CN114494882B - Winter wheat remote sensing identification analysis method and system based on random forest - Google Patents
Winter wheat remote sensing identification analysis method and system based on random forest Download PDFInfo
- Publication number
- CN114494882B CN114494882B CN202210116672.3A CN202210116672A CN114494882B CN 114494882 B CN114494882 B CN 114494882B CN 202210116672 A CN202210116672 A CN 202210116672A CN 114494882 B CN114494882 B CN 114494882B
- Authority
- CN
- China
- Prior art keywords
- image
- winter wheat
- remote sensing
- synthetic image
- precision
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Probability & Statistics with Applications (AREA)
- Image Processing (AREA)
Abstract
The invention provides a winter wheat remote sensing identification analysis method and system based on a random forest. The method comprises the following steps: generating index features by using the image set, then fusing the image set, the index features and the topographic features to obtain a second synthetic image, calculating texture features by using the second synthetic image, and finally fusing the texture features with the second synthetic image to obtain a third synthetic image; training a random forest model by using the training set and the third synthetic image as random forest classification model input; calculating a confusion matrix by using the test set to evaluate the precision of the classification result of the winter wheat at each growth period; and selecting the remote sensing image of the growth period stage with highest separability and classification result precision to generate a winter wheat spatial distribution map, obtaining an extraction area, and applying the extraction area and a ground real value to perform precision evaluation on the extraction area. The scheme provided by the invention has the advantages of high overall precision, good classification effect and highest remote sensing identification precision of winter wheat in the jointing and heading period.
Description
Technical Field
The invention belongs to the technical field of winter wheat remote sensing identification, and particularly relates to a winter wheat remote sensing identification analysis method and system based on a random forest.
Background
At present, researches are carried out to calculate separability between winter wheat in different growth stages and other land utilization coverage types through a Jeffries-Matusita (J-M) distance method, and finally determine the sentinel-2 image in the heading stage as the optimal period for extracting the areas of the winter wheat in north and middle regions of Anhui province.
The method also researches and utilizes a time polymerization technology, combines Landsat-8OLI and sentinel-2 data to explore the remote sensing identification condition of winter wheat in each growth period of Shandong province, finally determines that the data of the maturity period and the green turning period are more effective, and the remote sensing identification effect of the winter wheat is better.
The prior art has the following defects:
the researches use a small amount of features for remote sensing identification and classification, but neglect the influence of other features such as terrain, texture and the like on the remote sensing identification and classification of the winter wheat.
Disclosure of Invention
In order to solve the technical problems, the invention provides a technical scheme of a winter wheat remote sensing identification analysis method and system based on a random forest, so as to solve the technical problems.
The invention discloses a winter wheat remote sensing identification analysis method based on a random forest, which comprises the following steps:
step S1, sample data preparation: collecting winter wheat actual measurement sample data to obtain a training set and a test set;
s2, extracting and preprocessing a synthetic image: acquiring a 2A-level earth surface reflectivity product in a sentinel-2 remote sensing image, preprocessing the 2A-level earth surface reflectivity product to obtain an image set, and synthesizing the image set to obtain a first synthesized image;
step S3, divisibility calculation: calculating separability between different covers by using the first synthetic image, and evaluating the separability between the winter wheat at each growth period stage and the different covers;
s4, classification characteristic construction: generating an index feature by using the image set, then fusing the image set, the index feature and a topographic feature to obtain a second synthetic image, calculating a texture feature by using the second synthetic image, and finally fusing the texture feature and the second synthetic image to obtain a third synthetic image;
s5, constructing and training a random forest classification model: applying the training set and the third synthetic image as random forest classification model input to train the random forest model;
s6, precision evaluation of classification results: calculating a confusion matrix by using the test set to evaluate the precision of the classification result of the winter wheat at each growth period;
s7, spatial mapping and area extraction of winter wheat: selecting a remote sensing image in a growth period stage with highest separability and classification result precision to generate a winter wheat space distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period with highest separability and classification result precision to obtain an extraction area, and performing precision evaluation on the extraction area by using the extraction area and the ground true value.
According to the method of the first aspect of the present invention, in the step S1, the method of sample data preparation further includes:
collecting time data of winter wheat growth and development stages of an agricultural meteorological site, dividing each growth period of the winter wheat, and dividing the growth period of the winter wheat into 5 growth period stages.
According to the method of the first aspect of the present invention, in step S2, the specific method for obtaining the 2A-level surface reflectivity product in the remote sensing image of sentinel-2 includes:
obtaining 2A-level surface reflectivity products in sensory-2 remote sensing images of 5 growth stages in a research area;
the specific method for preprocessing the 2A-level surface reflectivity product to obtain the image set comprises the following steps:
carrying out cloud removal on the remote sensing image by using a quality control band mark, and then selecting an image set of 10 spectral bands with spatial resolution of 10 meters and 20 meters of the remote sensing image;
the specific method for synthesizing the image set to obtain the first synthesized image includes:
and respectively carrying out median synthesis on 10 spectral bands of the remote sensing images in the 5 growth periods, and cutting to an interested area to obtain five first synthetic images with 10 spectral bands.
According to the method of the first aspect of the present invention, in step S3, the specific method for calculating separability between different ground covers using the first composite image includes:
and calculating separability between different ground covers by using the first synthetic image through the J-M distance.
According to the method of the first aspect of the present invention, in step S4, the specific method for generating the index feature by using the image set includes:
and selecting corresponding bands from the 10 spectral bands to calculate and obtain commonly used 8 exponential features: a normalized vegetation index, an enhanced vegetation index, a soil conditioning vegetation index, a normalized water body index, a normalized difference water body index, a normalized construction index, a red edge normalized index, and a red edge position index;
the specific method for obtaining the second synthetic image by fusing the image set, the index features and the terrain features comprises the following steps:
adding the 8 indexes as independent wave bands to each image of 10 spectral wave bands respectively, and then performing median synthesis on each added image to obtain a first process synthetic image with 18 wave bands;
the altitude, the gradient, the slope direction and the 4 topographic features of the mountain shadow are taken as independent wave bands and added into the first process synthetic image, and a second synthetic image with 22 wave bands is generated.
According to the method of the first aspect of the present invention, in step S4, the specific method for obtaining a third synthetic image by calculating the texture feature using the second synthetic image and finally fusing the texture feature with the second synthetic image includes:
calculating by using the second synthetic image to obtain 6 texture features of angular second moment, contrast, correlation, variance, inverse difference moment and entropy, and adding the 6 texture features as independent wave bands to the second synthetic image to generate a second process synthetic image containing 28 features;
and resampling the second process synthetic image to 10m spatial resolution to obtain the third synthetic image.
According to the method of the first aspect of the present invention, in the step S6, the specific method for evaluating the accuracy of the classification result of winter wheat at each stage of the growing season by applying the test set to calculate the confusion matrix comprises:
selecting four evaluation indexes of user precision, drawing precision, overall precision and Kappa coefficient to comprehensively reflect the recognition condition of the random forest winter wheat, wherein the collective calculation formula is as follows:
wherein the content of the first and second substances,
UA i : user precision;
PA i : drawing precision;
OA: the overall accuracy;
kappa: a Kappa coefficient;
n: testing the total number of the concentrated samples;
m: the number of categories;
n i : predicting the number of the categories i in the categories i;
N i : predicting the total number of the categories i;
M i : the total number of classes i in the test set.
The invention discloses a winter wheat remote sensing identification analysis system based on random forest, the system includes:
the first processing module is configured to collect the actual measurement sample data of the winter wheat to obtain a training set and a test set;
the second processing module is configured to acquire a 2A-level ground surface reflectivity product in the sensory-2 remote sensing image, preprocess the 2A-level ground surface reflectivity product to obtain an image set, and synthesize the image set to obtain a first synthesized image;
a third processing module configured to calculate separability between different covers using the first composite image, and evaluate separability between the winter wheat at each growth stage and the different covers;
a fourth processing module, configured to generate an index feature by applying the image set, then fuse the image set, the index feature and a topographic feature to obtain a second synthetic image, then calculate a texture feature by applying the second synthetic image, and finally fuse the texture feature with the second synthetic image to obtain a third synthetic image;
a fifth processing module configured to apply the training set and the third synthetic image as random forest classification model inputs, training the random forest model;
a sixth processing module configured to apply the test set to calculate a confusion matrix to evaluate the accuracy of the classification result of winter wheat at each stage of the growing period;
the seventh processing module is configured to select a remote sensing image in a growth period stage with the highest separability and classification result precision to generate a winter wheat spatial distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period with highest separability and classification result precision to obtain an extraction area, and performing precision evaluation on the extraction area by using the extraction area and the ground true value.
A third aspect of the invention discloses an electronic device. The electronic equipment comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for identifying and analyzing the winter wheat based on the random forest remote sensing in the first aspect of the disclosure are realized.
A fourth aspect of the invention discloses a computer-readable storage medium. The computer readable storage medium stores thereon a computer program which, when executed by the processor, implements the steps of the method for remote sensing, identifying and analyzing winter wheat based on random forests according to any one of the first aspect of the present disclosure.
The scheme provided by the invention has the advantages that the random forest-based classification method is high in overall precision and good in classification effect, wherein the remote sensing identification precision of winter wheat in the heading stage of the jointing and heading stage is highest, and the remote sensing identification precision of the tillering stage of seedling emergence is lowest. The area extraction precision of the winter wheat in the jointing and heading period research area by the random forest method is high.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a winter wheat remote sensing identification analysis method based on a random forest according to an embodiment of the invention;
FIG. 2 is a 2019-2020 spatial distribution diagram of winter wheat in the main region of North Henan, generated by a random forest method according to an embodiment of the present invention;
FIG. 3 is a structural diagram of a remote sensing recognition analysis system for winter wheat based on a random forest according to an embodiment of the invention;
fig. 4 is a block diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention discloses a winter wheat remote sensing identification analysis method based on a random forest. Fig. 1 is a flowchart of a remote sensing recognition analysis method for winter wheat based on a random forest according to an embodiment of the present invention, as shown in fig. 1, the method includes:
step S1, sample data preparation: collecting actual measurement sample data of winter wheat to obtain a training set and a test set;
s2, extracting and preprocessing a synthetic image: acquiring a 2A-level earth surface reflectivity product in a sentinel-2 remote sensing image, preprocessing the 2A-level earth surface reflectivity product to obtain an image set, and synthesizing the image set to obtain a first synthesized image;
step S3, separability calculation: calculating separability between different ground covers by using the first synthetic image, and evaluating the separability between the winter wheat at each growth period stage and the different ground covers;
s4, classification characteristic construction: generating an index feature by using the image set, then fusing the image set, the index feature and a topographic feature to obtain a second synthetic image, calculating a texture feature by using the second synthetic image, and finally fusing the texture feature and the second synthetic image to obtain a third synthetic image;
s5, constructing and training a random forest classification model: applying the training set and the third synthetic image as random forest classification model input to train the random forest model;
s6, evaluating the precision of the classification result: calculating a confusion matrix by using the test set to evaluate the precision of the classification result of the winter wheat at each growth period;
s7, spatial mapping and area extraction of winter wheat: selecting a remote sensing image in a growth period stage with highest separability and classification result precision to generate a winter wheat space distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period with highest separability and classification result precision to obtain an extraction area, and performing precision evaluation on the extraction area by using the extraction area and the ground true value.
In step S1, sample data preparation: and collecting the actual measurement sample data of the winter wheat to obtain a training set and a test set.
In some embodiments, in the step S1, the method of sample data preparation further comprises:
collecting time data of winter wheat growth and development stages of an agricultural meteorological site, dividing each growth period of the winter wheat, and dividing the growth period of the winter wheat into 5 growth period stages.
Specifically, winter wheat sample data is mainly obtained through field actual measurement, and non-winter wheat sample data is obtained through visual interpretation based on a GEE platform besides field actual measurement. And exporting all the collected points into a KML format, and introducing the KML format into Google Earth for inspection to remove points with obvious errors. Randomly dividing the sample points into two parts, and making 70% of the sample points into a training set for training and classification; 30% are made into a test set for accuracy assessment.
And (3) acquiring data of the agricultural meteorological site, wherein the data of the agricultural meteorological site records the specific time of the growth and development stage of the winter wheat, so that the data can be used for dividing each growth period of the winter wheat. According to the data of 7 agricultural gas stations in the research area, the whole growth cycle of winter wheat in the research area from 10 months and 10 days in 2019 to 6 months and 10 days in 2020 is divided into five growth period stages.
In step S2, the extraction and preprocessing of the composite image: the method comprises the steps of obtaining a 2A-level earth surface reflectivity product in a sentinel-2 remote sensing image, preprocessing the 2A-level earth surface reflectivity product to obtain an image set, and synthesizing the image set to obtain a first synthesized image.
In some embodiments, in the step S2, the specific method for acquiring 2A-level surface reflectance products in the remote sensing image of sentinel-2 includes:
obtaining 2A-level surface reflectivity products in sensory-2 remote sensing images of 5 growth stages in a research area;
the specific method for preprocessing the 2A-level surface reflectivity product to obtain the image set comprises the following steps:
carrying out cloud removal on the remote sensing image by using a quality control band mark, and then selecting an image set of 10 spectral bands with spatial resolution of 10 meters and 20 meters of the remote sensing image;
the specific method for synthesizing the image set to obtain the first synthesized image includes:
and respectively carrying out median synthesis on 10 spectral bands of the remote sensing images in the 5 growth periods, and cutting to an interested area to obtain five first synthetic images with 10 spectral bands.
Specifically, based on the GEE platform, a (COPERNICUS/S2 _ SR) image set is called to respectively obtain a 2A-level surface reflectivity product in a sensory-2 remote sensing image of five growth periods of a research area, and the product is subjected to radiation correction, atmospheric correction and orthorectification.
And (3) utilizing a quality control waveband (QA 60) mark to realize cloud removal, and then selecting 10 spectral wavebands with spatial resolution of 10 meters and 20 meters of the sentinenl-2 remote sensing image. The 10 bands include red, green, blue, near infrared bands (band 4, band3, band2, band 8) with 10m spatial resolution, red edge 1, red edge 2, red edge 3, narrow near infrared, short wave infrared 1, short wave infrared 2 bands (band 5, band6, band7, band8A, band, band 12) with 20m spatial resolution.
And respectively carrying out median synthesis on 10 spectral bands of the remote sensing images in the 5 growth periods, and cutting to an interested area to obtain five first synthetic images with 10 spectral bands.
In step S3, separability calculation: calculating separability between different covers using the first composite image, and evaluating the separability between the winter wheat at each growth period stage and the different covers.
In some embodiments, in step S3, the specific method for calculating separability between different ground covers using the first composite image includes:
and calculating separability between different ground covers by using the first synthetic image through the J-M distance.
Specifically, five first synthetic images with 10 spectral bands are used for calculating the J-M distance between 5 soil covers in the GEE so as to compare and analyze the separability between winter wheat and other soil covers in different growth periods. Types of soil cover include winter wheat, buildings and roads, other vegetation (garlic, corn, vegetable fields, grass), forests and water. Each type takes 10% of the entire sample set (this step is implemented in ArcGIS) as samples for J-M distance calculation in an equal-proportion random sampling manner. Under the assumption that the features accord with normality (if the features do not accord with a normal distribution rule, the feature separability is considered to be poor, and classification is not considered), the J-M distance calculation formula is as follows:
J＝2(1-e -B )
in the formula: b represents a bus distance, m i And m j Mean and variance of spectral reflectivities of classes i and j, respectively. The value range of J is 0-2,0, which means that the two categories are almost inseparable, 2 means that the two categories can be completely separated, and the larger the J is, the better the separability between the ground covers is.
The separability of winter wheat, forest and water in each growth period of the research area is obtained through calculation; however, there is a large difference in separability between winter wheat and buildings and roads, other vegetation (corn, garlic, vegetable fields, grasslands). The separability between winter wheat and other cover types is highest at the heading stage, the J-M distance between the winter wheat and other covers is 1.98, and the separability is only 1.46 at the tillering stage of sowing. The results of the J-M distance indicate that the heading stage is the best stage to distinguish winter wheat from other cover types.
In step S4, the classification feature construction: and generating an index feature by using the image set, then fusing the image set, the index feature and the topographic feature to obtain a second synthetic image, calculating a texture feature by using the second synthetic image, and finally fusing the texture feature and the second synthetic image to obtain a third synthetic image.
In some embodiments, in step S4, the specific method for generating the index feature by applying the image set includes:
and selecting corresponding bands from the 10 spectral bands to calculate and obtain commonly used 8 exponential features: a normalized vegetation index, an enhanced vegetation index, a soil conditioning vegetation index, a normalized water body index, a normalized difference water body index, a normalized construction index, a red edge normalized index, and a red edge position index;
the specific method for obtaining the second synthetic image by fusing the image set, the index features and the terrain features comprises the following steps:
adding the 8 exponential features as independent wave bands to each image of 10 spectral wave bands respectively, and then performing median synthesis on each added image to obtain a first process synthetic image with 18 wave bands;
and adding the altitude, the slope, the sloping direction and4 topographic features of mountain shadow as independent wave bands into the first process synthetic image to generate a second synthetic image with 22 wave bands.
The specific method for obtaining a third synthetic image by applying the second synthetic image to calculate texture features and finally fusing the texture features with the second synthetic image comprises the following steps:
calculating by using the second synthetic image to obtain 6 texture features of angular second moment, contrast, correlation, variance, inverse difference moment and entropy, and adding the 6 texture features as independent wave bands to the second synthetic image to generate a second process synthetic image containing 28 features;
and resampling the second process synthetic image to 10m spatial resolution to obtain the third synthetic image.
In step S5, constructing and training a random forest classification model: and applying the training set and the third synthetic image as random forest classification model input to train the random forest model.
Specifically, the random forest is a classifier comprising a plurality of decision trees, two thirds of data are randomly extracted from a training set to create the training set by adopting a bootstrap step-by-step guided sampling strategy, a decision tree is respectively established for each training set, one third of data is not extracted every time of sampling, and the internal error unbiased estimation is carried out on the part of data, so that the out-of-bag error is generated and is used for evaluating the classification precision of the random forest.
The random forest implementation process comprises the following steps:
first, a step-by-step guided sampling technique is used to extract N training sets from the training set, wherein the size of each training set is about 2/3 of the training set.
Establishing a regression tree for each training set respectively to generate a forest consisting of N decision trees; in the growth process of each tree, M characteristic variables (M is less than or equal to M) are randomly selected from all M characteristic variables, and the optimal attribute is selected from the M attributes according to the minimum principle of the Gini index to carry out internal node branching, so that each tree grows fully, and pruning operation is not usually carried out.
And finally, collecting the prediction results of the N decision trees, and determining the category of the new sample by adopting a voting mode.
About 1/3 of the data is not decimated per sample, this portion of the data is commonly referred to as out-of-bag data (OOB); the out-of-bag data is used for internal error estimation, resulting in OOB errors, which are used to predict the accuracy of the classification.
The random forest classification model is constructed by calling the (ee.classic. Smiledrandomforest) function inside the GEE. And sending the training set and the third synthetic image into a random forest model for training. The following parameters are mainly set: the number of decision trees in the forest. The number of decision trees is set to 10 to 200 in the GEE, respectively, increased by 10 each time. And then respectively sending the training set and the third synthesized image into a random forest model for training, and verifying by using the test set to obtain the precision under the quantity of the decision trees. The number of compared decision trees is set to 200 for best results.
Other parameters are kept as defaults.
In step S6, the classification result precision is evaluated: and (3) calculating a confusion matrix by using the test set to evaluate the precision of the classification result of the winter wheat at each growth period stage.
In some embodiments, in step S6, the specific method for evaluating the accuracy of the classification result of winter wheat at each stage of the growth period by applying the test set to calculate the confusion matrix includes:
selecting four evaluation indexes of user precision, drawing precision, overall precision and Kappa coefficient to comprehensively reflect the recognition condition of the random forest winter wheat, wherein the collective calculation formula is as follows:
wherein the content of the first and second substances,
UA i : user precision;
PA i : drawing precision;
OA: the overall accuracy;
kappa: a Kappa coefficient;
n: testing the total number of the concentrated samples;
m: the number of categories;
n i : predicted as one of the categories i actually being the category iCounting;
N i : predicting the total number of the categories i;
M i : the total number of classes i in the test set.
In step S7, spatial mapping and area extraction of winter wheat: selecting a remote sensing image in a growth period stage with highest separability and classification result precision to generate a winter wheat space distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period stage with the highest separability and classification result precision to obtain an extraction area, and applying the extraction area and the ground true value to perform precision evaluation on the extraction area.
Specifically, the remote sensing identification conditions of winter wheat in each growth period stage are compared, and the fact that the classification precision of the heading period of the jointing is the highest and the result obtained by the J-M distance is consistent is found. And then selecting the remote sensing image with the highest precision in the jointing and heading period to generate a winter wheat spatial distribution map.
The winter wheat area was then counted in GEE using the ee. The areas of the winter wheat of each pixel element in the period are summed to calculate the extraction area of the winter wheat of the whole research area. The winter wheat seeding area precision is the ratio of the estimated extraction area of a research area to the ground true value, and by combining with the official agricultural statistics yearbook, the precision evaluation is carried out on the extraction area by adopting the following formula:
in the formula, P represents the area extraction precision, S represents the winter wheat planting area extracted by a random forest method, and S' represents the real winter wheat planting area on the ground.
The extraction areas of the winter wheat in the jointing and heading period research area by the random forest method are 979.67 kilo hectares respectively, and the area extraction precision is 96.72 percent.
As shown in fig. 2, it is a spatial distribution diagram of the winter wheat in the main area of north yunnan of 2019-2020, which is generated by using the random forest method in the present invention. The classification method based on the random forest has the advantages that the overall accuracy is over 91 percent, the Kappa coefficient is over 0.89, the classification effect is good, the remote sensing identification accuracy of the winter wheat in the heading stage is highest, the overall accuracy is 96.90 percent, and the Kappa coefficient is 0.96; the remote sensing identification precision of the tillering stage of seedling emergence is the lowest, and the overall precision and the Kappa coefficient are respectively 91.99 percent and 0.89 percent.
According to the scheme provided by the invention, the overall accuracy of the classification method based on the random forest is over 91 percent, the Kappa coefficients are over 0.89, and the classification effect is good, wherein the remote sensing identification accuracy of the winter wheat in the heading stage of the jointing is highest, the overall accuracy is 96.90 percent, and the Kappa coefficient is 0.96; the remote sensing identification precision of the tillering stage of seedling emergence is the lowest, and the overall precision and the Kappa coefficient are respectively 91.99 percent and 0.89 percent. The extraction areas of the winter wheat in the jointing and heading period research area by the random forest method are 979.67 kilo hectares respectively, and the area extraction precision is 96.72 percent.
In addition, compared with the previous research, the research adds topographic and textural features into random forest categories, so that the overall recognition precision and Kappa coefficient of the winter wheat in the heading stage are respectively improved by 1.71% and 0.02. Previous studies only used one classification method, and did not evaluate the use of other methods in the identification of winter wheat in different growth periods.
The research also uses a classification regression tree and a support vector machine classifier to evaluate the identification precision of winter wheat in the jointing and heading period, and the experimental results are shown in the following table 1,
TABLE 1 winter wheat precision evaluation indexes of different classifiers
| Classifier | User accuracy | Accuracy of producer | Overall accuracy | Kappa coefficient |
| Random forest | 0.97 | 0.95 | 0.97 | 0.96 |
| Categorizing regression trees | 0.94 | 0.94 | 0.93 | 0.92 |
| Support vector machine | 0.97 | 0.95 | 0.92 | 0.90 |
The results show that the random forest classifier works best compared to other classifiers.
The invention discloses a winter wheat remote sensing identification analysis system based on a random forest. FIG. 3 is a structural diagram of a remote sensing recognition analysis system for winter wheat based on a random forest according to an embodiment of the invention; as shown in fig. 3, the system 100 includes:
the first processing module 101 is configured to collect actual measurement sample data of winter wheat to obtain a training set and a test set;
the second processing module 102 is configured to obtain a 2A-level surface reflectivity product in a sensory-2 remote sensing image, preprocess the 2A-level surface reflectivity product to obtain an image set, and synthesize the image set to obtain a first synthesized image;
a third processing module 103 configured to calculate separability between different covers using the first composite image, and evaluate separability between the winter wheat at each growth stage and the different covers;
a fourth processing module 104, configured to apply the image set to generate an index feature, then fuse the image set, the index feature and a topographic feature to obtain a second synthetic image, then apply the second synthetic image to calculate a texture feature, and finally fuse the texture feature and the second synthetic image to obtain a third synthetic image;
a fifth processing module 105 configured to apply the training set and the third synthetic image as a random forest classification model input, training the random forest model;
a sixth processing module 106 configured to apply the test set to calculate a confusion matrix to evaluate the accuracy of the classification result of winter wheat at each stage of the growing season;
the seventh processing module 107 is configured to select a remote sensing image in a growth period stage with the highest separability and classification result precision to generate a winter wheat spatial distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period stage with the highest separability and classification result precision to obtain an extraction area, and applying the extraction area and the ground true value to perform precision evaluation on the extraction area.
According to the system of the second aspect of the present invention, the first processing module 101 is specifically configured to, the method for sample data preparation further includes:
collecting time data of winter wheat growth and development stages of an agricultural meteorological site, dividing each growth period of the winter wheat, and dividing the growth period of the winter wheat into 5 growth period stages.
Specifically, the data of the winter wheat sample actually measured is mainly obtained by field actual measurement, and the data of the non-winter wheat sample is obtained by visual interpretation based on a GEE platform besides the field actual measurement. And exporting all the collected points into a KML format, and introducing the KML format into Google Earth for inspection to remove points with obvious errors. Randomly dividing the sample points into two parts, and making 70% of the sample points into a training set for training and classification; 30% are made into a test set for accuracy assessment.
And (3) acquiring data of the agricultural meteorological site, wherein the data of the agricultural meteorological site records the specific time of the growth and development stage of the winter wheat, so that the data can be used for dividing each growth period of the winter wheat. According to the method, the whole growth cycle of winter wheat in the research area from 10 months and 10 days in 2019 to 6 months and 10 days in 2020 is divided into five growth period stages according to 7 agricultural gas station data in the research area.
According to the system of the second aspect of the present invention, the second processing module 102 is specifically configured to, the specific method for acquiring 2A-level surface reflectance products in a sensory-2 remote sensing image includes:
obtaining 2A-level surface reflectivity products in sensory-2 remote sensing images of 5 growth stages in a research area;
the specific method for preprocessing the 2A-level surface reflectivity product to obtain the image set comprises the following steps:
carrying out cloud removal on the remote sensing image by using a quality control band mark, and then selecting an image set of 10 spectral bands with spatial resolution of 10 meters and 20 meters of the remote sensing image;
the specific method for synthesizing the image set to obtain the first synthesized image includes:
and respectively carrying out median synthesis on 10 spectral bands of the remote sensing images in the 5 growth periods, and cutting to an interested area to obtain five first synthetic images with 10 spectral bands.
Specifically, based on the GEE platform, a (COPERNICUS/S2 _ SR) image set is called to respectively obtain a 2A-level surface reflectivity product in a sensory-2 remote sensing image of five growth periods of a research area, and the product is subjected to radiation correction, atmospheric correction and orthorectification.
And (3) utilizing a quality control waveband (QA 60) mark to realize cloud removal, and then selecting 10 spectral wavebands with spatial resolution of 10 meters and 20 meters of the sentinenl-2 remote sensing image. The 10 bands include red, green, blue, near infrared bands (band 4, band3, band2, band 8) with 10m spatial resolution, red edge 1, red edge 2, red edge 3, narrow near infrared, short wave infrared 1, short wave infrared 2 bands (band 5, band6, band7, band8A, band, band 12) with 20m spatial resolution.
And respectively carrying out median synthesis on 10 spectral bands of the remote sensing images in the 5 growth periods, and cutting to an interested area to obtain five first synthetic images with 10 spectral bands.
According to the system of the second aspect of the present invention, the third processing module 103 is specifically configured such that the specific method of calculating separability between different ground covers using the first composite image includes:
and calculating separability between different ground covers by using the first synthetic image through the J-M distance.
Specifically, five first synthetic images with 10 spectral bands are used for calculating the J-M distance between 5 soil covers in the GEE so as to compare and analyze the separability between winter wheat and other soil covers in different growth periods. Types of soil cover include winter wheat, buildings and roads, other vegetation (garlic, corn, vegetable fields, grass), forests and water. Each type takes 10% of the entire sample set (this step is implemented in ArcGIS) as samples for J-M distance calculation in an equal proportion random sampling. Under the assumption that the features accord with normality (if the features do not accord with a normal distribution rule, the feature separability is considered to be poor, and classification is not considered), the J-M distance calculation formula is as follows:
J＝2(1-e -B )
in the formula: b represents a bus distance, m i And m j Mean and variance of spectral reflectances of classes i and j, respectively. The value range of J is 0-2,0, which means that the two categories are almost inseparable, 2 means that the two categories can be completely separated, and the larger the J is, the better the separability between the ground covers is.
The separability of winter wheat, forest and water in each growth period of the research area is obtained through calculation; however, there is a large difference in separability between winter wheat and buildings and roads, other vegetation (corn, garlic, vegetable fields, grasslands). The separability between winter wheat and other cover types is highest at the heading stage, the J-M distance between the winter wheat and other covers is 1.98, and the separability is only 1.46 at the tillering stage of sowing. The results of the J-M distance indicate that the heading stage is the best stage to distinguish winter wheat from other cover types.
According to the system of the second aspect of the present invention, the fourth processing module 104 is specifically configured to, the specific method for generating the index feature by applying the image set includes:
and selecting corresponding bands from the 10 spectral bands to calculate and obtain commonly used 8 exponential features: a normalized vegetation index, an enhanced vegetation index, a soil conditioning vegetation index, a normalized water body index, a normalized difference water body index, a normalized construction index, a red edge normalized index, and a red edge position index;
the specific method for obtaining the second synthetic image by fusing the image set, the index features and the terrain features comprises the following steps:
adding the 8 indexes serving as independent wave bands to each image of 10 spectrum wave bands respectively, and then performing median synthesis on each added image to obtain a first process synthetic image with 18 wave bands;
and adding the altitude, the slope, the sloping direction and4 topographic features of mountain shadow as independent wave bands into the first process synthetic image to generate a second synthetic image with 22 wave bands.
The specific method for obtaining a third synthetic image by applying the second synthetic image to calculate texture features and finally fusing the texture features with the second synthetic image comprises the following steps:
calculating by using the second synthetic image to obtain 6 texture features of angular second moment, contrast, correlation, variance, inverse difference moment and entropy, and adding the 6 texture features as independent wave bands to the second synthetic image to generate a second process synthetic image containing 28 features;
and resampling the second process synthetic image to 10m spatial resolution to obtain the third synthetic image.
According to the system of the second aspect of the present invention, the fifth processing module 105 is specifically configured to, where the random forest is a classifier comprising a plurality of decision trees, randomly extract two thirds of data from the training set to create the training set by using a bootstrap step-by-step guided sampling strategy, and establish a decision tree for each training set, where one third of the data is not extracted every time sampling is performed, and perform an unbiased estimation of internal errors by using the part of data, so as to generate out-of-bag errors for evaluating the classification accuracy of the random forest.
The random forest implementation process comprises the following steps:
first, a step-by-step guided sampling technique is used to extract N training sets from the training set, wherein the size of each training set is about 2/3 of the training set.
Establishing a regression tree for each training set respectively to generate a forest consisting of N decision trees; in the growth process of each tree, M characteristic variables (M is less than or equal to M) are randomly selected from all M characteristic variables, and the optimal attribute is selected from the M attributes according to the minimum principle of the Gini index to carry out internal node branching, so that each tree grows fully, and pruning operation is not usually carried out.
And finally, collecting the prediction results of the N decision trees, and determining the category of the new sample by adopting a voting mode.
About 1/3 of the data is not decimated per sample, this portion of the data is commonly referred to as out-of-bag data (OOB); an internal error estimate is made using the out-of-bag data, resulting in OOB errors, which are used to predict the accuracy of the classification.
A random forest classification model is constructed by calling a (ee.classifier.smiledrandomforest) function inside the GEE. And sending the training set and the third synthetic image into a random forest model for training. The following parameters are mainly set: the number of decision trees in the forest. The number of decision trees is set to 10 to 200 in the GEE, respectively, increased by 10 each time. And then respectively sending the training set and the third synthetic image into a random forest model for training, and verifying by using the test set to obtain the precision under the quantity of the decision trees. The number of compared decision trees is set to 200 for best results.
Other parameters are kept as defaults.
According to the system of the second aspect of the present invention, the sixth processing module 106 is specifically configured such that the specific method for applying the test set to calculate a confusion matrix to evaluate the accuracy of the classification result of winter wheat at each stage of the growing season includes:
selecting four evaluation indexes of user precision, drawing precision, overall precision and Kappa coefficient to comprehensively reflect the recognition condition of the random forest winter wheat, wherein the collective calculation formula is as follows:
wherein, the first and the second end of the pipe are connected with each other,
UA i : user precision;
PA i : drawing precision;
OA: the overall accuracy;
kappa: a Kappa coefficient;
n: testing the total number of the concentrated samples;
m: the number of categories;
n i : predicting the number of the categories i in the categories i;
N i : predicting the total number of the categories i;
M i : the total number of classes i in the test set.
According to the system of the second aspect of the present invention, the seventh processing module 107 is specifically configured to, by comparing the remote sensing identification conditions of the winter wheat in each growth period, find that the classification precision of the heading period of the jointing is the highest and is consistent with the result obtained by the J-M distance. And then selecting the remote sensing image with the highest precision in the jointing and heading period to generate a winter wheat spatial distribution map.
The winter wheat area was then counted in GEE using the ee. The areas of the winter wheat of each pixel element in the period are summed to calculate the extraction area of the winter wheat of the whole research area. The winter wheat seeding area precision is the ratio of the estimated extraction area of a research area to the ground true value, and the precision evaluation is carried out on the extraction area by combining the official agricultural statistics yearbook and adopting the following formula:
in the formula, P represents the area extraction precision, S represents the winter wheat planting area extracted by a random forest method, and S' represents the real winter wheat planting area on the ground.
The extraction areas of the winter wheat in the jointing and heading period research area by the random forest method are 979.67 kilo hectares respectively, and the area extraction precision is 96.72 percent.
A third aspect of the invention discloses an electronic device. The electronic equipment comprises a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps of the method for remote sensing, identifying and analyzing the winter wheat based on the random forest in the first aspect of the disclosure are realized.
Fig. 4 is a block diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 4, the electronic device includes a processor, a memory, a communication interface, a display screen, and an input device, which are connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic equipment comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program to run on the non-volatile storage medium. The communication interface of the electronic device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, near Field Communication (NFC) or other technologies. The display screen of the electronic equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the electronic equipment, an external keyboard, a touch pad or a mouse and the like.
It will be understood by those skilled in the art that the structure shown in fig. 4 is only a partial block diagram related to the technical solution of the present disclosure, and does not constitute a limitation of the electronic device to which the solution of the present application is applied, and a specific electronic device may include more or less components than those shown in the drawings, or combine some components, or have a different arrangement of components.
A fourth aspect of the invention discloses a computer-readable storage medium. The computer readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the method for identifying and analyzing winter wheat remote sensing based on random forests in any one of the first aspect of the disclosure are realized.
Note that, the technical features of the above embodiments may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description in the present specification. The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent application shall be subject to the appended claims.
Claims (10)
1. A winter wheat remote sensing identification analysis method based on random forests is characterized by comprising the following steps:
step S1, sample data preparation: collecting actual measurement sample data of winter wheat to obtain a training set and a test set;
s2, extracting and preprocessing a synthetic image: acquiring a 2A-level earth surface reflectivity product in a sentinel-2 remote sensing image, preprocessing the 2A-level earth surface reflectivity product to obtain an image set, and synthesizing the image set to obtain a first synthesized image;
step S3, separability calculation: calculating separability between different covers by using the first synthetic image, and evaluating the separability between the winter wheat at each growth period stage and the different covers;
s4, classification characteristic construction: applying the image set to generate index features, then fusing the image set, the index features and topographic features to obtain a second synthetic image, applying the second synthetic image to calculate texture features, and finally fusing the texture features and the second synthetic image to obtain a third synthetic image;
s5, constructing and training a random forest classification model: applying the training set and the third synthetic image as random forest classification model input to train the random forest classification model;
s6, precision evaluation of classification results: calculating a confusion matrix by using the test set to evaluate the precision of the classification result of the winter wheat at each growth period;
s7, spatial mapping and area extraction of winter wheat: selecting a remote sensing image in a growth period stage with highest separability and classification result precision to generate a winter wheat space distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period with highest separability and classification result precision to obtain an extraction area, and performing precision evaluation on the extraction area by using the extraction area and the ground true value.
2. The remote sensing recognition analysis method for winter wheat based on random forest as claimed in claim 1, wherein in the step S1, the method for preparing sample data further comprises:
collecting time data of winter wheat growth and development stages of an agricultural meteorological site, dividing each growth period of the winter wheat, and dividing the growth period of the winter wheat into 5 growth period stages.
3. The method for remote sensing, recognizing and analyzing winter wheat based on the random forest as claimed in claim 2, wherein in the step S2, the specific method for obtaining the 2A-level surface reflectivity product in the sensory-2 remote sensing image comprises:
obtaining 2A-level surface reflectivity products in sensory-2 remote sensing images of 5 growth stages in a research area;
the specific method for preprocessing the 2A-level surface reflectivity product to obtain the image set comprises the following steps:
carrying out cloud removal on the remote sensing image by using a quality control band mark, and then selecting an image set of 10 spectral bands with spatial resolution of 10 meters and 20 meters of the remote sensing image;
the specific method for synthesizing the image set to obtain the first synthesized image includes:
and respectively carrying out median synthesis on 10 spectral bands of the remote sensing images in the 5 growth periods, and cutting to an interested area to obtain five first synthetic images with 10 spectral bands.
4. The method for remote sensing, recognizing and analyzing winter wheat based on random forest as claimed in claim 1, wherein in the step S3, the specific method for calculating separability between different ground covers by using the first synthetic image comprises:
and calculating separability between different ground covers by using the first synthetic image through the J-M distance.
5. The remote sensing recognition analysis method for winter wheat based on random forest as claimed in claim 3, wherein in said step S4, said specific method for generating index features by using said image set comprises:
selecting corresponding wave bands from the 10 spectrum wave bands to calculate and obtain commonly used 8 exponential characteristics: a normalized vegetation index, an enhanced vegetation index, a soil conditioning vegetation index, a normalized water body index, a normalized difference water body index, a normalized construction index, a red edge normalized index, and a red edge position index;
the specific method for obtaining the second synthetic image by fusing the image set, the index features and the terrain features comprises the following steps:
adding the 8 exponential features as independent wave bands to each image of 10 spectral wave bands respectively, and then performing median synthesis on each added image to obtain a first process synthetic image with 18 wave bands;
the altitude, the gradient, the slope direction and the 4 topographic features of the mountain shadow are taken as independent wave bands and added into the first process synthetic image, and a second synthetic image with 22 wave bands is generated.
6. The remote sensing recognition analysis method for winter wheat based on random forests as claimed in claim 5, wherein in the step S4, the specific method for calculating the texture features by using the second synthetic image and finally fusing the texture features and the second synthetic image to obtain a third synthetic image comprises:
calculating by using the second synthetic image to obtain 6 texture features of angular second moment, contrast, correlation, variance, inverse difference moment and entropy, and adding the 6 texture features as independent wave bands to the second synthetic image to generate a second process synthetic image containing 28 features;
and resampling the second process synthetic image to 10m spatial resolution to obtain the third synthetic image.
7. The remote sensing recognition analysis method for winter wheat based on random forests as claimed in claim 1, wherein in the step S6, the specific method for evaluating the classification result precision comprises the following steps:
selecting user precision, drawing precision and overall precision, and applying the test set to calculate a confusion matrix to evaluate four evaluation indexes of classification result precision Kappa coefficient of winter wheat at each growth period to comprehensively reflect the recognition condition of random forest winter wheat, wherein the collective calculation formula is as follows:
wherein the content of the first and second substances,
UA i : user precision;
PA i : drawing precision;
OA: the overall accuracy;
kappa: a Kappa coefficient;
n: testing the total number of the concentrated samples;
m: the number of categories;
n i : predicting the number of the categories i in the categories i;
N i : predicting the total number of the categories i;
M i : the total number of categories i in the test set.
8. A remote sensing, recognizing and analyzing system for winter wheat based on random forests, characterized in that the system comprises:
the first processing module is configured to collect the actual measurement sample data of the winter wheat to obtain a training set and a test set;
the second processing module is configured to obtain a 2A-level ground surface reflectivity product in a sensory-2 remote sensing image, preprocess the 2A-level ground surface reflectivity product to obtain an image set, and synthesize the image set to obtain a first synthesized image;
a third processing module configured to calculate separability between different covers using the first composite image, and evaluate separability between the winter wheat at each growth stage and the different covers;
a fourth processing module, configured to generate an index feature by applying the image set, then fuse the image set, the index feature and a topographic feature to obtain a second synthetic image, then calculate a texture feature by applying the second synthetic image, and finally fuse the texture feature with the second synthetic image to obtain a third synthetic image;
a fifth processing module configured to apply the training set and the third synthetic image as random forest classification model inputs, training the random forest classification model;
a sixth processing module configured to apply the test set to calculate a confusion matrix to evaluate the accuracy of the classification result of winter wheat at each stage of the growing period;
the seventh processing module is configured to select a remote sensing image in a growth period stage with the highest separability and classification result precision to generate a winter wheat spatial distribution map according to the results of the separability and the classification result precision; and counting the area of the winter wheat in the growth period stage with the highest separability and classification result precision to obtain an extraction area, and applying the extraction area and the ground true value to perform precision evaluation on the extraction area.
9. An electronic device, characterized in that the electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the remote sensing recognition analysis method for winter wheat based on random forests as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a computer program which, when being executed by a processor, carries out the steps of a method for remote sensing, identification and analysis of winter wheat based on random forests as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210116672.3A CN114494882B (en) | 2022-02-07 | 2022-02-07 | Winter wheat remote sensing identification analysis method and system based on random forest |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210116672.3A CN114494882B (en) | 2022-02-07 | 2022-02-07 | Winter wheat remote sensing identification analysis method and system based on random forest |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114494882A CN114494882A (en) | 2022-05-13 |
| CN114494882B true CN114494882B (en) | 2023-04-18 |
Family
ID=81479428
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210116672.3A Active CN114494882B (en) | 2022-02-07 | 2022-02-07 | Winter wheat remote sensing identification analysis method and system based on random forest |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114494882B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079672B (en) * | 2019-12-20 | 2023-10-13 | 中国科学院空天信息创新研究院 | Grape classification extraction method based on maximum entropy characteristics |
| CN115690585B (en) * | 2022-11-11 | 2023-06-06 | 中国科学院空天信息创新研究院 | Method and system for extracting wheat tillering number based on digital photo |
| CN117649598A (en) * | 2023-10-23 | 2024-03-05 | 广东省国土资源测绘院 | Offshore culture space distribution information monitoring method and system based on SAR (synthetic aperture radar) images |
| CN117853947B (en) * | 2024-03-06 | 2024-05-10 | 山东同圆数字科技有限公司 | Winter wheat remote sensing image automatic analysis system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109871884A (en) * | 2019-01-25 | 2019-06-11 | 曲阜师范大学 | A kind of support vector machines object-oriented Remote Image Classification merging multiple features |
| CN113033670A (en) * | 2021-03-29 | 2021-06-25 | 华南农业大学 | Method for extracting rice planting area based on Sentinel-2A/B data |
| CN113963222A (en) * | 2021-10-28 | 2022-01-21 | 中国电子科技集团公司第五十四研究所 | High-resolution remote sensing image change detection method based on multi-strategy combination |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN109919083B (en) * | 2019-03-06 | 2020-11-17 | 武汉大学 | Early automatic winter wheat mapping method based on Sentinel-2 image data |
| CN110321861A (en) * | 2019-07-09 | 2019-10-11 | 中国水利水电科学研究院 | A kind of main crops production moon scale Dynamic Extraction method |
| CN112395914A (en) * | 2019-08-15 | 2021-02-23 | 中国科学院遥感与数字地球研究所 | Method for identifying land parcel crops by fusing remote sensing image time sequence and textural features |
| CN111950530B (en) * | 2020-09-08 | 2024-04-12 | 中国水利水电科学研究院 | Multi-feature optimization and fusion method for crop planting structure extraction |
-
2022
- 2022-02-07 CN CN202210116672.3A patent/CN114494882B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109871884A (en) * | 2019-01-25 | 2019-06-11 | 曲阜师范大学 | A kind of support vector machines object-oriented Remote Image Classification merging multiple features |
| CN113033670A (en) * | 2021-03-29 | 2021-06-25 | 华南农业大学 | Method for extracting rice planting area based on Sentinel-2A/B data |
| CN113963222A (en) * | 2021-10-28 | 2022-01-21 | 中国电子科技集团公司第五十四研究所 | High-resolution remote sensing image change detection method based on multi-strategy combination |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114494882A (en) | 2022-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114494882B (en) | Winter wheat remote sensing identification analysis method and system based on random forest | |
| Blickensdörfer et al. | Mapping of crop types and crop sequences with combined time series of Sentinel-1, Sentinel-2 and Landsat 8 data for Germany | |
| Han et al. | Modeling maize above-ground biomass based on machine learning approaches using UAV remote-sensing data | |
| Hlatshwayo et al. | Mapping forest aboveground biomass in the reforested Buffelsdraai landfill site using texture combinations computed from SPOT-6 pan-sharpened imagery | |
| CN102033230B (en) | Grassland satellite remote sensing monitoring system and method | |
| Sun et al. | Monitoring maize lodging grades via unmanned aerial vehicle multispectral image | |
| Popescu et al. | Remote sensing in the analysis and characterization of spatial variability of the territory. a study case in Timis County, Romania. | |
| Zhang et al. | Estimating wheat yield by integrating the WheatGrow and PROSAIL models | |
| Adjorlolo et al. | Estimation of canopy nitrogen concentration across C3 and C4 grasslands using WorldView-2 multispectral data | |
| Abdolalizadeh et al. | Rangeland canopy cover estimation using Landsat OLI data and vegetation indices in Sabalan rangelands, Iran | |
| Kotsuki et al. | SACRA–a method for the estimation of global high-resolution crop calendars from a satellite-sensed NDVI | |
| CN114882359B (en) | Soybean planting area extraction method and system based on vegetation index time series spectrum characteristics | |
| Pereira et al. | Nitrogen variability assessment of pasture fields under an integrated crop-livestock system using UAV, PlanetScope, and Sentinel-2 data | |
| Huang et al. | Assimilating remotely sensed information with the wheatgrow model based on the ensemble square root filter for improving regional wheat yield forecasts | |
| 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 | |
| Choudhary et al. | Random Forest for rice yield mapping and prediction using Sentinel-2 data with Google Earth Engine | |
| Panda et al. | Distinguishing blueberry bushes from mixed vegetation land use using high resolution satellite imagery and geospatial techniques | |
| Shu et al. | Assessing maize lodging severity using multitemporal UAV-based digital images | |
| CN105930863A (en) | Determination method for spectral band setting of satellite camera | |
| CN114494909A (en) | Method and system for generating spatial distribution diagram of soybean growing season | |
| Makhamreh | Derivation of vegetation density and land-use type pattern in mountain regions of Jordan using multi-seasonal SPOT images | |
| CN116579521A (en) | Yield prediction time window determining method, device, equipment and readable storage medium | |
| CN110245694A (en) | A kind of the Northwest's agricultural film information analysis method based on high score Remote Sensing Image Segmentation | |
| CN115128013A (en) | Soil organic matter content space prediction evaluation method based on partition algorithm | |
| Wilkinson et al. | Change detection techniques for use in a statewide forest inventory program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |