CN105930863A - Determination method for spectral band setting of satellite camera - Google Patents
Determination method for spectral band setting of satellite camera Download PDFInfo
- Publication number
- CN105930863A CN105930863A CN201610232316.2A CN201610232316A CN105930863A CN 105930863 A CN105930863 A CN 105930863A CN 201610232316 A CN201610232316 A CN 201610232316A CN 105930863 A CN105930863 A CN 105930863A
- Authority
- CN
- China
- Prior art keywords
- classification
- spectral
- camera
- satellite
- deploying
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a determination method for spectral band setting of a satellite camera. The method comprises that multi-spectrum images obtained by an optical remote-sensing satellite are input and preprocessed; eCognition software is used to select classification and evaluation samples from the multi-spectrum images; monitoring classification of the same classification sample, the same classification algorithm and different spectral band characteristics is carried out on the multi-spectrum images; and a confusion matrix is used to calculate the precision of a classification result of the same evaluation sample and the different waveband characteristics, and a waveband setting determination result of the multispectral camera of the optical remote sensing satellite is given. According to the invention, whether waveband setting of the multispectral camera of the optical remote sensing satellite for different ground applications is reasonable is evaluated and determined effectively, optical waveband can be selected for different ground applications, and reference is provided for design and optimization of the waveband setting index of a spatial camera.
Description
Technical field
The invention belongs to Satellite Camera Index System Design field, relate generally to Satellite Camera spectral deploying really
Determine method, be specifically related to a kind of Optical remote satellite multispectral camera spectral coverage parameter based on land cover classification
The determination method arranged.
Background technology
It is panchromatic multispectral that the current Small Earth observation satellite of China mainly carries with optical imagery as remote sensing
Camera, the index Design of Satellite Camera is the emphasis in whole design of satellites and difficult point, Satellite Camera overall
Technical specification specifically includes that (spatial resolution, spectral resolution, radiation are differentiated for operating spectrum band, resolution
Rate, temporal resolution etc.), pointing capability requirement, certainty of measurement, MTF, dynamic range, signal to noise ratio etc..
The index of current camera mainly proposes on the basis of the observation mission analyzing user, as found, identifying
With confirm which kind of target and the positional precision etc. of target, be converted to the total technical index of space camera, then
According to the level of technology development, select the Technology Ways that camera is developed；According to inheritance and schedule requirement, really
Determine the scheme used by camera and components and parts, be finally completed the production of space camera, test and pay.
The remote sensing satellite developed according to this pattern can well meet the earth observation demand of user,
Especially the continuing to optimize and promoting of the index such as spatial resolution, revisiting period, fabric width so that remotely-sensed data
The most military investigation or civilian generaI investigation field, all play the important and pivotal role.But, along with fixed
Amount remote sensing and the development of remote sensing image information interpretation technology, Ground Application has been not limited solely to visual observation,
The angle extracted from terrestrial information designs and determines that the index of space camera more can meet following remote sensing satellite star
Ground integrated design concept and the application demand of remotely-sensed data.
Summary of the invention
Present invention solves the technical problem that and be: for how from remote sensing Ground Application angle, Satellite Camera to be set
The problem that meter index determines and optimizes, it is provided that a kind of Satellite Camera spectral deploying based on land cover classification
Determine method, it is possible to provide the spectral coverage that should preferentially ensure in design of satellites, for the optimization of satellite index Design
Reference is provided.
Technical scheme: a kind of determination method of Satellite Camera spectral deploying, comprises the steps of
1) multispectral image obtaining satellite spatial camera carries out data prediction；Described multispectral image bag
Include blue spectral coverage, green spectral coverage, red spectral coverage and near-infrared spectral coverage；
2) to through described step 1) process after multispectral image carry out multi-scale division；
3) to through described step 2) image after dividing processing, the classification system according to being previously set is entered
Row classification samples and the selection of evaluation sample；
4) to through described step 2) image after dividing processing, utilize step 3) classification samples chosen,
Carry out same category sample, same category algorithm, the supervised classification of different spectral coverage spectral signature；
5) to through described step 4) process after classification results, utilize step 3) the evaluation sample chosen,
Carry out the classification results precision evaluation of identical evaluation sample, identical evaluation algorithms；
6) according to step 5) process after precision evaluation result, relative analysis goes out satellite spatial camera for soil
The spectral deploying trap queuing of ground cover classification application, so that it is determined that Optical remote satellite multispectral camera is not to
Band setting with Ground Application is the most reasonable, and the optimum spectral coverage for Different Ground application, for space
Design and the optimization of camera band setting index provide reference.
Described step 1) in data prediction refer to satellite spatial camera obtain multispectral image system
System geometric correction, the operation of system radiant correction.
Described step 2) in multi-scale division refer to utilize the multi-scale division instrument of eCognition software,
Multi-scale division is carried out by arranging scale parameter, form parameter and compactness parameter.
Described step 3) in classification system refer to the classification system of land cover classification, including construction land,
Arable land, landscape ground, water body, unused land.
Described step 3) in classification samples select refer to utilize the samples selection instrument of eCognition software,
Be trained the selection of sample for each class in classification system, training sample is for follow-up supervised classification.
Described step 3) in evaluate samples selection refer to utilize the samples selection instrument of eCognition software,
It is evaluated the selection of sample for each class in classification system, evaluates sample for follow-up classification results
Precision evaluation.
Described step 4) in supervised classification refer to utilize nearest neighbour classification device, respectively with blue spectral coverage, green spectrum
Section, the spectrum average of red spectral coverage, near-infrared spectral coverage carry out four supervised classifications as feature.
Described step 5) in precision evaluation refer to for step 4) four classification results obtaining, utilize
Step 3) selected by evaluation sample, calculate different classes of Producer precision based on confusion matrix, Yong Hujing
Degree, overall accuracy and Kappa coefficient.
Described step 6) in spectral deploying trap queuing refer to according to step 5) different spectral coverage spectrum average
The overall accuracy of the 4 groups of classification results obtained and Kappa coefficient, divide at land cover pattern Satellite Camera spectral coverage
The quality of apoplexy due to endogenous wind sorts from high to low, determines that the spectral deploying of Satellite Camera and camera are preferential in manufacturing
The spectral coverage ensured；
Described step 6) in spectral deploying trap queuing refer to step 5) different spectral coverage spectrum is the most worth
To 4 groups of classification results in user's precision of certain classification and Producer precision sort from high to low, determine
The spectral deploying of Satellite Camera extracted for certain ground category information and camera manufacture in the preferential spectral coverage ensured.
Present invention advantage compared with prior art is: propose a kind of satellite based on land cover classification
The determination method of camera spectral deploying, counter pushes away camera index from the angle of application precision, and having filled up should from ground
With angle evaluation with determine that the method for Satellite Camera index Design is blank, solve satellite data Ground Application with
The problem that satellite index Design disconnects mutually.Certain spectral coverage is obtained to nicety of grading contribution if this method is analyzed
Little, can consider the when of camera spectral deploying to cancel；If certain spectral coverage is big to nicety of grading contribution, phase
Firmly to be ensured and to be improved the precision of radiation calibration when of machine spectral deploying.The present invention with existing from
The method of the angle-determining spectral deploying of optical design is compared, have application target clearly, clear principle, behaviour
Make the advantage that step is simple, cost is little, it is not necessary to large-scale outer field measurement and test, for follow-up space camera
The design of index and optimization provide feasible method, also analyze for Incorporateization design and provide a kind of method
Use for reference.
Accompanying drawing explanation
Fig. 1 is the flow chart of a kind of Satellite Camera spectral deploying evaluation methodology of the embodiment of the present invention；
Fig. 2 is the flow chart of supervised classification；
Fig. 3 is example GF-1 satellite 2m/8m camera multispectral image used by the present invention；
Fig. 4 is the classification system that example land cover classification used by the present invention is set up；
Fig. 5 is the comparison diagram of classification results precision evaluation in example used by the present invention；
Detailed description of the invention
The present invention, with number (GF-1) satellite 2m/8m camera multispectral image example of high score, illustrates a kind of satellite phase
Machine spectral deploying determines the detailed description of the invention of method.The shooting time of experiment image is in August, 2013, examination
Test district's thumbnail as shown in Figure 3.The present invention is further described below in conjunction with the accompanying drawings.
As it is shown in figure 1, be that the present invention implements Satellite Camera spectral deploying and determines the flow chart of method, the present embodiment
Comprise the steps:
Step 1, the multispectral image obtaining satellite spatial camera carry out data prediction；
Number multispectral data of the high score got is carried out geometric correction and the radiant correction of system, pretreatment
Result after completing is as shown in Figure 3；
Step 2, to through described step 1 process after multispectral image carry out multi-scale division；
Multi-scale division in described step 2 is that a kind of region from bottom to top started from single pixel object is closed
And technology.In step from bottom to top, little imaged object merges into big object, clusters two-by-two
Cheng Zhong, the heterogeneous weight of imaged object is minimized by the optimization process comprised.In each step, adjacent shadow
It is as if object meets the heterogeneous minimum growth conditions of regulation, the most merged.If minimum growth exceedes
During the threshold value that defined by scale parameter, this process just stops.Multi-scale division essence be exactly a local
The process optimized.Specifically comprise the following steps that
(1) partitioning parameters is set, i.e. stops potting gum condition including setting a yardstick threshold value.Image divides
The important parameter cut has figure layer weight, scale parameter, parameter color, smoothness and compactness etc..According to not
With figure layer to the importance of segmentation result and suitability, give different weights to each figure layer；According to classification
Task and target interested determine scale parameter；According to image information textural characteristics and extracted special topic letter
Breath requires the weight determining light spectroscopic factor with form factor；In form factor, according to most of ground species
Structure attribute determines the weight of light spectroscopic factor and form factor.
(2) segmentation is performed.Start centered by one in any image pixel of the segmentation of image to split, first
During secondary segmentation, single pixel is counted as the calculating of a heterogeneous value of minimum polygon object participation；For the first time
After having split, based on the polygon object generated, carry out second time split, same calculating heterogeneity value,
If it is less than given threshold value, then proceeds segmentation repeatedly, the most then stop the segmentation work of image,
Form the imaged object layer of a fixed ruler angle value.
This example utilizes the multi-scale division instrument of eCognition software, arrange scale parameter be 110, shape
Shape parameter is 0.1, compactness parameter is 0.5 to carry out multi-scale division.
Step 3, to the image after described step 2 dividing processing, according to the class complicated variant being previously set
System carries out classification samples and evaluates the selection of sample；
The data that this example uses are from GF-1 satellite, and the main purpose of this satellite is by Land_use change and adjusts
Look into, the present invention is directed to the feature of this satellite and set up taxonomic hierarchies and include construction land, arable land, landscape ground, water
Body, unused land etc., in order to improve the precision of classification, plough and be further divided into exposed arable land, vegetation
Highly dense arable land and the arable land of vegetation sparse, as shown in Figure 4；
After setting up classification system, it is necessary first to select training sample for each classification, then for each class
It is not evaluated the selection of sample.
Step 4, to the image after described step 2 dividing processing, utilize the classification sample that step 3 is chosen
This, carry out same category sample, same category algorithm, the supervised classification of different spectral coverage spectral signature；
Supervised classification in described step 4 refers to that the special characteristic according to object and expert knowledge library mate,
Obtain maximum similarity, then object is classified as such, realize the classification of image.Concretely comprise the following steps: first basis
The priori of classification determines discriminant function and corresponding criterion, wherein utilizes some known class
The observation of training sample determine the referred to as study of the process of undetermined parameter or training in discriminant function, so
After the sample observations of unknown classification is substituted into discriminant function, further according to the criterion affiliated class to this sample
Not decisioing making, the process of supervised classification is as shown in Figure 2.
This example utilizes the nearest neighbour classification device of eCognition software, respectively with blue spectral coverage, green spectral coverage, red
Spectral coverage, the spectrum average of near-infrared spectral coverage carry out four supervised classifications as feature, obtain four classification results.
Step 5, to through described step 4 process after classification results, utilize the evaluation sample that step 3 is chosen,
Carry out the classification results precision evaluation of identical evaluation sample, identical evaluation algorithms；
Nicety of grading evaluation in described step 5 have employed Accuracy Assessment based on confusion matrix.Obscure
Matrix is used to indicate that a kind of reference format of precision evaluation.Confusion matrix is the matrix of n row n row, wherein n generation
The quantity of table classification.Confusion matrix typically can be expressed as following form, as shown in table 1.This matrix column is
With reference to image information (ground truth i.e. obtained by human interpretation), behavior is evaluated image classification result letter
Breath (i.e. being automatically extracted the observation obtained by algorithm), the part that row and column intersects summarises and is categorized into and joins
Examine the number of samples in a certain particular category that classification is relevant, sample number can be pixel number or split right
As number.
Main diagonal element (the x of matrix11,x22,…xnn) be to be assigned to the sample number of correct classification, diagonal with
Outer element is the Classification in Remote Sensing Image mistake classification number relative to reference data, and rightmost string is that every classification is being divided
Total quantity on class figure, and a line of bottom is shown that every classification with reference to the total quantity on figure.Wherein, xij
It is the i-th class and the classification samples number of reference data type jth class in categorical data；For classification gained
The summation of the i-th class arrived；Summation for the jth class of reference data；N is for evaluating total sample number.
The confusion matrix table that table 1 is general
Based on confusion matrix, a series of evaluation index can be added up classification extraction result is evaluated, substantially
Evaluation index as follows:
(1) overall accuracy:
Overall accuracy is a statistic with probability meaning, and statement is to each random sample, institute
The probability that the actual type in region corresponding to result and the reference data of classification is consistent.
(2) user's precision:
User's accuracy representing is appointed from classification results and is taken a random sample, and its type being had and ground are real
The conditional probability that border type is identical.
(3) Producer precision:
Producer accuracy representing is relative to any one random sample in reference data, on classification chart samely
The classification results conditional probability consistent with it of point.
(4) Kappa coefficient
Overall accuracy, the objectivity of user's precision index depend on sample and method, are referring to these
After mark is analyzed, it is still necessary to a more objective index carrys out classification of assessment quality.It is a kind of quantitative that Kappa analyzes
Evaluating concordance or the method for precision between Classification in Remote Sensing Image figure and reference data, it uses discrete multiplex method,
More objectively classification of assessment quality, overcomes confusion matrix and excessively relies on the collection of sample and sample data
Journey.Kappa analyzes the evaluation index produced and is referred to as κhatStatistics, is that a kind of mensuration is coincide between two width figures
Degree or the index of precision.
In formula, n is total columns (the most total classification number) in confusion matrix；xiiBe the i-th row in confusion matrix,
The i-th upper sample size of row, the number of samples of i.e. correct classification；xi+And x+iBe respectively the i-th row and i-th row total
Sample size；N is total sample size for accuracy evaluation.Many scholars are for classification essence for a long time
The evaluation of degree is studied, and research conclusion is thought, κhatValue > 0.80 presentation class figure and ground reference information
Between concordance very big or precision is the highest, κhatAt 0.40-0.80, value represents that concordance is medium, κhatValue is less than
0.40 represents that concordance is very poor.Any negative κhatIt is worth all presentation class weak effects, but the scope of negative value depends on
Confusion matrix to be evaluated, therefore negative value size can not presentation class effect.
Confusion matrix is utilized to calculate the Producer precision of four spectral coverage classification results of this example, user's precision, total
Body precision and Kappa coefficient are as shown in table 2-table 5.
The blue spectral coverage nicety of grading of table 2 is evaluated
Table 3 green spectral coverage nicety of grading is evaluated
Table 4 red spectral coverage nicety of grading is evaluated
Table 5 near-infrared spectral coverage nicety of grading is evaluated
Step 6, according to step 5 process after precision evaluation result, relative analysis goes out satellite spatial camera pair
In the spectral deploying trap queuing of land cover classification application, provide the advisory opinion of space camera spectral coverage design；
This example draws comparison diagram such as Fig. 5 by utilizing each spectral coverage spectral signature precision of obtaining of classification that exercises supervision
Shown in, from overall accuracy and Kappa factor evaluation result it can be seen that near-infrared spectral coverage nicety of grading is the highest,
Overall accuracy has reached 87.6%, and Kappa coefficient is up to 0.849, and the precision of red spectral coverage is minimum, totally essence
Degree is 71.9%, Kappa coefficient only 0.666, therefore deduce that for land cover pattern information retrieval this
Ground Application, the sequence of importance of space camera spectral deploying is: near-infrared spectral coverage > blue spectral coverage > green spectral coverage >
Red spectral coverage, therefore when design is towards the space camera of territory application aspect, spectral deploying should include at least
Near-infrared spectral coverage, and should preferentially ensure performance and the precision of this spectral coverage radiation calibration of near-infrared spectral coverage,
Improve quality and the precision of the near-infrared spectra segment information that ground obtains, the most just can improve land cover pattern information
The precision extracted.
Claims (10)
1. the determination method of a Satellite Camera spectral deploying, it is characterised in that comprise the steps of
1) multispectral image obtaining satellite spatial camera carries out data prediction；Described multispectral image includes blue spectral coverage, green spectral coverage, red spectral coverage and near-infrared spectral coverage；
2) to through described step 1) process after multispectral image carry out multi-scale division；
3) to through described step 2) image after dividing processing, carry out classification samples according to the classification system being previously set and evaluate the selection of sample；
4) to through described step 2) image after dividing processing, utilize step 3) classification samples chosen, carry out same category sample, same category algorithm, the supervised classification of different spectral coverage spectral signature；
5) to through described step 4) process after classification results, utilize step 3) the evaluation sample chosen, carry out the classification results precision evaluation of identical evaluation sample, identical evaluation algorithms；
6) according to step 5) process after precision evaluation result, relative analysis goes out the spectral deploying trap queuing that satellite spatial camera is applied for land cover classification, so that it is determined that the band setting that Different Ground is applied by Optical remote satellite multispectral camera is the most reasonable, and the optimum spectral coverage for Different Ground application, design and optimization for space camera band setting index provide reference.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterised in that described step 1) in data prediction refer to satellite spatial camera obtain multispectral image carry out system geometric correction, system radiant correction operation.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 2) in multi-scale division refer to utilize the multi-scale division instrument of eCognition software, carry out multi-scale division by arranging scale parameter, form parameter and compactness parameter.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterised in that described step 3) in classification system refer to the classification system of land cover classification, including construction land, arable land, landscape ground, water body, unused land.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 3) in classification samples select refer to utilize the samples selection instrument of eCognition software, be trained the selection of sample for each class in classification system, training sample is for follow-up supervised classification.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 3) in evaluate samples selection refer to utilize the samples selection instrument of eCognition software, it is evaluated the selection of sample for each class in classification system, evaluates sample for follow-up classification results precision evaluation.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 4) in supervised classification refer to utilize nearest neighbour classification device, carry out four supervised classifications using blue spectral coverage, green spectral coverage, red spectral coverage, the spectrum average of near-infrared spectral coverage as feature respectively.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 5) in precision evaluation refer to for step 4) four classification results obtaining, utilize step 3) selected by evaluation sample, calculate different classes of Producer precision based on confusion matrix, user's precision, overall accuracy and Kappa coefficient.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 6) in spectral deploying trap queuing refer to according to step 5) overall accuracy of 4 groups of classification results that is all worth to of different spectral coverage spectrum and Kappa coefficient, Satellite Camera spectral coverage quality in land cover classification is sorted from high to low, determine the spectral deploying of Satellite Camera and camera manufacture in the preferential spectral coverage ensured.
The determination method of a kind of Satellite Camera spectral deploying the most as claimed in claim 1, it is characterized in that, described step 6) in spectral deploying trap queuing refer to step 5) user's precision and the Producer precision of certain classification sort from high to low in 4 groups of classification results being all worth to of different spectral coverage spectrum, determine the spectral deploying of the Satellite Camera extracted for certain ground category information and camera manufacture in the preferential spectral coverage ensured.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610232316.2A CN105930863A (en) | 2016-04-14 | 2016-04-14 | Determination method for spectral band setting of satellite camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610232316.2A CN105930863A (en) | 2016-04-14 | 2016-04-14 | Determination method for spectral band setting of satellite camera |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105930863A true CN105930863A (en) | 2016-09-07 |
Family
ID=56838200
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610232316.2A Pending CN105930863A (en) | 2016-04-14 | 2016-04-14 | Determination method for spectral band setting of satellite camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105930863A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108282631A (en) * | 2017-01-06 | 2018-07-13 | 北京比兴科技有限公司 | Integrated space camera automatization test system |
| CN109271949A (en) * | 2018-09-28 | 2019-01-25 | 中国科学院长春光学精密机械与物理研究所 | Multispectral image data extraction method, device, equipment and readable storage medium storing program for executing |
| CN109472294A (en) * | 2018-10-15 | 2019-03-15 | 广州地理研究所 | A kind of recognition methods of urban water-body, device, storage medium and equipment |
| CN111898503A (en) * | 2020-07-20 | 2020-11-06 | 中国农业科学院农业资源与农业区划研究所 | Crop identification method and system based on cloud coverage remote sensing image and deep learning |
| CN115082452A (en) * | 2022-07-26 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Cloud and shadow based quantitative evaluation method for quality of remote sensing image |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102289673A (en) * | 2011-06-22 | 2011-12-21 | 复旦大学 | Method for selecting hyperspectral remote sensing image bands based on partial least squares |
-
2016
- 2016-04-14 CN CN201610232316.2A patent/CN105930863A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102289673A (en) * | 2011-06-22 | 2011-12-21 | 复旦大学 | Method for selecting hyperspectral remote sensing image bands based on partial least squares |
Non-Patent Citations (3)
| Title |
|---|
| PAUL SHANE FRAZIER ETAL.: "Water Body Detection and Delineation with Landsat TM Data", 《PHOTOGRAMMETRIC ENGINEERING &REMOTE SENSING》 * |
| 刘顺喜等: "面向土地资源管理的我国后续资源卫星谱段设置方法研究", 《地理与地理信息科学》 * |
| 许海雨: "基于高分辨率遥感影像的地类区划", 《中国优秀硕士学位论文数据库农业科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108282631A (en) * | 2017-01-06 | 2018-07-13 | 北京比兴科技有限公司 | Integrated space camera automatization test system |
| CN109271949A (en) * | 2018-09-28 | 2019-01-25 | 中国科学院长春光学精密机械与物理研究所 | Multispectral image data extraction method, device, equipment and readable storage medium storing program for executing |
| CN109472294A (en) * | 2018-10-15 | 2019-03-15 | 广州地理研究所 | A kind of recognition methods of urban water-body, device, storage medium and equipment |
| CN111898503A (en) * | 2020-07-20 | 2020-11-06 | 中国农业科学院农业资源与农业区划研究所 | Crop identification method and system based on cloud coverage remote sensing image and deep learning |
| CN115082452A (en) * | 2022-07-26 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Cloud and shadow based quantitative evaluation method for quality of remote sensing image |
| CN115082452B (en) * | 2022-07-26 | 2022-11-04 | 北京数慧时空信息技术有限公司 | Cloud and shadow based quantitative evaluation method for quality of remote sensing image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hu et al. | A phenology-based spectral and temporal feature selection method for crop mapping from satellite time series | |
| Cutler et al. | Estimating tropical forest biomass with a combination of SAR image texture and Landsat TM data: An assessment of predictions between regions | |
| CN103440505B (en) | The Classification of hyperspectral remote sensing image method of space neighborhood information weighting | |
| CN109063754B (en) | Remote sensing image multi-feature joint classification method based on OpenStreetMap | |
| CN103679675B (en) | Remote sensing image fusion method oriented to water quality quantitative remote sensing application | |
| CN109325431B (en) | Method and device for detecting vegetation coverage in feeding path of grassland grazing sheep | |
| CN105930863A (en) | Determination method for spectral band setting of satellite camera | |
| CN105930772A (en) | City impervious surface extraction method based on fusion of SAR image and optical remote sensing image | |
| Su et al. | Machine learning-assisted region merging for remote sensing image segmentation | |
| Sharma et al. | Geospatial technology in agroforestry: status, prospects, and constraints | |
| LUO et al. | Comparison of machine learning algorithms for mapping mango plantations based on Gaofen-1 imagery | |
| Shu et al. | Improving the estimation accuracy of SPAD values for maize leaves by removing UAV hyperspectral image backgrounds | |
| Zhang et al. | Mapping freshwater marsh species in the wetlands of Lake Okeechobee using very high-resolution aerial photography and lidar data | |
| Pan et al. | Mapping cropland distributions using a hard and soft classification model | |
| CN114266961A (en) | Method for integrating, learning and classifying marsh vegetation stacks by integrating hyperspectral and multiband fully-polarized SAR images | |
| CN113657158A (en) | Google Earth Engine-based large-scale soybean planting region extraction algorithm | |
| Jónsson | RGB and Multispectral UAV image classification of agricultural fields using a machine learning algorithm | |
| CN114332534B (en) | Hyperspectral image small sample classification method | |
| CN113936214B (en) | Karst wetland vegetation community classification method based on fusion of aerospace remote sensing images | |
| Asming et al. | Processing and classification of landsat and sentinel images for oil palm plantation detection | |
| CN113205014B (en) | Time sequence data farmland extraction method based on image sharpening | |
| Küchler et al. | Combining remotely sensed spectral data and digital surface models for fine-scale modelling of mire ecosystems | |
| Friedl et al. | An overview of uncertainty in optical remotely sensed data for ecological applications | |
| CN104392454A (en) | Merging method of membership scoring based on ground object categories under spatial-spectral combined classification frame for hyper-spectral remote sensing images | |
| Qu et al. | Mapping large area tea plantations using progressive random forest and Google Earth Engine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160907 |
|
| RJ01 | Rejection of invention patent application after publication |