CN105005782A - Fine method for global vegetation classification based on multi-temporal remote sensing data and spectroscopic data - Google Patents
Fine method for global vegetation classification based on multi-temporal remote sensing data and spectroscopic data Download PDFInfo
- Publication number
- CN105005782A CN105005782A CN201410166106.9A CN201410166106A CN105005782A CN 105005782 A CN105005782 A CN 105005782A CN 201410166106 A CN201410166106 A CN 201410166106A CN 105005782 A CN105005782 A CN 105005782A
- Authority
- CN
- China
- Prior art keywords
- data
- classification
- land
- vegetation
- land classification
- 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
Landscapes
- Image Processing (AREA)
Abstract
Provided is a fine method for global vegetation classification. The method comprises: firstly, obtaining spectroscopic data and remote sensing data of vegetation with a first main component; secondly, obtaining an area where land classification changes through the remarkable statistical test of average values; thirdly, performing minimum range monitoring classification of the area where land classification changes by taking an area where land classification doesn't change as a monitoring classification sample; fourthly, performing minimum range monitoring classification of a vegetation part of obtained data of new land classification data by utilizing spectroscopic data to obtain fine vegetation classification data; and finally, combining the obtained fine vegetation classification data with the non-vegetation data to obtain a fine vegetation classification result map. The method is applicable to departments of territorial resources, agriculture, forestry, and the like for fast updating a huge amount of land and vegetation classification data, and for fast accurately monitoring condition of vegetation growth and situation of change of land use.
Description
Technical field
Based on the meticulous vegetative breakdown method in the whole world of multi-temporal remote sensing data and spectroscopic data, belong to digital image processing field, particularly vegetation cover type change detection techniques and digital image processing techniques.
Background technology
Vegetation cover type refers to due to the natural quality in soil or the vegetation cover situation that formed by the effect of human activity, as forest, grassland, crop, bare area etc.Meticulous vegetative breakdown is the description to vegetation cover type refinement more, as type such as forest cover such as grade is subdivided into the type such as deciduous broad-leaved forest, evergreen coniferous forest, and agrotype is subdivided into the types such as paddy rice, wheat, corn.Meticulous vegetative breakdown is an important content of land classification research, and vegetation pattern change affects other change of properties of earth system, to a great extent as Productivity of Ecological System, bio-diversity, biogeochemical cycle etc.
Remote sensing has the advantage that can provide planar, synchronous earth's surface information, remote sensing technology develop into global meticulous vegetative breakdown and vegetation pattern Changeement provides new tool, people are enable to pass through science and effective means more, obtain the meticulous vegetation pattern information of global range and situation of change thereof, for solve along with population sharply expand and environmental pollution that urbanization process is accelerated to bring, land deterioration, vegetation deterioration, species diversity disappearance and resource plaque be weary etc. that a series of great global environmental problem provides powerful power-assisted.
In the past few decades, experts and scholars both domestic and external are devoted to study meticulous vegetative breakdown technology and method always and improve the precision of by remote sensing image, vegetation being carried out to sophisticated category.The various non-supervisory and supervised classification technology of initial development, existing a lot of advanced sorting algorithm is widely used at present, comprises minimum distance classification, maximum likelihood classification, neuroid etc.
Minimum distance classification is the conventional supervised classification method based on image statistics, principle of classification is the class each pixel dot-dash being grouped into the class center place apart from its minimum distance, its nicety of grading depends on the precision of adding up understanding and the training of known atural object classification, also relevant with training sample quantity.
Maximum likelihood classification, also referred to as Bayes (Bayes) classification, is the conventional supervised classification based on image statistics equally.This sorting technique is based on bayesian criterion, assuming that training sample data are in the distribution Normal Distribution of spectral space, calculate the probability density isoline of sample, and set up discriminant classification function, by each pixel of point by point scanning, the proper vector of pixel is substituted into discriminant classification function, obtains it and belong to all kinds of probability, and pixel is included into the classification of maximum probability.
Neuroid classification is a kind of sorting technique with artificial intelligence grown up in recent years.Mainly comprise the neural network classification methods such as BP neural network, radial base neural net, fuzzy neural network, wavelet neural network.Wherein BP neural network model (feed-forward network model) is neural network model most widely used at present.It is made up of input layer, hidden layer, output layer three part, and learning process is made up of forward-propagating process and back-propagation process.BP neural network model can be converted into a nonlinear optimal problem the I/O problem of one group of sample, usually can obtain classifying quality more better than general statistical method, but also there is the shortcomings such as pace of learning slowly, is not easily restrained, efficiency is not high.
In addition the sorting algorithms such as support vector machine (SVM), decision tree are also had.From researching and analysing above, these algorithms do not need the hypothesis following data fit normal distribution, are therefore more suitable for single or polynary remotely-sensed data to incorporate in vegetative breakdown algorithm.But, because these sorting algorithms need enough training samples and machine learning time, specific classification problem is also needed to the different nonparametric classification algorithm determined, therefore in the production of meticulous vegetative breakdown product, the sorting parameter being difficult to adopt in the world same to be applicable to and sorting algorithm, its algorithm universality is poor, also the Time and place resolution accomplishing simultaneously to keep meticulous vegetative breakdown product is difficult to, the meticulous vegetative breakdown technology in the current whole world is caused not meet whole world change, grain is assessed, carbon cycle, the demand of the applications such as land cover pattern variation monitoring, therefore rapid automatizedly produce the meticulous vegetative breakdown product of global arbitrary region as required, improve its temporal resolution, there is its irreplaceable significance.
Summary of the invention
(1) technical matters that will solve
The invention provides a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data, remote sensing image is only relied on to cause global classifying different parts algorithm universality poor as data source in order to solve in existing vegetative breakdown technology, and existing vegetative breakdown product temporal resolution is low, the problem of the meticulous vegetative breakdown product of global arbitrary region can not be obtained fast as required.
(2) technical scheme
Based on the meticulous vegetative breakdown method in the whole world of multi-temporal remote sensing data and spectroscopic data, it is characterized in that comprising following concrete steps:
Step one, remotely-sensed data and typical vegetation spectroscopic data obtain.
According to the difference of sophisticated category survey region, by remote sensing data sharing platform, airborne remote sensing platform real data, obtain the multi-temporal remote sensing data message in this region; By land classification data sharing platform, land classification enquiry data etc., obtain the history land classification data message in this region; By spectroscopic data shared platform and actual measurement vegetation spectroscopic data, obtain the spectroscopic data information of the typical vegetation sample in this region.
Step 2, remotely-sensed data and spectroscopic data pre-service.
According to the difference of sophisticated category survey region, by being numbered of remote sensing history land classification data, carry out the process of land classification code assignment for each land type, obtain the history land classification code figure in this region; The time series remotely-sensed data of the new period is merged with the form of principal component transform, obtains the first principal component remotely-sensed data after merging.To some spectroscopic datas of each typical vegetation sample in this region, merge with the form of principal component transform, obtain the first principal component spectroscopic data after merging.
Step 3, feature information extraction.
First principal component remotely-sensed data is superposed with history land classification code figure, generate a land classification data structure body, in order to record pixel numbering, history land classification code value, remotely-sensed data value, and the average of the remotely-sensed data value of each land classification and standard deviation, this structure comprises the characteristic information of remotely-sensed data and land classification data two kinds of data sources.Generate a spectroscopic data structure, in order to spectra re-recorded numbering, vegetative breakdown code value, longitude, latitude, wavelength value, the corresponding reflectivity of wavelength, this structure comprises the characteristic information of first principal component spectroscopic data.
Step 4, mean value conspicuousness statistical test change land type.
Utilize the land classification data structure body obtained in step 3, according to history land classification code value, the remotely-sensed data value of each land classification is averaged the statistical test of value conspicuousness, detect the pixel of remotely-sensed data value generation marked change in each land classification, generate a change pixel array, whether record the change of the remotely-sensed data value of each pixel numbering.
Step 5, land type change information is utilized to carry out land classification.
The change pixel array utilizing step 4 to obtain, using the remotely-sensed data value of part unchanged in each land classification as supervised classification sample, minimum distance method supervised classification is carried out to the remotely-sensed data value of the part changed, the classification results obtained by region of variation merges with the history land classification data of non-region of variation, obtains new land classification data.
Step 6, spectroscopic data information is utilized to carry out meticulous vegetative breakdown.
The first principal component spectroscopic data utilizing step 3 to obtain is as supervised classification sample, in new land classification data step 4 obtained, land type is that the part of vegetation carries out minor increment supervised classification and post-classification comparison, obtains new vegetation sophisticated category figure.
(3) beneficial effect
The present invention adopts multi-temporal remote sensing data, avoids the erroneous judgement situation of the land classification adopting single phase remotely-sensed data to bring because of cloud, shade and other noise informations, at utmost ensure that the reliability of grouped data; Adopt history land classification data to carry out as basis of classification data, make full use of existing resource, avoid a large amount of repeated manpower, financial resources and time waste; Adopt the change detecting method of mean value conspicuousness statistical test change, the region do not changed to land classification, without the need to reclassifying, improves efficiency and the accuracy of algorithm; Adopt typical vegetation spectroscopic data as the sample of vegetation sophisticated category, avoid the criteria for classification caused because of global zones of different vegetation pattern difference and upgrowth situation difference not to be suitable in the world, improve algorithm universality.Algorithm can be applicable to the departments such as land resources and agricultural, forestry and upgrades magnanimity soil and vegetative breakdown data fast, quick and precisely monitor vegetation growth status and land use change survey situation, for the research such as interphase interaction of the climate change in the whole world, bio-diversity, evolution of ecological environment and people and environment provides true and analyzes Data Source widely, there is wide market outlook and using value.
Accompanying drawing explanation
Fig. 1: based on the meticulous vegetative breakdown method flow diagram in the whole world of multi-temporal remote sensing data and spectroscopic data;
Fig. 2: mean value conspicuousness statistical test change land type process flow diagram;
Fig. 3: invention test findings Local map.Wherein Fig. 3 a is the central arid belt in Ningxia first quarter in 2011 vegetation sophisticated category figure, Fig. 3 b is the central arid belt in Ningxia vegetation sophisticated category figure second quarter in 2011, Fig. 3 c is the central arid belt in Ningxia vegetation sophisticated category figure third season in 2011, Fig. 3 d is central arid belt in Ningxia vegetation sophisticated category figure fourth quarter in 2011.
Embodiment
In order to understand technical scheme of the present invention better, introduce the present invention in detail below in conjunction with the drawings and the specific embodiments.
The present invention is a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data, and the method mainly comprises following step:
1. remotely-sensed data and typical vegetation spectroscopic data obtain;
2. remotely-sensed data and spectroscopic data pre-service;
3. feature information extraction;
4. mean value conspicuousness statistical test change land type;
5. utilize land type change information to carry out land classification;
6. utilize spectroscopic data information to carry out meticulous vegetative breakdown.
As shown in Figure 1, the concrete implementation detail of each several part is as follows for specific implementation flow process of the present invention:
1. remotely-sensed data and typical vegetation spectroscopic data obtain
The acquisition of remotely-sensed data and exemplary spectrum data, is divided into following step to realize:
(1) according to the difference of sophisticated category survey region, by land classification data sharing platform, land classification enquiry data etc., the history land classification data message in this region is obtained;
(2) by remote sensing data sharing platform, airborne remote sensing platform real data, obtain the multi-temporal remote sensing data message in this region, remotely-sensed data need meet multidate feature, acquisition time is when coming after land classification data message, and the time interval is consistent, as synthesis reflectivity product (MOD09A1) in 8 days of MODIS500m resolution, MODIS500m resolution 16 days vegetation index sinteticses (MOD13A1), Landsat8 land imager 30m resolution remote sense images (revisiting period 16 days) etc.;
(3) by spectroscopic data shared platform and actual measurement vegetation spectroscopic data, obtain the spectroscopic data information of the typical vegetation sample in this region, this spectroscopic data spectral resolution need be not less than the spectral resolution of remotely-sensed data, spectroscopic data acquisition time is consistent with remotely-sensed data acquisition time, and can reflect this region typical vegetation upgrowth situation.
2. remotely-sensed data and spectroscopic data pre-service
For the remotely-sensed data of acquisition in step 1 and the pre-service of typical vegetation spectroscopic data, following step is divided into realize:
(1) the history land classification Data Source owing to adopting differs, not unified standard, first need to carry out pre-service numbering chemical industry to do, the process of land classification code assignment is carried out for each land type, obtain the history land classification code figure in this region, make this code consistent with IGBP (IGBP) global vegetative breakdown scheme: water (0), evergreen coniferous forest (1), evergreen broadleaf forest (2), fallen leaves coniferous forest (3), deciduous broad-leaved forest (4), mixed forest (5), closing shrubbery (6), open shrubbery (7), the grassland (8) of many trees, savanna (9), grassland (10), permanent wetland (11), crop (12), city and built-up areas (13), the compact land (14) of crop and natural vegetation, ice and snow (15) and bare area or low vegetative coverage ground (16),
(2) the time series remotely-sensed data of the new period is merged with the form of principal component transform, obtain the first principal component remotely-sensed data that can reflect maximum quantity of information after merging;
(3) to some spectroscopic datas of each typical vegetation sample in this region, merge with the form of principal component transform, obtain the first principal component spectroscopic data that can reflect maximum quantity of information after merging.
3. feature information extraction
For the history land classification code figure obtained in step 2, first principal component remotely-sensed data and first principal component spectroscopic data, generate land classification data structure body and spectroscopic data structure respectively, be divided into following step to realize:
(1) land classification data structure body is built, in order to store: pixel is numbered, history land classification code value, remotely-sensed data value, and the average of the remotely-sensed data value of each land classification and standard deviation, this structure comprises the characteristic information of remotely-sensed data and land classification data two kinds of data sources;
(2) spectroscopic data structure is built, in order to store: spectra number, vegetative breakdown code value, longitude, latitude, wavelength value, the corresponding reflectivity of wavelength, this structure comprises the characteristic information of first principal component spectroscopic data;
(3) to first principal component remotely-sensed data and history land classification code figure, read each pixel successively, in land classification data structure body, record pixel numbering, history land classification code value, remotely-sensed data value;
(4) the first principal component spectroscopic data of each typical vegetation sample is read successively, spectra re-recorded numbering in spectroscopic data structure, vegetative breakdown code value, longitude, latitude, wavelength value and the corresponding reflectivity of this wavelength.
4. mean value conspicuousness statistical test change land type
For the land classification data structure body built in step 3, adopt mean value conspicuousness statistical test change land type, specific implementation flow process as shown in Figure 2, is divided into following step to realize:
(1) choose the land classification data structure body of a certain land type, add up this land classification data mean value μ and this land classification data standard difference σ;
(2) whether build change pixel array Data_Change, change in order to the land classification marked corresponding to each pixel, initial value is 0, and namely land classification does not change;
(3) i=0 is made;
(4) for the data value Xi of i-th pixel of this land classification, whether calculate Xi is in outside the σ scope of μ ± 2, namely | Xi-μ | whether >2 σ sets up, if set up, then the numerical value of Data_Change [i] becomes 1 from 0, namely represent that land classification changes, if be false, then Data_Change [i] remains unchanged;
(5) if Xi is not last pixel, then i=i+1, return step (4), otherwise enter next step;
(6) according to change pixel array Data_Change, the data that the data be divided into land classification to change land classification data structure body and land classification do not change.
5. utilize land type change information to carry out land classification
Utilize the data that the land classification obtained in step 4 does not change, and the spectroscopic data structure obtained in step 3, meticulous vegetative breakdown is carried out to the data that the land classification obtained in step 4 changes, is divided into following step to realize:
(1) in step 4 data of the unchanged part of each land classification as supervised classification sample, minimum distance method supervised classification is carried out to the data of the part that land classification in step 4 changes, the classification results obtained by region of variation merges with the history land classification data of non-region of variation, obtains new land classification data;
(2) by step (1) originally according to the new land classification data of IGBP whole world vegetative breakdown scheme according to the fresh code system recompile in global vegetation sophisticated category system, original code be 12 crop first remain unchanged, obtain new land classification data, whole world vegetation sophisticated category system code is as follows: paddy rice (11), wheat (12), corn (13), the compact land (14) of crop and natural vegetation, other Crop Group (16), deciduous broad-leaved forest (21), evergreen broadleaf forest (22), fallen leaves coniferous forest (23), evergreen coniferous forest (24), mixed forest (25), grassland (31), the grassland (32) of many trees, savanna (33), open shrubbery (41), closing shrubbery (42), permanent wetland (51), water body (61), city and built-up areas (71), bare area and low vegetative coverage ground (81), ice and snow (91),
6. utilize spectroscopic data information to carry out meticulous vegetative breakdown
(1) utilize first principal component spectroscopic data in the spectroscopic data structure obtained in step 3 as supervised classification sample, the crop part being 12 by land type code in the new land classification data obtained in step 5 carries out minor increment supervised classification, agrotype is subdivided into paddy rice (11), wheat (12), corn (13), by the method arranging threshold value, the pixel being greater than setting threshold value is divided into other Crop Group (16);
(2) post-classification comparison, completes image and exports.
The present invention can be applicable to the departments such as land resources and agricultural, forestry and upgrades magnanimity soil and vegetative breakdown data fast, quick and precisely monitor vegetation growth status and land use change survey situation, for the research such as interphase interaction of the climate change in the whole world, bio-diversity, evolution of ecological environment and people and environment provides true and analyzes Data Source widely, there is wide market outlook and using value.
Claims (6)
1., based on the meticulous vegetative breakdown method in the whole world of multi-temporal remote sensing data and spectroscopic data, it is characterized in that comprising the following steps:
(1) the spectroscopic data information of the multi-temporal remote sensing data message of sophisticated category survey region, history land classification data message and typical vegetation sample is obtained;
(2) being numbered of remote sensing history land classification data will obtained in step (1), obtain the history land classification code figure in this region, the time series remotely-sensed data of the new period obtained in step (1) and the spectroscopic data of typical vegetation sample are carried out principal component transform, obtains first principal component remotely-sensed data and first principal component spectroscopic data;
(3) the first principal component remotely-sensed data obtained in step (2) is superposed with history land classification code figure, generate a land classification data structure body, the first principal component spectroscopic data obtained in step (2) is generated a spectroscopic data structure;
(4) to the land classification data structure body obtained in step (3), the statistical test of the value that is averaged conspicuousness, detect the pixel of remotely-sensed data value generation marked change in each land classification, generate a change pixel array, whether record the change of the remotely-sensed data value of each pixel numbering;
(5) the change pixel array in step (3) is utilized, using the remotely-sensed data value of part unchanged in each land classification as supervised classification sample, minimum distance method supervised classification is carried out to the remotely-sensed data value of the part changed, the classification results obtained by region of variation merges with the history land classification data of non-region of variation, obtains the land classification data of new coding;
(6) utilize first principal component spectroscopic data that step (3) obtains as supervised classification sample, the part being vegetation by land type in land classification data new in step (5) carries out minor increment supervised classification and post-classification comparison, obtains new vegetation sophisticated category figure.
2. a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data according to claim 1, it is characterized in that: in described step (3), generate land classification data structure body, refer to record pixel numbering, history land classification code value, remotely-sensed data value, and the average of the remotely-sensed data value of each land classification and the data structure body of standard deviation, this structure comprises the characteristic information of remotely-sensed data and land classification data two kinds of data sources.
3. a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data according to claim 1, it is characterized in that: in described step (3), generate spectroscopic data structure, refer to and to number in order to spectra re-recorded, vegetative breakdown code value, longitude, latitude, wavelength value, the data structure body of the corresponding reflectivity of wavelength, this structure comprises the characteristic information of first principal component spectroscopic data.
4. a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data according to claim 1, it is characterized in that: the mean value conspicuousness statistical test in described step (4), its concrete steps are as follows:
A () chooses the land classification data structure body of a certain land type, add up this land classification data mean value μ and this land classification data standard difference σ;
B () builds change pixel array Data_Change, whether change in order to the land classification marked corresponding to each pixel, initial value is 0, and namely land classification does not change;
C () makes i=0;
D () is for the data value Xi of i-th pixel of this land classification, whether calculate Xi is in outside the σ scope of μ ± 2, namely | Xi-μ | whether >2 σ sets up, if set up, then the numerical value of Data_Change [i] becomes 1 from 0, namely represent that land classification changes, if be false, then Data_Change [i] remains unchanged;
If e () Xi is not last pixel, then i=i+1, returns step (4), otherwise enter next step;
F () is according to changing pixel array Data_Change, the data that the data be divided into land classification to change land classification data structure body and land classification do not change.
5. a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data according to claim 1, it is characterized in that: the land classification data of the new coding in described step (5), refer to the history land classification data of classification results and the non-region of variation that region of variation is obtained merge after land classification data according to the fresh code system recompile in global vegetation sophisticated category system, original code be 12 crop first remain unchanged, the land classification data encoding obtaining new coding is as follows:
Paddy rice (11), wheat (12), corn (13), the compact land (14) of crop and natural vegetation, other Crop Group (16), deciduous broad-leaved forest (21), evergreen broadleaf forest (22), fallen leaves coniferous forest (23), evergreen coniferous forest (24), mixed forest (25), grassland (31), the grassland (32) of many trees, savanna (33), open shrubbery (41), closing shrubbery (42), permanent wetland (51), water body (61), city and built-up areas (71), bare area and low vegetative coverage ground (81), ice and snow (91).
6. a kind of meticulous vegetative breakdown method in the whole world based on multi-temporal remote sensing data and spectroscopic data according to claim 1, it is characterized in that: in described step (6), minor increment supervised classification is carried out to the part that land type in new land classification data is vegetation, refer to that first principal component spectroscopic data in the spectroscopic data structure utilizing and obtain in step (3) is as supervised classification sample, the crop part being 12 by land type code in the new land classification data obtained in step (5) carries out minor increment supervised classification, agrotype is subdivided into paddy rice (11), wheat (12), corn (13), by the method arranging threshold value, the pixel being greater than setting threshold value is divided into other Crop Group (16).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410166106.9A CN105005782A (en) | 2014-04-24 | 2014-04-24 | Fine method for global vegetation classification based on multi-temporal remote sensing data and spectroscopic data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410166106.9A CN105005782A (en) | 2014-04-24 | 2014-04-24 | Fine method for global vegetation classification based on multi-temporal remote sensing data and spectroscopic data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105005782A true CN105005782A (en) | 2015-10-28 |
Family
ID=54378447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410166106.9A Pending CN105005782A (en) | 2014-04-24 | 2014-04-24 | Fine method for global vegetation classification based on multi-temporal remote sensing data and spectroscopic data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105005782A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224938A (en) * | 2015-11-06 | 2016-01-06 | 电子科技大学 | The modification method of remote sensing images land use classes result |
| CN105740759A (en) * | 2016-01-15 | 2016-07-06 | 武汉珈和科技有限公司 | Middle-season rice information decision tree classification method based on multi-temporal data feature extraction |
| CN106339423A (en) * | 2016-08-14 | 2017-01-18 | 李宇翔 | Method and device for dynamically updating sugarcane planting information |
| CN106355143A (en) * | 2016-08-25 | 2017-01-25 | 中国农业大学 | Seed maize field identification method and system based on multi-source and multi-temporal high resolution remote sensing data |
| CN106373150A (en) * | 2016-08-15 | 2017-02-01 | 中国农业大学 | Identification method of corns for seed on the basis of high resolution remote sensing data texture analysis |
| CN106682624A (en) * | 2016-12-29 | 2017-05-17 | 中国科学院深圳先进技术研究院 | Built-up area extraction method and apparatus based on time sequence remote sensing information |
| CN107480634A (en) * | 2017-08-12 | 2017-12-15 | 天津市测绘院 | A kind of geographical national conditions ground mulching monitoring method based on multistage target classification |
| CN108241905A (en) * | 2016-12-23 | 2018-07-03 | 赫克斯冈技术中心 | For predicting the method for soil and/or plant situation |
| CN108399400A (en) * | 2018-03-23 | 2018-08-14 | 中国农业科学院农业资源与农业区划研究所 | A kind of early stage crop recognition methods and system based on high-definition remote sensing data |
| CN108647568A (en) * | 2018-03-30 | 2018-10-12 | 电子科技大学 | Grassland degeneration extraction method based on full convolutional neural networks |
| CN111630551A (en) * | 2018-01-23 | 2020-09-04 | X开发有限责任公司 | Crop type classification in images |
| CN112258793A (en) * | 2020-10-19 | 2021-01-22 | 深圳前海微众银行股份有限公司 | Disaster monitoring and early warning method, device, equipment and medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510374A (en) * | 2009-03-12 | 2009-08-19 | 中国科学院遥感应用研究所 | Method for automatically updating land-use map based on historical data and remote sensing data |
| CN103679202A (en) * | 2013-12-17 | 2014-03-26 | 中国测绘科学研究院 | Method and device suitable for classifying vegetation through optical remote sensing satellite images |
-
2014
- 2014-04-24 CN CN201410166106.9A patent/CN105005782A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510374A (en) * | 2009-03-12 | 2009-08-19 | 中国科学院遥感应用研究所 | Method for automatically updating land-use map based on historical data and remote sensing data |
| CN103679202A (en) * | 2013-12-17 | 2014-03-26 | 中国测绘科学研究院 | Method and device suitable for classifying vegetation through optical remote sensing satellite images |
Non-Patent Citations (2)
| Title |
|---|
| 吴见等: "基于空间信息的高光谱遥感植被分类技术", 《农业工程学报》 * |
| 竞霞等: "基于分区和多时相遥感数据的山区植被分类研究", 《遥感技术与应用》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105224938B (en) * | 2015-11-06 | 2018-11-09 | 电子科技大学 | The modification method of remote sensing images land use classes result |
| CN105224938A (en) * | 2015-11-06 | 2016-01-06 | 电子科技大学 | The modification method of remote sensing images land use classes result |
| CN105740759A (en) * | 2016-01-15 | 2016-07-06 | 武汉珈和科技有限公司 | Middle-season rice information decision tree classification method based on multi-temporal data feature extraction |
| CN105740759B (en) * | 2016-01-15 | 2019-01-15 | 武汉珈和科技有限公司 | Semilate rice information decision tree classification approach based on feature extraction in multi-temporal data |
| CN106339423A (en) * | 2016-08-14 | 2017-01-18 | 李宇翔 | Method and device for dynamically updating sugarcane planting information |
| CN106373150A (en) * | 2016-08-15 | 2017-02-01 | 中国农业大学 | Identification method of corns for seed on the basis of high resolution remote sensing data texture analysis |
| CN106355143A (en) * | 2016-08-25 | 2017-01-25 | 中国农业大学 | Seed maize field identification method and system based on multi-source and multi-temporal high resolution remote sensing data |
| US11432469B2 (en) | 2016-12-23 | 2022-09-06 | Hexagon Technology Center Gmbh | Method for prediction of soil and/or plant condition |
| CN108241905A (en) * | 2016-12-23 | 2018-07-03 | 赫克斯冈技术中心 | For predicting the method for soil and/or plant situation |
| CN106682624B (en) * | 2016-12-29 | 2019-08-02 | 中国科学院深圳先进技术研究院 | Built-up areas extracting method and device based on time series remote sensing information |
| CN106682624A (en) * | 2016-12-29 | 2017-05-17 | 中国科学院深圳先进技术研究院 | Built-up area extraction method and apparatus based on time sequence remote sensing information |
| CN107480634A (en) * | 2017-08-12 | 2017-12-15 | 天津市测绘院 | A kind of geographical national conditions ground mulching monitoring method based on multistage target classification |
| CN111630551A (en) * | 2018-01-23 | 2020-09-04 | X开发有限责任公司 | Crop type classification in images |
| US11321943B2 (en) | 2018-01-23 | 2022-05-03 | X Development Llc | Crop type classification in images |
| CN108399400A (en) * | 2018-03-23 | 2018-08-14 | 中国农业科学院农业资源与农业区划研究所 | A kind of early stage crop recognition methods and system based on high-definition remote sensing data |
| CN108647568A (en) * | 2018-03-30 | 2018-10-12 | 电子科技大学 | Grassland degeneration extraction method based on full convolutional neural networks |
| CN112258793A (en) * | 2020-10-19 | 2021-01-22 | 深圳前海微众银行股份有限公司 | Disaster monitoring and early warning method, device, equipment and medium |
| CN112258793B (en) * | 2020-10-19 | 2022-10-28 | 深圳前海微众银行股份有限公司 | Disaster monitoring and early warning method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105005782A (en) | Fine method for global vegetation classification based on multi-temporal remote sensing data and spectroscopic data | |
| Pham et al. | Biomass estimation of Sonneratia caseolaris (l.) Engler at a coastal area of Hai Phong city (Vietnam) using ALOS-2 PALSAR imagery and GIS-based multi-layer perceptron neural networks | |
| Gleason et al. | Forest biomass estimation from airborne LiDAR data using machine learning approaches | |
| Raitsos et al. | Identifying four phytoplankton functional types from space: An ecological approach | |
| Chen et al. | Estimating soil moisture over winter wheat fields during growing season using machine-learning methods | |
| Reddy et al. | Survey on weather prediction using big data analystics | |
| Vijayabaskar et al. | Crop prediction using predictive analytics | |
| Wong et al. | Combining eo-1 hyperion and envisat asar data for mangrove species classification in Mai Po Ramsar Site, Hong Kong | |
| Zhao et al. | Toward a better understanding of coastal salt marsh mapping: A case from China using dual-temporal images | |
| Do et al. | Monitoring landscape fragmentation and aboveground biomass estimation in Can Gio Mangrove Biosphere Reserve over the past 20 years | |
| CN101276420A (en) | Classification method for syncretizing optical spectrum information and multi-point simulation space information | |
| Shukla et al. | Performance analysis of different predictive models for crop classification across an aridic to ustic area of Indian states | |
| Jayanth et al. | Identification of land cover changes in the coastal area of Dakshina Kannada district, South India during the year 2004–2008 | |
| Vintrou et al. | Mapping fragmented agricultural systems in the Sudano-Sahelian environments of Africa using random forest and ensemble metrics of coarse resolution MODIS imagery | |
| Dong et al. | Digital mapping of soil available phosphorus supported by AI technology for precision agriculture | |
| Sardar et al. | Understanding the dynamics of landscape of greater Sundarban area using multi-layer perceptron Markov chain and landscape statistics approach | |
| Talukdar et al. | Wetland‐inundated area modeling and monitoring using supervised and machine learning classifiers | |
| Billah et al. | Random forest classifications for landuse mapping to assess rapid flood damage using Sentinel-1 and Sentinel-2 data | |
| Melati | The use of remote sensing data to monitor land use systems and forest variables of the tropical rainforest landscape under transformation in Jambi Province, Sumatra, Indonesia | |
| Latifi | Characterizing forest structure by means of remote sensing: a review | |
| Ayanlade | Remote sensing approaches for land use and land surface temperature assessment: a review of methods | |
| Gallo et al. | In-season and dynamic crop mapping using 3D convolution neural networks and sentinel-2 time series | |
| Jiang et al. | Monitoring the coastal environment using remote sensing and GIS techniques | |
| Maltamo et al. | Assessing dead wood by airborne laser scanning | |
| Brennan et al. | Comparing stability in random forest models to map Northern Great Plains plant communities in pastures occupied by prairie dogs using Pleiades imagery |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20151028 |
|
| WD01 | Invention patent application deemed withdrawn after publication |