index characteristic function, textural characteristics function, shape facility function；
S5：Rule classification preview；
S6：Crop acreage information retrieval；
S7：Preserve classification results.
Using technique scheme, the present invention/invention can at least obtain following technique effects：
The present invention is due to having carried out multi-scale segmentation according to atural object texture, shape facility etc., so it is able to ensure that same plot
Height homogeneity (or heterogeneous minimum), and be classified according to membership function, improve Objects recognition precision；Such as
This, the present invention possesses advantage of the object oriented classification algorithm in high score image is processed, and simultaneously as orthography tool
Possess the abundant spectral information advantage of multi_temporal images for multidate, greatly improve county domain scale in remote sensing farming species
Plant area information extraction accuracy.
Wherein, phasor and phasor during three images when phasor, the second image when orthophotoquad can be divided into the first image.
Fig. 2 is multidate terrain classification flow chart, understands with reference to Fig. 2, and the present embodiment is preferred, and orthophotoquad includes the
Phasor and phasor during three images when phasor, the second image during one image；
Phasor and the 3rd when phasor, the second image when the circulation of crop acreage information extracting method is directed to the first image
During image, phasor carries out crop acreage information retrieval, specifically includes following steps：
Phasor during the first image of input, based on object oriented classification algorithm, carries out farmland information extraction for the first time, ploughs with non-
Ground data separation comes；
Phasor during second the second image of input, extracts result in conjunction with farmland information, based on object oriented classification algorithm, base
In spectral signature, textural characteristics and shape facility, nonirrigated farmland information retrieval is carried out, arable land layer is divided into rice terrace and nonirrigated farmland；
Phasor during third time three image of input, in conjunction with nonirrigated farmland information retrieval result, based on phasor difference during three images
The spectral signature of crop and vegetation index difference, nonirrigated farmland information is at least refined as the first dry crop and the second dry crop.
That is, repeat step S1-S7, with object oriented classification algorithm, from based on spectral signature, textural characteristics and shape facility, different works
The aspect such as the spectral signature of thing and vegetation index difference is processed to remote sensing image data, more rapidly and accurately identifies county
The crop acreage information of domain scale in remote sensing.
The present embodiment is preferred, merges all of extraction result, obtains accurate crops and type of ground objects information, classification
As a result it is that crops include Oryza sativa L., Semen Maydiss and Semen sojae atricolor, type of ground objects includes water body, building, bare area, forest land and road..Obtain
Accurate crops and type of ground objects information.Wherein, phase 1 is phasor during the first image, the like；Wherein, dry crop 1
That is the first dry crop, the like.
The present embodiment is preferred, selects optimal segmentation yardstick to include process the step of carrying out multi-scale segmentation：
Segmentation scale parameter setting：
The digital picture of one width M N array is divided into several folded regions of mutually disjointing；
With multi-scale division algorithm to generate the Image Segmentation region of height homogeneity (or heterogeneous minimum), so as to suitable
In optimal separation and expression ground object target；
The major parameter that segmentation is arranged at least includes out to out number, the yardstick interval of segmentation, out to out, color, puts down
Slippery, invalid value are arranged；
Segmentation effect is checked, if effect fails to reach default standard, the step of repeat size is split.Image
Cut zone is imaged object, and multi-scale division is the patented technology extracted by imaged object；Image fusion segmentation is adopted
With heterogeneous minimum region merging algorithm.Segmentation effect is checked, the deviation that can reduce error in subsequent step expands
Bigization, specifically, if where there is error, then after setting up the step process such as membership function, will be enlarged by of error,
Then information retrieval result is caused to there is relatively large deviation.
Multi-scale division is the patented technology extracted by imaged object.It can with different scale, high-quality carry
Take imaged object (thick and thin rank).This technology is suitable for the image with texture information, such as SAR, high-resolution satellite
Image or aeronautical data.It is suitable for extracting significant original data object from image data according to specific task.
Multi-scale division is to proceed by a region merging technique technology from bottom to top from the object of a pixel, and little imaged object can
To be merged in slightly larger object.
Image fusion segmentation is started from using heterogeneous minimum region merging algorithm, the merging of pixel in Image Segmentation
Single pixel is first merged into less imaged object, then less imaged object is merged into by any one pixel in image
Larger polygon object, in cutting procedure, polygon object is ever-increasing heterogeneous minimum.He be one from top to bottom, by
The process that level merges.
The basic thought of region merging method be will have like property gather composition area polygonal, first to per
Individual need segmentation region look for a seed pixel as growth starting point, then by seed pixel surrounding neighbors with seed picture
The pixel that there is same or similar property in unit is merged in the region at seed pixel place, by these new pixels as new seed
Pixel proceeds process above, and until not meeting the pixel of condition, such a region is generated as.In order to ensure shadow
As the homogeneity inside the polygon object that segmentation is generated and the heterogeneous suitable degree of adjacent polygons object, in region merging technique
Cutting procedure in need to consider two standards：The criterion and the condition for determining that stopping pixel merging for being similar to pixel merging is set,
The two conditions control the ownership of pixel in cutting procedure, and therefore the whether reasonable of standard setting directly affects impact after segmentation
The effectiveness of object.
The purpose of region merging algorithm is the weight heterogeneity minimum of imaged object after realization segmentation, only considers spectrum
After heterogeneous minimum can cause to split, the Polygonal Boundary of imaged object is relatively crushed, therefore, usually spectrum heterogeneity standard
Use cooperatively with special heterogeneity standard.It needs to be determined that impact heterogeneous and compactness heterogeneity before segmentation.Only guaranteed light
Spectrum heterogeneity, smoothness is heterogeneous, compactness is heterogeneous minimum, can just make the average heterogeneity of all objects of view picture image most
Little.
The heterogeneous f of any one imaged object be calculated by four variables obtained from：Wcolor (spectral information because
Son) wshape (the shape information factor) hcolor (spectrum heterogeneity) hshape (shape heterogeneity), and wcolor+wshape=
1.Among lower, w is the spectral information factor, and span is：0-1.
F=w hcolor+(1-w)hshape(0.1)；
Spectrum heterogeneity hcolor is not only relevant with the pixel number of composition object, additionally depends on the standard deviation of each wave band
(formula 0.2)
WhereinFor the standard deviation of pixel interior pel value, it is worth to according to the pixel for constituting object, n is pixel number.
Shape includes two sub- factors：Smoothness hsmoothWith compactness hcompct(formula 0.3).
hshape=wcmpct·hcmpct+(1-wcmpct)·hsmooth(0.3)；
hmoothAnd hcmpctDepending on the pixel number n of composition object, the minimum side length b of polygonal length of side l and same area
(formula 0.4).
Image fusion segmentation step is：
(1) partitioning parameters are set, including the weight of each wave band, i.e. importance of the single wave band in cutting procedure；One
Yardstick threshold value come determine pixel merge stop condition；Require to determine spectrum according to image texture feature and extracted thematic information
The weight of the factor and form factor；In form factor, the structure attribute according to most of atural object classifications determines the compactness factor
Weight with the smoothness factor.
(2) start segmentation centered on any one pixel in image, when splitting for the first time, single pixel is counted as one
Minimum polygon object participates in the calculating of heterogeneous value；After the completion of splitting for the first time, based on the polygon object for generating
Second segmentation is carried out, the same heterogeneity that calculates is worth, and judges the difference between the threshold value of f and reservation, if f is less than threshold value s, continues
The continuous segmentation work for carrying out multiple segmentation, then stopping image on the contrary, forms the imaged object layer of a fixed ruler angle value.
After multi-scale division, the elementary cell of image has been no longer single pixel, but by homogeneity pixel constitute polygon
Shape object, each polygon object not only includes spectral information, but also includes shape information, texture information, neighborhood information, right
For the similar ground class of spectral information, just easily can be extracted by the difference of other attributes of polygon object.Many
Multi-scale segmentation not only generates significant imaged object, and also by the imaged object Information expansion of former resolution to different chis
On degree, the multiple dimensioned description of image information is realized.The vision relaxation of people is analogous to, as yardstick is incrementally increased, to image
Carry out progressively comprehensive process.
The present embodiment is preferred, includes process the step of set up membership function：
Luminance factor, the average factor, variance of unit weight are provided；
The area features factor, boundary index characterization factor, boundary length characterization factor, degree of compacting characterization factor, length are provided
Characterization factor, the quant's sign factor, the aspect ratio features factor and shape index characterization factor；
The neighbouring relations factor, the syntopy factor, the filiation factor are provided；
The texture factor is provided；Support assistance data；
Self-defined editor's formula forms characterization factor；
The step of rule classification preview, includes process：Inspection-classification effect, if effect fails to reach default mark
Standard, then return the step of setting up membership function.Include for the characteristics of objects of object oriented classification：Spectral signature, shape spy
Levy, textural characteristics.After having executed multi-scale division generation imaged object homogeneity map, feature list and characteristics of objects value list are quilt
Activation.Feature list is a very powerful instrument, for finding to distinguish the feature of different images object class；And classified
Validity check, reaches a standard and could continue next step, can improve the precision of information retrieval.
Below the various features in feature list are carried out with a simple introduction：
(1) average
1) brightness
Calculate in selector channel, the average of subject pixels brightness value.
2) maximum interlayer difference
Equal value difference of the object in any two passage is calculated, and maximum difference is obtained, with maximum difference divided by brightness value be
For maximum interlayer difference.
3) Layer 0,1,2 ... ..., k；
Passage 0,1 ... ..., the object average of k.
(2) standard deviation
Passage 0,1 ... ..., the objective metric of k is poor.
(3) pixel
1) ratio
The ratio of passage reflects contribution degree of the passage to total luminance value.New " ratio " need to be created according to channel value.
2) minimum pixel value
New " minimum pixel value " need to be created according to channel value.
3) max pixel value
New " max pixel value " need to be created according to channel value.
4) with neighborhood contrast
Apart from d to the minimum enclosed rectangle border of external expansion existing object, to calculate in spreading range and reject point in object
The average that puts afterwards.New " with neighborhood contrast " need to be created according to channel value.As shown in Figure 3 with neighborhood contrast figure.
5) with neighborhood standard deviation
Apart from d to the minimum enclosed rectangle border of external expansion existing object, to calculate in spreading range and reject point in object
The standard variance that puts afterwards.New " with neighborhood standard deviation " need to be created according to channel value.
(4) with current layer relation
1) with the equal value difference of current layer
The difference of imaged object and current layer average.New " with the equal value difference of current layer " need to be created according to channel value.
2) with current layer ratio
The ratio of current layer average and imaged object.New " with current layer ratio " need to be created according to channel value.
Shape facility：
(1) commonly use
1) area
For the image without Geographic Reference, it is pixel number in object；There is the image of Geographic Reference, be that object is covered
True area.
2) boundary index
As boundary length and minimum enclosed rectangle girth ratioWherein, bv is the border of imaged object
Long, lv is the length of imaged object, and wv is the width of imaged object.Boundary index figure is as shown in Figure 4.
3) boundary length
The boundary length of one imaged object is defined as the total border length of its image adjacent with other or view picture image
Border long.In the image without Geographic Reference, the border length of a pixel is defined as 1.Boundary length figure is as shown in Figure 5.
4) degree of compacting
The degree of compacting of imaged object is the ratio of product of the length of object with width and total number of pixels.
5) density
6) long
It is defined as the square root of total number of pixels and the product of the length-width ratio of minimum enclosed rectangle.
7) length/width
Approximately tried to achieve by minimum external world's rectangle.
8) shape index
Shape index is defined as image boundary length with the subduplicate ratio of 4 times of imagery zoneShape index
Figure is as shown in Figure 6.
9) wide
The most important analysis method of textural characteristics is based on gray level co-occurrence matrixes (Gray Level Co-occurrence
Matrix,GLCM).GLCM is described in image, on θ direction distance for d be respectively provided with gray scale i for a pair and the pixel of j goes out
Existing probability.It is assumed that texture region to be studied is rectangle, which has Nx resolution in the horizontal direction, has Ny in vertical direction
Individual resolution, the gray level of image is N., Lx={ 0,1,2 ..., Nx-1 }, Ly={ 0,1,2 ..., Ny-1 } be respectively level and
Vertical spatial domain.Gray scale is θ for a pair of the pixel locality of i and j, and distance is designated as pi, j (d, θ) for the probability of d, specifically
Computing formula is：
It is comprehensive in terms of direction, adjacent spaces and amplitude of variation that the gray level co-occurrence matrixes of piece image reflect gradation of image
Conjunction information.Relation in image expressed by GLCM between pixel pair, gradation of image co-occurrence matrix figure is as shown in Figure 7.
Object oriented classification software is mainly taken seven kinds of features of GLCM to carry out the classification of textural characteristics, i.e.,：GLCM is same
Matter, GLCM contrast, GLCM diversity, GLCM entropy, GLCM angle (second moment), GLCM average, GLCM variance.
(1) GLCM homogeneity：
Homogeneity is the tolerance to image texture localized variation size.f1Between the zones of different of the bigger explanation image texture of value
Lack change, locally highly uniform.
(2) GLCM contrast：
The big pixel of texture medium contrast is many, then contrast is bigger.For open grain, Pi,jIt is attached compared with leading diagonal is concentrated on
Closely, the value of contrast is less.
(3) GLCM diversity：
Diversity describes the different degree in gray level co-occurrence matrixes between row or column element, is the degree of gray scale linear relationship
Amount.
(4) GLCM entropy：
The quantity of information of image is represented, is represented the complexity of texture, be the tolerance of picture material randomness.Texture-free entropy
It is zero, textured entropy maximum.
(5) GLCM angle (second moment)：
Angular second moment is the tolerance of gradation of image uniformity, works as Pi,jDistribution value is concentrated near leading diagonal, and local is described
The gradation of image distribution of neighborhood is uniform, and image assumes thicker texture, and the value is accordingly larger.
(6) GLCM average：
(7) GLCM standard deviation：
Variance reflects the cycle of texture.Value is bigger, shows that the cycle of texture is bigger.Standard deviation is the square root of variance.
The present embodiment is preferred, preserves in classification results, and result type includes two kinds of forms of vector grid.
The present embodiment is preferred, according to features such as atural object texture, shapes, sets segmentation scale parameter, selects optimal dividing
Cutting yardstick carries out multi-scale segmentation so that also include step after the step of plot is same terrain object attribute after segmentation：
Segmentation effect is checked, if effect fails default standard is reached, the step of repeat size is split, that is, is weighed
Multiple step S2 is until segmentation effect is up to standard；
The step of rule classification preview, includes process：Inspection-classification effect, if effect fails to reach default mark
Standard, then return the step of setting up membership function, that is, read step S4 and S5 again until classifying quality is up to standard.Wherein, segmentation is checked
As a result in, the image after segmentation is checked, the homogeneity being to ensure that inside the polygon object of Image Segmentation generation
Heterogeneous suitable degree with adjacent polygons object.Wherein, the validity check of rule classification preview can use sight check,
Additive method can also be used.
Embodiment two：
Fig. 8 be the present invention a kind of employ the arbitrary disclosed crop acreage information extracting method of the present invention towards
The schematic diagram of the crop acreage information extracting device of county domain scale in remote sensing, the device includes：
Module 10 is just being penetrated, for preparing the orthophotoquad after ortho-rectification is processed；
Multi-scale segmentation module 20, for according to atural object texture, shape facility, setting segmentation scale parameter, selects optimal
Segmentation yardstick carries out multi-scale segmentation so that after segmentation, plot is same terrain object attribute；
Module 30 chosen by sample, for selecting the plot with Typical Representative as sample object；
Membership function sets up module 40, for setting up membership function, at least includes vegetation index characteristic function, texture
Characteristic function, shape facility function；
Classification previewing module 50, for rule classification preview；
Information extraction modules 60, for crop acreage information retrieval；
Memory module 70, for preserving classification results.
Beneficial effects of the present invention：
Using technique scheme, the present invention/invention can at least obtain following technique effects：
The present invention is due to having carried out multi-scale segmentation according to atural object texture, shape facility etc., so it is able to ensure that same plot
Height homogeneity (or heterogeneous minimum), and be classified according to membership function, improve Objects recognition precision；Such as
This, the present invention possesses advantage of the object oriented classification algorithm in high score image is processed, and simultaneously as orthography tool
Possess the abundant spectral information advantage of multi_temporal images for multidate, greatly improve county domain scale in remote sensing farming species
Plant area information extraction accuracy.
It should be noted that the various embodiments of the disclosure as above are generally related to input data to a certain extent
Process and output data generation.This input data is processed and output data generation can be in hardware or soft with combination of hardware
Realize in part.For example, can in mobile device or similar or related circuit using specific electronic components for realize with
The function that the various embodiments of the disclosure are associated as mentioned above.Alternatively, according to that instructs to operate for being stored or more
Multiple processors can achieve the function of associating with the various embodiments of the disclosure as described above.If it is, then these instructions
Can be stored on one or more non-transitory processor readable mediums, this be in the scope of the present disclosure.Processor can
The example for reading medium includes read only memory (ROM), random access memory (RAM), CD-ROM, tape, floppy disk and optics number
According to storage device.In addition, can be by disclosure art for realizing functional computer program, instruction and the instruction segment of the disclosure
Programmer easily explain.
Various embodiments although with reference to the disclosure illustrate and describe the disclosure, but those skilled in the art will manage
Solution, in the case of without departing from the spirit and scope of the present disclosure being defined by the appended claims and the equivalents thereof, can enter to which
Various changes in row form and details.
Claims (7)
1. a kind of crop acreage information extracting method towards county domain scale in remote sensing, it is characterised in that including step：
Prepare the orthophotoquad after processing through ortho-rectification；
According to atural object texture, shape facility, segmentation scale parameter is set, select optimal segmentation yardstick that multi-scale segmentation is carried out, make
After must splitting, plot is same terrain object attribute；
The plot with Typical Representative is selected as sample object；
Membership function is set up, at least includes vegetation index characteristic function, textural characteristics function, shape facility function；
Rule classification preview；
Crop acreage information retrieval；
Preserve classification results.
2. crop acreage information extracting method as claimed in claim 1, it is characterised in that the orthophotoquad bag
Phasor and phasor during three images when phasor, the second image when including the first image；The crop acreage information retrieval side
Method circulation for during the first image when phasor, the second image phasor and during three images phasor carry out crop acreage information
Extract, specifically include following steps：
Phasor during the first image of input, based on object oriented classification algorithm, is carried out farmland information extraction, is believed with bare place for the first time
Breath makes a distinction；
Phasor during second the second image of input, extracts result in conjunction with farmland information, based on object oriented classification algorithm, based on light
Spectrum signature, textural characteristics and shape facility, carry out nonirrigated farmland information retrieval, and arable land layer is divided into rice terrace and nonirrigated farmland；
Phasor during third time three image of input, in conjunction with nonirrigated farmland information retrieval result, based on phasor Different Crop during three images
Spectral signature and vegetation index difference, nonirrigated farmland information is at least refined as the first dry crop and the second dry crop.
3. crop acreage information extracting method as claimed in claim 2, it is characterised in that merge and all of extract knot
Really, the classification results for obtaining at least include：Crops include：Oryza sativa L., Semen Maydiss and Semen sojae atricolor, the type of ground objects of bare place information includes
Water body, building, bare area, forest land and road.
4. crop acreage information extracting method as claimed in claim 1, it is characterised in that the selection optimal point
Cutting the step of yardstick carries out multi-scale segmentation includes process：
Segmentation scale parameter setting：
The digital picture of one width M N array is divided into several folded regions of mutually disjointing；
With multi-scale division algorithm to generate the Image Segmentation region of height homogeneity (or heterogeneous minimum), so as to be suitable to most
Good separation and expression ground object target；
Segmentation arrange major parameter at least include out to out number, segmentation yardstick interval, out to out, color, smoothness,
Invalid value is arranged；
Segmentation effect is checked, if effect fails to reach default standard, the step of repeat size is split.
5. crop acreage information extracting method as claimed in claim 1, it is characterised in that described set up degree of membership letter
The step of number includes process：
Luminance factor, the average factor, variance of unit weight are provided；
The area features factor, boundary index characterization factor, boundary length characterization factor, degree of compacting characterization factor, long feature are provided
The factor, the quant's sign factor, the aspect ratio features factor and shape index characterization factor；
The neighbouring relations factor, the syntopy factor, the filiation factor are provided；
The texture factor is provided；
Support assistance data；
Self-defined editor's formula forms characterization factor；
The step of rule classification preview, includes process：Inspection-classification effect, if effect fails to reach default standard,
The step of membership function is set up in return.
6. crop acreage information extracting method as claimed in claim 1, it is characterised in that the preservation classification results
In, result type includes two kinds of forms of vector grid.
7. a kind of employ as described in claim 1-6 is arbitrary crop acreage information extracting method towards the remote sensing of county domain
The crop acreage information extracting device of yardstick, it is characterised in that include：
Module is just being penetrated, for preparing the orthophotoquad after ortho-rectification is processed；
Multi-scale segmentation module, for according to atural object texture, shape facility, setting segmentation scale parameter, selects optimal segmentation chi
Degree carries out multi-scale segmentation so that after segmentation, plot is same terrain object attribute；
Module chosen by sample, for selecting the plot with Typical Representative as sample object；
Membership function sets up module, for setting up membership function, at least includes vegetation index characteristic function, textural characteristics letter
Number, shape facility function；
Classification previewing module, for rule classification preview；
Information extraction modules, for crop acreage information retrieval；
Memory module, for preserving classification results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611006469.1A CN106446875A (en) | 2016-11-16 | 2016-11-16 | County remote-sensing scale-oriented crop planting area information extraction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611006469.1A CN106446875A (en) | 2016-11-16 | 2016-11-16 | County remote-sensing scale-oriented crop planting area information extraction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106446875A true CN106446875A (en) | 2017-02-22 |
Family
ID=58208632
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611006469.1A Pending CN106446875A (en) | 2016-11-16 | 2016-11-16 | County remote-sensing scale-oriented crop planting area information extraction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106446875A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106971156A (en) * | 2017-03-22 | 2017-07-21 | 中国地质科学院矿产资源研究所 | Rare earth mining area remote sensing information extraction method based on object-oriented classification |
| CN108960089A (en) * | 2018-06-20 | 2018-12-07 | 中科禾信遥感科技（苏州）有限公司 | A kind of method and device of tea grove remote sensing monitoring |
| CN109033937A (en) * | 2018-06-01 | 2018-12-18 | 武汉禾大科技有限公司 | Number of plant statistical method and system are carried out by unmanned plane image |
| CN109101952A (en) * | 2018-09-04 | 2018-12-28 | 北京邮电大学 | Objects recognition method, apparatus, system and server based on mobile Internet |
| CN109635731A (en) * | 2018-12-12 | 2019-04-16 | 中国科学院深圳先进技术研究院 | It is a kind of to identify method and device, storage medium and the processor effectively ploughed |
| CN109800296A (en) * | 2019-01-21 | 2019-05-24 | 四川长虹电器股份有限公司 | A kind of meaning of one's words fuzzy recognition method based on user's true intention |
| CN111489092A (en) * | 2020-04-15 | 2020-08-04 | 云南户外图科技有限公司 | Method and system for evaluating suitable growing area of plant cultivation and planting environment |
| CN112308047A (en) * | 2020-12-02 | 2021-02-02 | 宁夏回族自治区自然资源勘测调查院 | Method and system for monitoring cultivated land quality |
| CN113160183A (en) * | 2021-04-26 | 2021-07-23 | 山东深蓝智谱数字科技有限公司 | Hyperspectral data processing method, device and medium |
| CN113283281A (en) * | 2021-02-26 | 2021-08-20 | 中科禾信遥感科技（苏州）有限公司 | Zizania latifolia planting area extraction method based on multi-temporal remote sensing image |
| CN113392759A (en) * | 2021-06-11 | 2021-09-14 | 河南大学 | Overwintering crop planting area identification method based on multi-source full-time-phase satellite image under cloud computing platform |
| CN113673339A (en) * | 2021-07-16 | 2021-11-19 | 北京农业信息技术研究中心 | Method and device for extracting standing straw based on unmanned aerial vehicle image |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140064554A1 (en) * | 2011-11-14 | 2014-03-06 | San Diego State University Research Foundation | Image station matching, preprocessing, spatial registration and change detection with multi-temporal remotely-sensed imagery |
| CN104866852A (en) * | 2015-04-01 | 2015-08-26 | 环境保护部卫星环境应用中心 | Method and apparatus for extracting land cover information in remote sensing image |
| CN105447494A (en) * | 2015-12-01 | 2016-03-30 | 二十一世纪空间技术应用股份有限公司 | Vegetable field monitoring method based on multi-source multi-temporal remote sensing image data |
| CN105608474A (en) * | 2015-12-30 | 2016-05-25 | 二十一世纪空间技术应用股份有限公司 | High-resolution-image-based regional adaptive cultivated land extraction method |
-
2016
- 2016-11-16 CN CN201611006469.1A patent/CN106446875A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140064554A1 (en) * | 2011-11-14 | 2014-03-06 | San Diego State University Research Foundation | Image station matching, preprocessing, spatial registration and change detection with multi-temporal remotely-sensed imagery |
| CN104866852A (en) * | 2015-04-01 | 2015-08-26 | 环境保护部卫星环境应用中心 | Method and apparatus for extracting land cover information in remote sensing image |
| CN105447494A (en) * | 2015-12-01 | 2016-03-30 | 二十一世纪空间技术应用股份有限公司 | Vegetable field monitoring method based on multi-source multi-temporal remote sensing image data |
| CN105608474A (en) * | 2015-12-30 | 2016-05-25 | 二十一世纪空间技术应用股份有限公司 | High-resolution-image-based regional adaptive cultivated land extraction method |
Non-Patent Citations (1)
| Title |
|---|
| 于静: "《遥感影像语义结构在城市规划中的应用》", 31 July 2009, 东北大学出版社 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106971156A (en) * | 2017-03-22 | 2017-07-21 | 中国地质科学院矿产资源研究所 | Rare earth mining area remote sensing information extraction method based on object-oriented classification |
| CN109033937A (en) * | 2018-06-01 | 2018-12-18 | 武汉禾大科技有限公司 | Number of plant statistical method and system are carried out by unmanned plane image |
| CN109033937B (en) * | 2018-06-01 | 2021-11-05 | 武汉禾大科技有限公司 | Method and system for counting plant number through unmanned aerial vehicle image |
| CN108960089A (en) * | 2018-06-20 | 2018-12-07 | 中科禾信遥感科技（苏州）有限公司 | A kind of method and device of tea grove remote sensing monitoring |
| CN109101952A (en) * | 2018-09-04 | 2018-12-28 | 北京邮电大学 | Objects recognition method, apparatus, system and server based on mobile Internet |
| CN109635731A (en) * | 2018-12-12 | 2019-04-16 | 中国科学院深圳先进技术研究院 | It is a kind of to identify method and device, storage medium and the processor effectively ploughed |
| CN109800296A (en) * | 2019-01-21 | 2019-05-24 | 四川长虹电器股份有限公司 | A kind of meaning of one's words fuzzy recognition method based on user's true intention |
| CN109800296B (en) * | 2019-01-21 | 2022-03-01 | 四川长虹电器股份有限公司 | Semantic fuzzy recognition method based on user real intention |
| CN111489092A (en) * | 2020-04-15 | 2020-08-04 | 云南户外图科技有限公司 | Method and system for evaluating suitable growing area of plant cultivation and planting environment |
| CN111489092B (en) * | 2020-04-15 | 2023-01-31 | 云南户外图科技有限公司 | Method and system for evaluating suitable growing area of plant cultivation and planting environment |
| CN112308047A (en) * | 2020-12-02 | 2021-02-02 | 宁夏回族自治区自然资源勘测调查院 | Method and system for monitoring cultivated land quality |
| CN112308047B (en) * | 2020-12-02 | 2022-06-17 | 宁夏回族自治区自然资源勘测调查院 | Method and system for monitoring cultivated land quality |
| CN113283281A (en) * | 2021-02-26 | 2021-08-20 | 中科禾信遥感科技（苏州）有限公司 | Zizania latifolia planting area extraction method based on multi-temporal remote sensing image |
| CN113160183A (en) * | 2021-04-26 | 2021-07-23 | 山东深蓝智谱数字科技有限公司 | Hyperspectral data processing method, device and medium |
| CN113392759A (en) * | 2021-06-11 | 2021-09-14 | 河南大学 | Overwintering crop planting area identification method based on multi-source full-time-phase satellite image under cloud computing platform |
| CN113673339A (en) * | 2021-07-16 | 2021-11-19 | 北京农业信息技术研究中心 | Method and device for extracting standing straw based on unmanned aerial vehicle image |
| CN113673339B (en) * | 2021-07-16 | 2024-02-23 | 北京农业信息技术研究中心 | Unmanned aerial vehicle image-based on-site straw extraction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106446875A (en) | County remote-sensing scale-oriented crop planting area information extraction method and device | |
| Neupane et al. | Deep learning based banana plant detection and counting using high-resolution red-green-blue (RGB) images collected from unmanned aerial vehicle (UAV) | |
| US11521380B2 (en) | Shadow and cloud masking for remote sensing images in agriculture applications using a multilayer perceptron | |
| US9202259B2 (en) | Automated geospatial image mosaic generation with automatic source selection | |
| CN106384081B (en) | Slope farmland extraction method and system based on high-resolution remote sensing image | |
| Packalen et al. | Predicting the spatial pattern of trees by airborne laser scanning | |
| CN112381013B (en) | Urban vegetation inversion method and system based on high-resolution remote sensing image | |
| US20160321501A1 (en) | Automated geospatial image mosaic generation with multiple zoom level support | |
| Samat et al. | Classification of VHR multispectral images using extratrees and maximally stable extremal region-guided morphological profile | |
| US11574464B2 (en) | Practical method for landslide detection in large space | |
| Karami et al. | Gully erosion mapping using object-based and pixel-based image classification methods | |
| CN107330875A (en) | Based on the forward and reverse heterogeneous water body surrounding enviroment change detecting method of remote sensing images | |
| Rodriguez-Galiano et al. | Incorporating the downscaled Landsat TM thermal band in land-cover classification using random forest | |
| Ge et al. | Deep residual network-based fusion framework for hyperspectral and LiDAR data | |
| CN114563378B (en) | Method, device, medium and equipment for quantitatively describing space distribution of cyanobacterial bloom in lakes and reservoirs | |
| Zaabar et al. | Application of convolutional neural networks with object-based image analysis for land cover and land use mapping in coastal areas: A case study in Ain Témouchent, Algeria | |
| CN115965812B (en) | Evaluation method for classification of unmanned aerial vehicle images on wetland vegetation species and land features | |
| Uribeetxebarria et al. | Assessing ranked set sampling and ancillary data to improve fruit load estimates in peach orchards | |
| Hong et al. | Multitemporal greenhouse mapping for high-resolution remote sensing imagery based on an improved YOLOX | |
| Batista et al. | On the classification of remote sensing high spatial resolution image data | |
| CN110427961B (en) | Building information extraction method and system based on rule and sample fusion | |
| CN116994029A (en) | Fusion classification method and system for multi-source data | |
| CN115049920A (en) | GEE combined growth period attribute and object-oriented large-area crop extraction method | |
| Priego et al. | ECAS-II: A hybrid algorithm for the construction of multidimensional image segmenters | |
| Fan et al. | Classification of hyperspectral image by preprocessing method based relation network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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: 20170222 |