CN112861435A - Mangrove forest quality remote sensing retrieval method and intelligent terminal - Google Patents
Mangrove forest quality remote sensing retrieval method and intelligent terminal Download PDFInfo
- Publication number
- CN112861435A CN112861435A CN202110178986.1A CN202110178986A CN112861435A CN 112861435 A CN112861435 A CN 112861435A CN 202110178986 A CN202110178986 A CN 202110178986A CN 112861435 A CN112861435 A CN 112861435A
- Authority
- CN
- China
- Prior art keywords
- data
- remote sensing
- inversion
- quality
- model
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- 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
- Y02A20/00—Water conservation; Efficient water supply; Efficient water use
- Y02A20/20—Controlling water pollution; Waste water treatment
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a mangrove forest quality remote sensing retrieval method and an intelligent terminal, wherein the method comprises the following steps: acquiring target remote sensing data of a region to be monitored and monitoring data corresponding to a plurality of groups of monitoring samples, wherein each monitoring data comprises environmental data and quality data; establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environment data, wherein the inversion model group comprises a plurality of inversion models, and the inversion model group comprises a remote sensing inversion model and a collaborative inversion model; calculating the precision of each inversion model, and selecting the inversion model with the highest precision as a target model; and drawing a mangrove forest quality map of the area to be monitored based on the target model. The invention can improve the quality monitoring precision of mangrove forest region.
Description
Technical Field
The invention relates to the field of plant monitoring, in particular to a mangrove forest quality remote sensing retrieval method and an intelligent terminal.
Background
The mangrove forest is woody plant community in the intertidal zone of tropical and subtropical beach, and has the functions of promoting silt to make land, protecting dike, maintaining biological diversity and global carbon balance, etc. In recent decades, the area of mangrove forest has decreased dramatically, subject to global climate change and human activity. The sharp reduction of mangrove forest area not only affects the material exchange between sea and land ecosystems and the global atmosphere composition, but also weakens the resistance of land to the rise of sea level. Therefore, protection, monitoring and management of mangrove forest are attracting attention from many parties. Wherein, the monitoring of the mangrove forest quality directly influences the deployment and the expansion of the subsequent mangrove forest protection work and the management work.
At present, remote sensing monitoring research of mangrove forest mainly takes wetland dynamics and driving force, community distribution and biophysical parameters (such as leaf area index, biomass, average crown width and tree height) monitoring as main points. With the development of the technology for quantitatively inverting the quality of crops, land forests and grass plants by using the remote sensing technology, some scholars also monitor the quality of mangroves by using the remote sensing technology. However, these models are poorly interpreted and difficult to reflect sensitive variable information that affects plant quality variation. The main body is as follows:
(1) most of the current inversion monitors the current quality based on a certain period, and the environment of the default mangrove forest is not changed at each stage. In fact, although species and season are important factors influencing the remote sensing inversion accuracy of the mangrove plant quality, environmental factors also have important influence on the accuracy of the inversion result.
(2) The hyperspectral image has rich wave band information, but the spatial resolution is low, so that accurate spatial mapping of the mangrove plant quality in a small area is difficult to realize. Mangrove forest grows in intertidal zone, the difficulty of obtaining high spectral data of field observation canopy is high, most researches utilize indoor leaf high spectral data to invert mangrove plant quality, but landscape scale inversion is affected more complexly, so that the indoor leaf scale inversion model cannot be popularized to landscape scale.
(3) The inversion method is mainly based on an empirical model, the inversion mechanism research is weak, the existing methods are all empirical models based on statistical regression or machine learning, the superiority and inferiority of each model are not sufficiently researched, the influence of external environment factors on the models is not considered, the inversion mechanism research is lacked, and the model applicability is poor.
Therefore, the current method for inverting the mangrove forest based on the remote sensing data still has room for improvement in inversion accuracy.
Disclosure of Invention
The invention mainly aims to provide a mangrove forest quality remote sensing retrieval method and an intelligent terminal, and aims to solve the problem that mangrove forest monitoring precision is low in the prior art.
In order to achieve the aim, the invention provides a mangrove forest quality remote sensing inversion method, which comprises the following steps:
acquiring target remote sensing data of a region to be monitored and monitoring data corresponding to a plurality of groups of monitoring samples, wherein each monitoring data comprises environmental data and quality data;
establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environment data, wherein the inversion model group comprises a plurality of inversion models, and the inversion model group comprises a remote sensing inversion model and a collaborative inversion model;
calculating the precision of each inversion model, and selecting the inversion model with the highest precision as a target model;
and drawing a mangrove forest quality map of the area to be monitored based on the target model.
Optionally, in the remote sensing inversion method for mangrove forest quality, for each group of the monitoring samples, the environmental data corresponding to the monitoring samples includes monitoring values corresponding to a plurality of environmental factors, where the environmental factors include seawater salinity, seawater pollution, soil salinity, soil nutrition, and/or topography of the area to be monitored.
Optionally, in the mangrove forest quality remote sensing inversion method, for each group of the monitoring samples, the quality data corresponding to the monitoring samples includes leaf nitrogen content and leaf phosphorus content.
Optionally, the mangrove forest quality remote sensing inversion method, wherein when the inversion model is a remote sensing inversion model, establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environmental data specifically includes:
constructing a vegetation index based on the target remote sensing data;
and regressing the quality data according to the band reflectivity corresponding to each band in the target remote sensing data and the vegetation index to generate the remote sensing inversion model.
Optionally, the mangrove forest quality remote sensing inversion method, wherein when the inversion model is a collaborative inversion model, establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environmental data specifically includes:
constructing a vegetation index based on the target remote sensing data;
and regressing the quality data according to the waveband reflectivity, the vegetation index and the environment data to generate the collaborative inversion model.
Optionally, in the mangrove forest quality remote sensing inversion method, the regression mode of the quality data is random forest regression.
Optionally, the mangrove forest quality remote sensing inversion method, wherein the calculating the precision of each inversion model and selecting the inversion model with the highest precision as the target model specifically includes:
randomly selecting a plurality of samples in the monitoring samples as verification samples;
calculating a corresponding predicted value in each verification sample according to each inversion model;
determining the corresponding precision of the inversion model according to the predicted value and the quality data corresponding to the predicted value;
and selecting the inversion model with the highest precision as the target model.
Optionally, the mangrove forest quality remote sensing inversion method, wherein after calculating the precision of each inversion model and selecting the inversion model with the highest precision as the target model, further comprises:
for each environmental factor, dividing the monitoring data into a plurality of classification groups according to the size of the monitoring value corresponding to the environmental factor;
aiming at each classification group, establishing a grouping inversion model corresponding to the classification group according to variable data, the waveband reflectivity and the vegetation index, wherein the variable data are inspection values corresponding to environmental factors except the environmental factors;
and calculating the precision of each grouped inversion model, and determining the corresponding generalization of the collaborative inversion model according to the precision.
In addition, to achieve the above object, the present invention further provides an intelligent terminal, wherein the intelligent terminal includes: the mangrove forest quality remote sensing inversion method comprises a memory, a processor and a mangrove forest quality remote sensing inversion program which is stored on the memory and can run on the processor, wherein the mangrove forest quality remote sensing inversion program realizes the steps of the mangrove forest quality remote sensing inversion method when being executed by the processor.
In addition, in order to achieve the above object, the present invention further provides a computer readable storage medium, wherein the computer readable storage medium stores a mangrove forest quality remote sensing inversion program, and when the mangrove forest quality remote sensing inversion program is executed by a processor, the steps of the mangrove forest quality remote sensing inversion method are implemented.
The invention discloses a mangrove forest quality remote sensing inversion method, an intelligent terminal and a computer readable storage medium, wherein environmental factors such as seawater salinity, seawater chemical oxygen demand, soil salinity, soil organic matters, elevation and gradient are important factors influencing mangrove plant quality space variation. And selecting a model with the highest precision from the inversion models to invert the mangrove forest quality of the area to be monitored so as to obtain a mangrove forest quality map of the area. In addition, when an inversion model is established, a random forest machine learning method with easy interpretability and high precision of statistical regression is utilized to realize spatial mapping of the mangrove plant quality in a regional scale, and a theoretical and method basis is provided for accurate inversion of the mangrove plant quality.
Drawings
FIG. 1 is a flow chart of a preferred embodiment provided by the remote sensing inversion method for mangrove forest quality of the present invention;
FIG. 2 is an overall flow chart of a preferred embodiment provided by the mangrove forest quality remote sensing inversion method of the present invention;
fig. 3 is a schematic operating environment diagram of an intelligent terminal according to a preferred embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
According to the mangrove forest quality remote sensing inversion method, the mangrove forest quality remote sensing inversion method can be executed through an intelligent terminal, and the intelligent terminal comprises terminals such as a smart phone and a computer. In this embodiment, a computer is taken as an example to describe a mangrove forest quality remote sensing inversion process, and a certain software or a certain plug-in the computer may be used to specifically execute the process. As shown in fig. 1 and fig. 2, the mangrove forest quality remote sensing inversion method comprises the following steps:
and S100, acquiring target remote sensing data of a region to be monitored and monitoring data corresponding to a plurality of groups of monitoring samples.
Specifically, a monitoring sample of an area to be monitored is collected in advance. In this embodiment, the monitoring samples include a community sample of the mangrove forest and an environment sample, the community sample is a sample indicating characteristics of the growth state of the mangrove forest, and the environment sample is a sample of the environment where the mangrove forest is located and related to the growth of the mangrove forest. For each group of monitoring samples, the environmental data corresponding to the monitoring samples comprises monitoring numerical values corresponding to a plurality of environmental factors.
Each monitoring data includes environmental data and quality data. The environment data is mainly data obtained from an environment sample, and the quality data is mainly data obtained from a community sample.
The quality data mainly refers to the quality parameters for evaluating the mangrove forest, and can be determined by the ecological chemometrics characteristics, plant height data, breast height data and the like. Wherein the ecological chemometrics characteristics include the content of chemical components such as carbon, nitrogen, phosphorus and the like in the mangrove forest rhizome leaves. Because of the diversity of the species of the plants in the mangrove forest community, the collection of representative species can be adopted as the sampling object. The sampling species used for collecting the samples in this example were the communities Kandelia candel, Aegiceras corniculata and Avicennia marina, and the ecological stoichiometric characteristics were monitored by the nitrogen and phosphorus content in the leaves of these species. Thus, the mass data in each set of monitored samples included leaf nitrogen content and leaf phosphorus content. In addition, data such as the chlorophyll content in the leaf can be used as the quality data.
The environmental samples in this embodiment include seawater samples and soil samples, and the mangrove forest is suitable for growing in the impact saline soil or saline sand soil, so the salt content in the environment can also be used as an index of the growth state of the mangrove forest, and the seawater samples and the soil samples are collected. In this embodiment, the monitoring value for representing the environmental pollution condition is the chemical oxygen demand in the seawater sample. Chemical oxygen demand is a rapid and important parameter for organic contamination in seawater. The monitored value for characterizing soil nutrition is organic matter content. Organic matter content is an important indicator of soil fertility. In addition, other indicators that may be used to characterize environmental pollution and soil nutrition may be substituted for use in this embodiment. Thus, the environmental factors in this example include seawater salinity, seawater pollution, soil salinity, and soil nutrition. In addition, in practical application, the environmental factors can be deleted to achieve better effect.
In addition, the terrain plays an important role in the development condition of a community, and the elevation, the gradient, the altitude and the like of the community can influence the development of the community. Therefore, in this embodiment, the environmental factor further includes a terrain of the area to be monitored, and specifically may include an elevation, a slope, and the like. The terrain data may be obtained from a terrain Model such as a Digital Elevation Model (DEM).
When community samples and environment samples are collected, 75 3x 3-meter samples are randomly arranged in a sampling area for any sampling species, and the distance between any two samples is larger than 30 meters, so that the cross interference of data is reduced. Finally, 75 monitoring samples are collected, wherein each monitoring sample comprises a leaf sample, a seawater sample and a soil sample.
In this embodiment, the nitrogen and phosphorus content in each leaf sample in each group of monitoring samples is analyzed to obtain the quality data in the group of monitoring samples. For each set of seawater samples, salinity and chemical oxygen demand in the seawater samples were measured by chemical instruments. For the soil samples in each set of monitoring samples, salinity and organic matter content in the soil samples are monitored.
And sorting the obtained data to obtain monitoring data corresponding to a plurality of groups of monitoring samples. The monitoring data are input into a computer, and the computer acquires the monitoring data through reading or scanning.
Meanwhile, remote sensing data of the area to be monitored are obtained. The optical remote sensing records the reflection spectrum information of the ground object. The reflection spectrum curve of the vegetation has obvious characteristics from visible light to infrared bands, reflects the chlorophyll content and the growth condition of the vegetation, and the chlorophyll content is related to the biomass of the leaves, which is related to the biomass of the community. Therefore, the living biomass of the vegetation can be estimated by using the remote sensing data according to the characteristics of the reflection spectrum of the vegetation. Structural information of the forest canopy surface is intelligently monitored through optical remote sensing carding, the forest canopy surface does not have penetrability, cannot monitor the texture part occupying a vegetation biomass main body, and is easily influenced by weather. Therefore, after the original remote sensing data is obtained, certain preprocessing needs to be performed on the original remote sensing data to obtain subsequently usable target remote sensing data. The preprocessing in this embodiment includes radiometric calibration, atmospheric correction, and/or resampling to eliminate the effects of the atmosphere and terrain on the image and obtain the true reflectivity of the earth's surface. The remote sensing data collected by the embodiment is remote sensing data collected based on a satellite WorldView-3, and WorldView-2 data and the like can be selected optical remote sensing data. Further, to maintain consistency with the collection of the monitoring samples, the collected telemetry data is data from the collection day, i.e., the date the monitoring samples were collected, and some time before and some time after the collection.
And S200, establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environment data.
Specifically, on the basis of remote sensing data, on the inversion model level, three types of models are currently used for remote sensing inversion of mangrove plant quality, namely a vegetation index model, a multivariate statistical regression model and a machine learning model. The Vegetation Index model is a common model for quantitative inversion of mangrove plant quality, such as Normalized Difference Vegetation Index (NDVI), rededgeposition Index (REPI), and Soil adjusted Vegetation Index (msaiv), and has a simple form and easy explanation. On the basis of the former, in order to fully utilize hundreds of high spectral band information, Stepwise Multiple Linear Regression (SMLR) and Partial Least Squares Regression (PLSR) are used to invert the quality of mangrove plants. In addition, Artificial Neural Network (ANN) and Support Vector Regression (SVR) are common machine learning models that have been used to establish a nonlinear relationship between mangrove plant quality and hyperspectral reflectance.
On the remote sensing scale level, the current scholars perform remote sensing inversion on the quality of mangrove plants from the leaf scale and the landscape scale, and the remote sensing inversion method specifically comprises the following steps:
(1) the blade size is as follows: the hyperspectral reflectivity of fresh leaves and the content of leaf components of the mangrove plants are usually measured under constant laboratory conditions, the method is mainly used for researching the theoretical possibility of remote sensing inversion of the quality of the mangrove plants, and is the basis of research on canopy and landscape scale;
(2) landscape scale: at present, three multispectral satellite images, such as remote sensing images acquired by satellites Landsat5, Landsat 8 and Worldview-2, are used for quantitative inversion and mapping of regional mangrove plant quality.
However, the correlation of the models to the variables is poor in interpretability, and sensitive waveband information and sources influencing the plant quality are difficult to reflect. On the basis, the vegetation index is constructed according to the target remote sensing data. The Vegetation Index (VI) is a numerical value obtained by performing linear operation and nonlinear operation on spectral data acquired by a remote sensor, and has a certain referential property for covering a Vegetation cover. In this embodiment, the constructed vegetation index may be any one or more of normalized vegetation index, optimized soil conditioning vegetation index, differential ratio vegetation index, and the like. The reflectivity curve form of the same object reflects different reflectivity of different wave bands, so that the reflectivity of different wave bands is compared with the quality data, the rule between the numerical values of the remote sensing image data and the quality data can be reflected, and the inversion is carried out on the quality distribution of the mangrove forest.
Therefore, in this embodiment, an inversion model based on remote sensing data can be obtained by performing regression analysis according to the band reflectivity and the vegetation index corresponding to each band.
Because the environmental factors also have certain influence on the quality distribution of the mangrove forest, the quality data is regressed by combining the wave band reflectivity, the vegetation index and the environmental data, so that a collaborative inversion model of the collaborative environmental factors and the remote sensing data is obtained.
In addition, since there are a plurality of environmental factors in this embodiment, the collaborative inversion model may be an inversion model constructed by combining all the environmental factors, or may be an inversion model constructed based on a part of the environmental factors.
Generally, when an inversion model is built, due to the correlation and diversity among data, the calculation amount of the data is reduced, and the dimension reduction processing is often performed on the data. For example, in remote sensing data, each waveband is highly correlated, and the displayed visual effects are similar. A commonly used dimension reduction method is Principal Component Analysis (PCA), which can remove redundant information between bands and compress multi-band image information to a few more effective converted bands than the original bands, thereby reducing data dimensions and data calculation amount. Similarly, there is a correlation between environmental factors, such as salinity in seawater and soil salinity. Therefore, in the first implementation manner of this embodiment, in the modeling process, the dimension reduction processing may be performed on the target remote sensing data and the environmental data to obtain the remote sensing characteristic value and the environmental characteristic value, and then the regression analysis may be performed on the quality data based on the remote sensing characteristic value, the plant index and/or the environmental characteristic value to obtain the remote sensing inversion model and/or the collaborative inversion model.
Although the data dimensionality reduction can reflect the trend of the data to a certain extent and reduce the data calculation amount, the data dimensionality reduction has certain data loss. In a second implementation manner of this embodiment, a random forest regression is used to realize regression on quality data. The random forest regression model is formed by presetting a plurality of decision trees, wherein each decision tree has no correlation, but the final decision result is output based on each decision tree. The randomness of the random forest is represented by the randomness of the samples used to train the decision tree, as well as the randomness of the features.
Aiming at each decision tree, an indefinite number of monitoring data are randomly selected from a plurality of groups of monitoring data in a replacement mode to serve as training data. And then, based on the training data, segmenting each node in the decision tree to obtain the trained decision tree. In the splitting process, a plurality of features are randomly selected, and then the optimal splitting point is selected from the selected features to realize the splitting of the left and right subtrees. All the characteristics in this embodiment are the respective bands, the vegetation index and the respective environmental factors. And obtaining a trained decision tree, namely an inversion model after the division is finished.
The remote sensing inversion model and the collaborative inversion model can be obtained by adopting a similar method, except that the selected characteristics have differences.
Therefore, an inversion model group comprising a plurality of inversion models is established in the mode, and the inversion model group at least comprises a remote sensing inversion model and a collaborative inversion model.
In addition, after the remote sensing inversion model and the collaborative inversion model are obtained, the variable projection importance in the remote sensing inversion model and the collaborative inversion model can be extracted to measure the importance of each characteristic, namely the importance of the environmental factor. If the regression method is a random forest regression method, the importance of the obtained variable projection can be obtained by adopting a Gini index method, a precision reduction method and the like.
And step S300, calculating the precision of each inversion model, and selecting the inversion model with the highest precision as a target model.
Specifically, after a plurality of inversion models are obtained, due to different scenarios, the accuracy of each model may be different. It is therefore necessary to determine the accuracy of each inverse model based on the current data.
The accuracy of the inversion model can be determined by Root Mean Square Error (RMSE) and coefficient of determination (R)2) And the like. The decision coefficient is mainly used for linear regression, and the larger the numerical value of the decision coefficient is, the stronger the correlation between the analog value and the actual value obtained based on the inversion model is, and the more accurate the model is. The smaller the RMSE between the simulated and actual values, the better the inverse model will perform. In addition, indexes used for evaluating the regression model, such as Mean Absolute Error (MAE), Mean Square Error (MSE), and the like, can also be used as the accuracy of the inversion model. The specific calculation process is as follows:
randomly selecting a plurality of samples in the monitoring samples as verification samples;
calculating a corresponding predicted value in each verification sample according to each inversion model;
determining the corresponding precision of the inversion model according to the predicted value and the quality data corresponding to the predicted value;
and selecting the inversion model with the highest precision as the target model.
Specifically, several of the monitoring samples are randomly selected as verification samples. Part of the monitoring samples can be selected as verification samples, and all of the monitoring samples can also be selected as verification samples. And then calculating corresponding predicted values in the verification samples according to each inversion model, such as the remote sensing inversion model in the foregoing. For example, for the verification sample A, the reflectivity of each wave band and the corresponding vegetation index, which are obtained by calculating the target remote sensing data corresponding to the verification sample A, are input into the obtained inversion model, and the inversion model can calculate the predicted value corresponding to the verification sample A. And then, according to the predicted value and the quality data of the verification sample A, calculating and evaluating relevant indexes of the regression model, such as the average variance, the absolute coefficient and the like, so as to obtain the corresponding precision of the inversion model. And finally, selecting the inversion model with the highest precision as the target model.
In addition, after the target model is obtained, the applicability of the model can be verified. The transitivity of the target model among different environmental factors is explored, and therefore the generalization of the target model is analyzed. The specific process is as follows:
for each environmental factor, dividing the quality data into a plurality of classification groups according to the magnitude of the monitoring value corresponding to the environmental factor;
aiming at each classification group, establishing a grouping inversion model corresponding to the classification group according to variable data, the waveband reflectivity and the vegetation index, wherein the variable data are inspection values corresponding to other environment factors except the environment factor;
and calculating the precision of each grouped inversion model, and determining the corresponding generalization of the collaborative inversion model according to the precision.
Specifically, for each environmental factor, the monitoring data is divided into a plurality of classification groups according to the size of the monitoring value corresponding to the environmental factor. For example, the seawater salinity is divided into a plurality of classification groups according to the numerical value of the seawater salinity in each monitoring data, for example, the monitoring data A, B and C, and is divided into two classification groups according to the seawater salinity corresponding to each quality data, wherein the first classification group is the monitoring data A and the monitoring data B, and the second classification group is the monitoring data C.
And then establishing a grouping inversion model corresponding to each classification group based on the wave band reflectivity and the vegetation index corresponding to each classification group and data corresponding to environmental factors except the environmental factor of seawater salinity. And then calculating the corresponding precision of each grouped inversion model. The method for creating the inversion model and calculating the accuracy is introduced above and will not be described herein. By grouping the accuracy of the inversion model, the generalization of remote sensing inversion can be determined. And calculating the similarity between the accuracies of the grouped inversion models, and if the similarity exceeds a preset similarity threshold, determining that the remote sensing inversion has strong popularization. The collaborative inversion model is divided into a plurality of classification groups aiming at the same environmental factor, such as seawater salinity, and if the precision of each classification model is relatively close to each other, the collaborative inversion model has stable prediction performance under the condition of different environmental factors, so that the collaborative inversion model has relatively strong popularization.
And S400, drawing a mangrove forest quality map of the area to be monitored based on the target model.
Specifically, after a target model is obtained, a mangrove forest quality map of the area to be monitored is drawn based on the target model. The mangrove quality map refers to the mangrove quality corresponding to each spatial position in the region to be monitored, and can also be regarded as a mangrove quality spatial distribution map. The mangrove forest quality map comprises the prediction quality values corresponding to all positions in the area to be monitored.
When the target model is a remote sensing inversion model, the remote sensing data of each position area is contained in the target remote sensing data, so that the corresponding waveband reflectivity and the vegetation index can be calculated and obtained based on any position coordinate, and the waveband reflectivity and the vegetation index corresponding to each position can be obtained based on the currently obtained target remote sensing data. And then, inputting the band reflectivity and the vegetation index into a target model, thereby outputting a prediction quality value corresponding to the position. And obtaining the mangrove forest quality map of the area to be monitored after obtaining the prediction quality values corresponding to all the positions.
When the target model is a collaborative inversion model, the amount of the environmental samples is limited, so that the data of seawater salinity, seawater pollution, soil salinity and soil pollution which depend on the actually acquired samples cannot be directly obtained. In this embodiment, to implement the drawing of the mangrove quality map, based on the currently obtained environmental data, a spatial interpolation method is adopted to obtain the environmental data corresponding to each position, and the adopted spatial interpolation method includes a kriging interpolation method and the like. The interpolation method is an algorithm for converting limited discrete data into a continuous data surface, so that the simulation of the environmental data of the whole area to be monitored is realized on the basis of the data of each current environmental sample. After the environmental data, the vegetation index and the waveband reflectivity corresponding to each position are obtained, the environmental data, the vegetation index and the waveband reflectivity are input into the collaborative inversion model, and the prediction quality value corresponding to the position is obtained.
Further, as shown in fig. 3, based on the remote sensing inversion method for mangrove forest quality, the invention further provides an intelligent terminal, and the intelligent terminal comprises a processor 10, a memory 20 and a display 30. Fig. 3 shows only some of the components of the smart terminal, but it should be understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead.
The memory 20 may be an internal storage unit of the intelligent terminal in some embodiments, such as a hard disk or a memory of the intelligent terminal. The memory 20 may also be an external storage device of the Smart terminal in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the Smart terminal. Further, the memory 20 may also include both an internal storage unit and an external storage device of the smart terminal. The memory 20 is used for storing application software installed in the intelligent terminal and various data, such as program codes of the installed intelligent terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 20 stores a mangrove forest quality remote sensing inversion program 40, and the mangrove forest quality remote sensing inversion program 40 can be executed by the processor 10, so as to realize the mangrove forest quality remote sensing inversion method in the application.
The processor 10 may be, in some embodiments, a Central Processing Unit (CPU), a microprocessor or other data Processing chip, and is configured to run program codes stored in the memory 20 or process data, such as executing the mangrove forest quality remote sensing inversion method.
The display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch panel, or the like in some embodiments. The display 30 is used for displaying information at the intelligent terminal and for displaying a visual user interface. The components 10-30 of the intelligent terminal communicate with each other via a system bus.
In one embodiment, the following steps are implemented when the processor 10 executes the mangrove forest quality remote sensing inversion program 40 in the memory 20:
acquiring target remote sensing data of a region to be monitored and monitoring data corresponding to a plurality of groups of monitoring samples, wherein each monitoring data comprises environmental data and quality data; establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environment data, wherein the inversion model group comprises a plurality of inversion models, and the inversion model group comprises a remote sensing inversion model and a collaborative inversion model; calculating the precision of each inversion model, and selecting the inversion model with the highest precision as a target model; and drawing a mangrove forest quality map of the area to be monitored based on the target model. The invention can improve the quality monitoring precision of mangrove forest region.
The invention also provides a computer readable storage medium, wherein the computer readable storage medium stores a mangrove forest quality remote sensing inversion program, and the mangrove forest quality remote sensing inversion program realizes the steps of the mangrove forest quality remote sensing inversion method when being executed by a processor.
Of course, it will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer readable storage medium, and when executed, the program can include the processes of the embodiments of the methods described above. The computer readable storage medium may be a memory, a magnetic disk, an optical disk, etc.
It is to be understood that the invention is not limited to the examples described above, but that modifications and variations may be effected thereto by those of ordinary skill in the art in light of the foregoing description, and that all such modifications and variations are intended to be within the scope of the invention as defined by the appended claims.
Claims (10)
1. A mangrove forest quality remote sensing inversion method is characterized by comprising the following steps:
acquiring target remote sensing data of a region to be monitored and monitoring data corresponding to a plurality of groups of monitoring samples, wherein each monitoring data comprises environmental data and quality data;
establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environment data, wherein the inversion model group comprises a plurality of inversion models, and the inversion model group comprises a remote sensing inversion model and a collaborative inversion model;
calculating the precision of each inversion model, and selecting the inversion model with the highest precision as a target model;
and drawing a mangrove forest quality map of the area to be monitored based on the target model.
2. The remote sensing inversion method for mangrove forest quality according to claim 1, characterized in that for each group of the monitoring samples, the environmental data corresponding to the monitoring samples comprises monitoring values corresponding to a plurality of environmental factors, wherein the environmental factors comprise seawater salinity, seawater pollution, soil salinity, soil nutrition and/or topography of the area to be monitored.
3. The remote sensing inversion method of mangrove forest quality according to claim 1, characterized in that for each group of said monitoring samples, the quality data corresponding to the monitoring samples comprise leaf nitrogen content and leaf phosphorus content.
4. The remote sensing inversion method of mangrove forest quality according to claim 2, wherein when the inversion model is a remote sensing inversion model, establishing an inversion model group corresponding to the quality data according to the target remote sensing data and the environmental data, specifically comprising:
constructing a vegetation index based on the target remote sensing data;
and regressing the quality data according to the band reflectivity corresponding to each band in the target remote sensing data and the vegetation index to generate the remote sensing inversion model.
5. The remote sensing inversion method of mangrove forest quality according to claim 4, wherein when the inversion model is a collaborative inversion model, the establishing of the inversion model group corresponding to the quality data according to the target remote sensing data and the environmental data specifically comprises:
constructing a vegetation index based on the target remote sensing data;
and regressing the quality data according to the waveband reflectivity, the vegetation index and the environment data to generate the collaborative inversion model.
6. The remote sensing inversion method of mangrove forest quality according to claim 5, characterized in that the regression of the quality data is random forest regression.
7. The remote sensing inversion method of mangrove forest quality according to claim 4, wherein the calculating the accuracy of each said inversion model, and selecting the inversion model with the highest accuracy as the target model specifically comprises:
randomly selecting a plurality of samples in the monitoring samples as verification samples;
calculating a corresponding predicted value in each verification sample according to each inversion model;
determining the corresponding precision of the inversion model according to the predicted value and the quality data corresponding to the predicted value;
and selecting the inversion model with the highest precision as the target model.
8. The remote sensing inversion method of mangrove forest quality according to claim 7, wherein after calculating the accuracy of each said inversion model and selecting the inversion model with the highest accuracy as the target model, further comprising:
for each environmental factor, dividing the monitoring data into a plurality of classification groups according to the size of the monitoring value corresponding to the environmental factor;
aiming at each classification group, establishing a grouping inversion model corresponding to the classification group according to variable data, the waveband reflectivity and the vegetation index, wherein the variable data are inspection values corresponding to environmental factors except the environmental factors;
and calculating the precision of each grouped inversion model, and determining the corresponding generalization of the collaborative inversion model according to the precision.
9. An intelligent terminal, characterized in that, intelligent terminal includes: a memory, a processor and a mangrove forest quality remote sensing inversion program stored on the memory and operable on the processor, the mangrove forest quality remote sensing inversion program when executed by the processor implementing the steps of the mangrove forest quality remote sensing inversion method according to any one of claims 1 to 8.
10. A computer readable storage medium, characterized in that the computer readable storage medium stores a mangrove forest quality remote sensing inversion program, which when executed by a processor implements the steps of the mangrove forest quality remote sensing inversion method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110178986.1A CN112861435B (en) | 2021-02-09 | 2021-02-09 | Mangrove quality remote sensing inversion method and intelligent terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110178986.1A CN112861435B (en) | 2021-02-09 | 2021-02-09 | Mangrove quality remote sensing inversion method and intelligent terminal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112861435A true CN112861435A (en) | 2021-05-28 |
| CN112861435B CN112861435B (en) | 2023-07-18 |
Family
ID=75988132
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110178986.1A Active CN112861435B (en) | 2021-02-09 | 2021-02-09 | Mangrove quality remote sensing inversion method and intelligent terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112861435B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113536937A (en) * | 2021-06-17 | 2021-10-22 | 海南省林业科学研究院(海南省红树林研究院) | Mangrove forest ecological quantitative inversion method based on remote sensing technology |
| CN113552079A (en) * | 2021-06-17 | 2021-10-26 | 海南省林业科学研究院(海南省红树林研究院) | Mangrove forest quantitative inversion system based on aviation hyperspectral data |
| CN114924034A (en) * | 2022-05-06 | 2022-08-19 | 贵州师范大学 | Forestry carbon measurement system based on ecological process model |
| CN116229289A (en) * | 2022-12-13 | 2023-06-06 | 河南省农业科学院农业经济与信息研究所 | Unmanned aerial vehicle-based winter wheat biomass rapid and high-precision estimation method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102269576A (en) * | 2010-06-03 | 2011-12-07 | 曹春香 | Active and passive joint inversion method for forest coverage and effective leaf area index |
| CN102495412A (en) * | 2011-11-23 | 2012-06-13 | 中国科学院地理科学与资源研究所 | Multi-scale impervious surface information collaborative remote sensing inversion method |
| CN104483271A (en) * | 2014-12-19 | 2015-04-01 | 武汉大学 | Forest biomass amount retrieval method based on collaboration of optical reflection model and microwave scattering model |
| CN107885909A (en) * | 2017-10-18 | 2018-04-06 | 中国水利水电科学研究院 | A kind of weight of water quality remote sensing multi-model collaboration inverting assembles optimization method and device |
| CN109344550A (en) * | 2018-11-26 | 2019-02-15 | 国智恒北斗科技集团股份有限公司 | A kind of forest reserves inversion method and system based on domestic high score satellite remote sensing date |
-
2021
- 2021-02-09 CN CN202110178986.1A patent/CN112861435B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102269576A (en) * | 2010-06-03 | 2011-12-07 | 曹春香 | Active and passive joint inversion method for forest coverage and effective leaf area index |
| CN102495412A (en) * | 2011-11-23 | 2012-06-13 | 中国科学院地理科学与资源研究所 | Multi-scale impervious surface information collaborative remote sensing inversion method |
| CN104483271A (en) * | 2014-12-19 | 2015-04-01 | 武汉大学 | Forest biomass amount retrieval method based on collaboration of optical reflection model and microwave scattering model |
| CN107885909A (en) * | 2017-10-18 | 2018-04-06 | 中国水利水电科学研究院 | A kind of weight of water quality remote sensing multi-model collaboration inverting assembles optimization method and device |
| CN109344550A (en) * | 2018-11-26 | 2019-02-15 | 国智恒北斗科技集团股份有限公司 | A kind of forest reserves inversion method and system based on domestic high score satellite remote sensing date |
Non-Patent Citations (13)
| Title |
|---|
| HUIYI SU .ETAL: "Machine learning and geostatistical approaches for estimating aboveground biomass in Chinese subtropical forests", 《WWW.REAEARCHAQUARE.COM/ARTICLE/RS-25148/V3》 * |
| HUIYI SU .ETAL: "Machine learning and geostatistical approaches for estimating aboveground biomass in Chinese subtropical forests", 《WWW.REAEARCHAQUARE.COM/ARTICLE/RS-25148/V3》, 6 October 2020 (2020-10-06), pages 1 - 36 * |
| 冯雪力;刘全明;: "基于多源遥感协同反演的区域性土壤盐渍化监测", 农业机械学报, no. 07, pages 134 - 140 * |
| 冯雪力等: "基于多源遥感协同反演的区域性土壤盐渍化监测", 《农业机械学报》 * |
| 冯雪力等: "基于多源遥感协同反演的区域性土壤盐渍化监测", 《农业机械学报》, no. 07, 16 May 2018 (2018-05-16), pages 134 - 140 * |
| 吴剑;陈鹏;傅世锋;陈庆辉;潘翔;宋志晓;: "基于无人机多源遥感数据海岛植被叶面积指数协同反演研究", 海洋技术学报, no. 06, pages 1 - 6 * |
| 吴剑等: "基于无人机多源遥感数据海岛植被叶面积指数协同反演研究", 《海洋技术学报》 * |
| 吴剑等: "基于无人机多源遥感数据海岛植被叶面积指数协同反演研究", 《海洋技术学报》, no. 06, 15 December 2019 (2019-12-15), pages 1 - 6 * |
| 李雨秦 等: "基于随机森林的红树林种群可分性度量研究", 《陕西理工大学学报》 * |
| 李雨秦 等: "基于随机森林的红树林种群可分性度量研究", 《陕西理工大学学报》, vol. 36, no. 2, 30 April 2020 (2020-04-30), pages 64 - 70 * |
| 武高洁 等: "红树林湿地地上生物量遥感估算研究进展_武高洁", 《南方农业学报》 * |
| 武高洁 等: "红树林湿地地上生物量遥感估算研究进展_武高洁", 《南方农业学报》, 30 April 2013 (2013-04-30), pages 693 - 696 * |
| 王宗梅;岳彩荣;刘琦;胡振华;柴凡一;: "基于光学和微波遥感数据的森林蓄积量估测模型研究", 西南农业学报, no. 08, pages 175 - 179 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113536937A (en) * | 2021-06-17 | 2021-10-22 | 海南省林业科学研究院(海南省红树林研究院) | Mangrove forest ecological quantitative inversion method based on remote sensing technology |
| CN113552079A (en) * | 2021-06-17 | 2021-10-26 | 海南省林业科学研究院(海南省红树林研究院) | Mangrove forest quantitative inversion system based on aviation hyperspectral data |
| CN114924034A (en) * | 2022-05-06 | 2022-08-19 | 贵州师范大学 | Forestry carbon measurement system based on ecological process model |
| CN114924034B (en) * | 2022-05-06 | 2024-05-14 | 贵州师范大学 | Forestry carbon metering system based on ecological process model |
| CN116229289A (en) * | 2022-12-13 | 2023-06-06 | 河南省农业科学院农业经济与信息研究所 | Unmanned aerial vehicle-based winter wheat biomass rapid and high-precision estimation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112861435B (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112861435B (en) | Mangrove quality remote sensing inversion method and intelligent terminal | |
| Duarte et al. | QPhenoMetrics: An open source software application to assess vegetation phenology metrics | |
| Peña-Barragán et al. | Object-based crop identification using multiple vegetation indices, textural features and crop phenology | |
| Fazakas et al. | Regional forest biomass and wood volume estimation using satellite data and ancillary data | |
| Viedma et al. | Modeling plant species richness using reflectance and texture data derived from QuickBird in a recently burned area of Central Spain | |
| CN109711102B (en) | Method for rapidly evaluating crop disaster loss | |
| CN108458978B (en) | Sensitive waveband and waveband combination optimal tree species multispectral remote sensing identification method | |
| CN114740180A (en) | Soil organic carbon estimation method and device based on multi-source remote sensing data | |
| Odebiri et al. | Modelling soil organic carbon stock distribution across different land-uses in South Africa: A remote sensing and deep learning approach | |
| Setiawan et al. | Characterizing temporal vegetation dynamics of land use in regional scale of Java Island, Indonesia | |
| CN112364681B (en) | Vegetation coverage estimation method and device based on two-dimensional table | |
| Bayat et al. | Ten-year estimation of Oriental beech (Fagus orientalis Lipsky) volume increment in natural forests: A comparison of an artificial neural networks model, multiple linear regression and actual increment | |
| Dalmayne et al. | Assessment of fine-scale plant species beta diversity using WorldView-2 satellite spectral dissimilarity | |
| Patenaude et al. | Integrating remote sensing datasets into ecological modelling: a Bayesian approach | |
| CN112903606A (en) | Mangrove forest ecological restoration force assessment method based on unmanned aerial vehicle hyperspectrum | |
| Yue et al. | Estimating fractional coverage of crop, crop residue, and bare soil using shortwave infrared angle index and Sentinel-2 MSI | |
| Rakuasa et al. | Analysis of Vegetation Index in Ambon City Using Sentinel-2 Satellite Image Data with Normalized Difference Vegetation Index (NDVI) Method based on Google Earth Engine | |
| CN114494909A (en) | Method and system for generating spatial distribution diagram of soybean growing season | |
| WO2023131949A1 (en) | A versatile crop yield estimator | |
| CN114611699A (en) | Soil moisture downscaling method and device, electronic equipment and storage medium | |
| Atzberger et al. | Portability of neural nets modelling regional winter crop acreages using AVHRR time series | |
| Kozoderov et al. | Models of pattern recognition and forest state estimation based on hyperspectral remote sensing data | |
| Cai | Vegetation Observation in the Big Data Era: Sentinel-2 data for mapping the seasonality of land vegetation | |
| Assiri et al. | Leaf area index and aboveground biomass estimation of an alpine peatland with a UAV multi-sensor approach | |
| Na et al. | Evalution of random forest ensemble classification for land cover mapping using TM and ancillary geographical data |
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 |