CN119007023A - Land degradation trend analysis method and system based on remote sensing - Google Patents
Land degradation trend analysis method and system based on remote sensing Download PDFInfo
- Publication number
- CN119007023A CN119007023A CN202411488480.0A CN202411488480A CN119007023A CN 119007023 A CN119007023 A CN 119007023A CN 202411488480 A CN202411488480 A CN 202411488480A CN 119007023 A CN119007023 A CN 119007023A
- Authority
- CN
- China
- Prior art keywords
- data
- time
- point
- trend
- land
- 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.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
- G06V10/143—Sensing or illuminating at different wavelengths
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/62—Extraction of image or video features relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
-
- 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)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
本发明涉及图像分析技术领域，具体为一种基于遥感的土地退化趋势分析方法及系统，包括以下步骤：基于目标区域的多时相遥感数据，对每个像素点执行光谱分析，计算像素点在多个时间点的光谱反射率变化。本发明中，通过对多时相遥感数据进行细致的光谱分析和时间序列分析，提升了土地退化监测的精确度，通过精确计算光谱反射率数据的时间变化，帮助尽早发现土地退化迹象，利用移动平均值分析主趋势，有效减少随机波动的干扰，提高了结果的稳定性，变点检测能够标记关键变化点，提升了对土地退化动态的理解，并有助于快速调整土地管理策略，通过进一步的光谱差异数据分类和特征提取，能够预测未来土地覆盖类型变化，优化资源分配和规划的准确性。
The present invention relates to the field of image analysis technology, specifically to a land degradation trend analysis method and system based on remote sensing, comprising the following steps: based on multi-temporal remote sensing data of a target area, performing spectral analysis on each pixel point, and calculating the spectral reflectance changes of the pixel point at multiple time points. In the present invention, the accuracy of land degradation monitoring is improved by performing detailed spectral analysis and time series analysis on multi-temporal remote sensing data, and the time changes of spectral reflectance data are accurately calculated to help find signs of land degradation as early as possible. The main trend is analyzed by using a moving average value, which effectively reduces the interference of random fluctuations and improves the stability of the results. Change point detection can mark key change points, improve the understanding of land degradation dynamics, and help to quickly adjust land management strategies. Through further spectral difference data classification and feature extraction, future land cover type changes can be predicted, and the accuracy of resource allocation and planning can be optimized.
Description
Technical Field
The invention relates to the technical field of image analysis, in particular to a land degradation trend analysis method and system based on remote sensing.
Background
The field of image analysis technology covers a wide range of technologies from basic image processing to advanced image recognition, detection and classification. These techniques aim at extracting useful information from a digital image and converting it into structured data that understand the content of the image. Methods of image analysis include, but are not limited to, edge detection, feature extraction, pattern recognition, and application of machine learning algorithms, all for enhancing, interpreting, and quantifying image data. The technology in the field is widely applied to medical imaging, satellite remote sensing, industrial detection, safety monitoring and automatic driving systems, and provides a powerful tool for supporting decision making and analysis.
Among them, the land degradation trend analysis method refers to monitoring and evaluating the change of land coverage and land utilization with time using image analysis technology, and particularly focuses on the phenomenon of land degradation. This method typically uses remote sensing images to analyze the change in land coverage by comparing the surface images at different points in time, thereby identifying degradation trends. Uses of such techniques include environmental monitoring, agricultural management, urban planning and natural resource management, by providing timely data to support land management decisions and environmental protection measures.
While the prior art supports basic image analysis, a common problem when processing multi-temporal data is that long-term trends are not well interpreted and capture capability for subtle changes is lacking. The method is insensitive to land degradation monitoring due to the deficiency in spectral data analysis, and early signals of degradation cannot be identified in time. As in agricultural management, failure to timely identify land degradation can lead to reduced crop yields, affecting food safety and agricultural sustainability. In urban planning and resource management, this lack of technology may lead to planning errors, increasing the risk of damage to the ecosystem.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides a land degradation trend analysis method and system based on remote sensing.
In order to achieve the above purpose, the present invention adopts the following technical scheme: a land degradation trend analysis method based on remote sensing comprises the following steps:
S1: performing spectral analysis on each pixel point based on multi-temporal remote sensing data of the target area, calculating spectral reflectivity changes of the pixel point at a plurality of time points, recording corresponding time sequence data, and generating initial time sequence difference data;
S2: applying time sequence analysis to the initial time sequence difference data, calculating a moving average value of each pixel point, extracting trend components of each data point, and generating a trend analysis result;
S3: detecting the change points by using the trend analysis result, identifying the change points in the time sequence of each pixel point, recording the time position of each change point, comparing the data difference before and after the change points, and generating a change point detection result;
S4: and sorting spectrum difference data before and after the variable points by utilizing the variable point detection result, performing pixel classification and feature extraction by combining trend slope and seasonal parameters, and predicting future land coverage type change to generate a land coverage prediction model.
As a further aspect of the present invention, the step of obtaining the initial time sequence difference data specifically includes:
S111: acquiring multi-temporal remote sensing data in a target area, performing spectrometer measurement, acquiring spectral reflectivity data of each pixel at a plurality of time points, and generating a multi-temporal spectral data set;
S112: analyzing the multi-temporal spectrum data set, calculating spectrum change between any two continuous time points, including spectrum reflectivity difference of each pixel point, and recording spectrum change values;
s113: integrating the spectrum variation value, and adopting the formula:
；
calculating to obtain the time sequence spectrum difference of each pixel point, and generating initial time sequence difference data;
Wherein, The time-series difference data is represented,Representative time pointIs used for the measurement of the spectral reflectance data of (a),The difference in absolute value is represented by a difference in absolute value,In order to participate in the number of points in time of the calculation,The representation takes the square root to smooth the difference.
As a further aspect of the present invention, the step of obtaining the trend analysis result specifically includes:
S211: selecting each pixel point data extracted from the initial time sequence difference data, applying a time sequence analysis method, executing moving average calculation, and generating moving average data by weighting and smoothing data points;
s212: extracting key trend changes from the moving average data by calculating the long-term trend of each time point, and generating main trend data;
s213: and synthesizing the main trend data, and adopting the formula:
；
estimating a trend in a future time period, and generating a trend analysis result;
Wherein, The results of the trend analysis are indicated,Before representationThe main trend data of the time points,Representing the weighting factor, ensures that the farther the time is from the current, the less the weight,As a total number of weights, the weight,For enhancing the impact of the long-term data.
As a further aspect of the present invention, the step of obtaining the change point detection result specifically includes:
s311: applying the trend analysis result, deploying a variable point detection algorithm to identify the change in the time sequence, calibrating potential variable points by comparing the data fluctuation of continuous time points, and generating a preliminary variable point list;
S312: determining the time position of each variable point from the preliminary variable point list, recording the data difference before and after each variable point, drawing the time and the associated influence of the variable point, and generating a variable point time position record;
s313: integrating the time and position records of the variable points, calculating the significance of each variable point, and adopting the formula:
；
Generating a change point detection result;
Wherein, Indicating a point in timeIs used for the data value of (a),As weight, highlight time pointAndThe significance of the difference between them,Is a constant, is used for balancing denominator,And (5) representing the significance scores of the variable points, and ensuring that the scores reflect the significance of the variable points.
As a further aspect of the present invention, the step of obtaining the land coverage prediction model specifically includes:
s411: sorting the variable point detection results, extracting spectrum difference data of each pixel point, including spectrum reflectivity differences before and after the variable point, and generating spectrum difference data;
s412: based on the spectrum difference data, a time sequence decomposition algorithm is adopted to analyze trend slope and seasonal parameters, pixels are classified, each class represents trend and periodic behavior of a land coverage type, and classification and feature data are generated;
S413: and processing the classification and characteristic data, predicting land coverage type change in a future time period, and adopting the formula:
；
generating a land cover prediction model;
Wherein, The trend slope data is represented as such,Representing the parameter of the season,Features representing neighboring pixels, enhancing the predictive ability of the model to future changes,Representing the output of the predictive model,、、The weights of the multiple factors in the prediction are adjusted.
The time sequence decomposition algorithm decomposes the spectrum difference data into three parts of long-term trend, seasonal period and noise according to the formula:
；
calculating trend slope and seasonal parameters of the land cover type;
Wherein, Representative time pointIs used for the measurement of the spectrum data of the (a),Representative time pointIs used for the long-term trend of (a),Representative time pointIs used as a component of the composition,Representative time pointIs a function of the noise of the (c),Is a coefficient for adjusting the influence degree of noise, is used for scaling and adjusting the noise,Is and period ofAn associated adjustment factor for controlling the intensity of the influence of the seasonal period,Is a noise smoothing coefficient for attenuating the effect of noise when its influence is small,Representing the length of the seasonal period.
A remote sensing-based land degradation trend analysis system for performing the above-described remote sensing-based land degradation trend analysis method, the system comprising:
the multi-time phase data processing module acquires multi-time phase image data of a target area from a remote sensing satellite, reads spectral reflectance data of each pixel point in the image, records spectral reflectance changes of a plurality of time points, calculates initial time sequence difference data, and generates a time sequence spectrum difference record;
The trend analysis module executes moving average calculation based on the time sequence spectrum difference record, executes main trend extraction on each data point, and applies linear regression analysis on the time sequence data to obtain a main trend analysis result;
The variable point detection module uses the main trend analysis result to detect variable points, identifies significant change points in the time sequence of each pixel point through statistical tests, records the time position of each variable point, compares the data before and after the variable points, analyzes the intensity and direction of data fluctuation, and generates variable point time records and variable intensity analysis;
And the land coverage prediction module integrates seasonal change parameters of the land by utilizing the change point time record and the change intensity analysis, classifies pixels by cluster analysis, extracts characteristic data, establishes a prediction model of the land coverage type and generates a land coverage prediction result.
Compared with the prior art, the invention has the advantages and positive effects that:
According to the invention, the precision of land degradation monitoring is remarkably improved by carrying out fine spectrum analysis and time sequence analysis on the multi-temporal remote sensing data. By accurately calculating the time variation of the spectral reflectance data, early detection of signs of land degradation is facilitated. And the moving average value is utilized to analyze the main trend, so that the interference of random fluctuation is effectively reduced, and the stability of the result is improved. The change point detection can mark key change points, so that the understanding of the land degradation dynamics is improved, and the land management strategy can be adjusted quickly. Through further spectrum difference data classification and feature extraction, the scheme can predict future land coverage type changes, and the accuracy of resource allocation and planning is optimized.
Drawings
FIG. 1 is a schematic workflow diagram of the present invention;
FIG. 2 is a flowchart showing the steps for acquiring initial time-series difference data according to the present invention;
FIG. 3 is a flowchart showing steps for obtaining trend analysis results according to the present invention;
FIG. 4 is a flowchart showing the steps for obtaining the detection result of the change point according to the present invention;
FIG. 5 is a flowchart showing the steps for obtaining a land cover prediction model according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the description of the present invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate describing the present invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, in the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
Embodiment one: referring to fig. 1, the present invention provides a technical solution: a land degradation trend analysis method based on remote sensing comprises the following steps:
S1: performing spectral analysis on each pixel point based on multi-temporal remote sensing data of the target area, calculating spectral reflectivity changes of the pixel point at a plurality of time points, recording corresponding time sequence data, and generating initial time sequence difference data;
S2: applying time sequence analysis to the initial time sequence difference data, calculating a moving average value of each pixel point, extracting trend components of each data point, and generating a trend analysis result;
S3: detecting variable points by using trend analysis results, identifying the variable points in the time sequence of each pixel point, recording the time position of each variable point, comparing the data difference before and after the variable points, and generating variable point detection results;
S4: and sorting spectrum difference data before and after the variable points by utilizing the variable point detection results, performing pixel classification and feature extraction by combining trend slope and seasonal parameters, and predicting future land coverage type changes to generate a land coverage prediction model.
The initial time sequence difference data comprises a time point mark and a spectrum reflectance difference record, the trend analysis result comprises a moving average value and a main trend, the change point detection result specifically refers to change point time positions and data difference comparison, and the land coverage prediction model comprises pixel classification, feature extraction, trend slope and seasonal parameters.
Referring to fig. 2, the initial time sequence difference data is obtained by the steps of:
S111: acquiring multi-temporal remote sensing data in a target area, performing spectrometer measurement, acquiring spectral reflectivity data of each pixel at a plurality of time points, and generating a multi-temporal spectral data set;
After selecting a target area and acquiring multi-phase remote sensing data, performing spectrometer measurement to acquire spectral reflectance data of each pixel point at a plurality of time points, wherein the multi-phase spectral data set is obtained through real-time monitoring and comprises original data obtained from a remote sensing satellite, the spectral analysis needs to conduct time sequence comparison on the data, the material composition change under a plurality of time points is identified through the comparison of the spectral resolution, the spectral data of each pixel point comprises spectral intensity and wavelength information of the material composition change, and the spectral data of each pixel point is obtained through direct measurement of a spectrometer and comprises visible light to near infrared wave bands, so that a foundation is laid for the next step of spectral difference data calculation.
S112: analyzing the multi-phase spectrum data set, calculating the spectrum change between any two continuous time points, including the spectrum reflectivity difference of each pixel point, and recording the spectrum change value;
In analysing a multi-temporal spectral data set, firstly the spectral reflectance data of each pixel point between two consecutive time points is extracted and recorded, this step ensures that the data of each time point can be captured, then the direction and amplitude of the change are assessed by comparing the spectral differences between consecutive time points, focusing on the amount of change of the pixel points, this method reveals potential trends and abnormal changes in the data, the recording of the spectral changes provides the necessary basis for the subsequent analysis of the data, for identifying and analysing pattern changes in the spectral data, and finally the recorded spectral change values will be used for iterative analysis and model building, providing a basis for the prediction of future spectral behaviour.
S113: integrating the spectrum change value, adopting the formula:
；
calculating to obtain the time sequence spectrum difference of each pixel point, and generating initial time sequence difference data;
Wherein, The time-series difference data is represented,Representative time pointIs used for the measurement of the spectral reflectance data of (a),The difference in absolute value is represented by a difference in absolute value,In order to participate in the number of points in time of the calculation,The representation takes the square root to smooth the difference.
The formula:
；
The formula has the advantages that extreme changes in the spectrum data are smoothed through a root opening mode of square sums, and the influence of abnormal points on overall difference evaluation is reduced.
Formula details and formula calculation derivation process:
is arranged at a certain target pixel point and a time point Spectral reflectance data of (2)150, At the point in timeSpectral reflectance data of (2)160, The difference:
；
if there are 5 such consecutive time points, the sum of squares of the differences is:
；
Time-series difference data:
；
the results show that the degree of change of the spectral reflectance data between each time point is 4.47 units in average during the investigation period, and the numerical assistance analyzes the average intensity of the spectral difference in the time sequence, so that the method can be used for monitoring environmental changes or other key surface characteristics.
Referring to fig. 3, the trend analysis result obtaining steps specifically include:
S211: selecting each pixel point data extracted from the initial time sequence difference data, applying a time sequence analysis method, executing moving average calculation, and generating moving average data by weighting and smoothing data points;
Selecting initial time sequence difference data, executing a moving average calculation method, smoothing the influence of data points to reduce random fluctuation by a weighting method, wherein in the calculation process, the influence of the size of a reference time window on a result is critical, the time window is too small to effectively smooth the data, and the critical trend change is covered if the time window is too large, and the size of a window which is best matched is determined by the frequency distribution and the fluctuation analysis of actual data, so that the smooth trend of each pixel point can be obtained by the weighting moving average method, and the trend is used as a basis for analyzing long-term trends.
S212: from the moving average data, extracting key trend changes by calculating long-term trends of each time point, and generating main trend data;
Based on moving average data, extracting main trend, extracting key trend change by calculating long-term trend components, determining dominant trend of the data, detecting seasonal and periodical components in actual data, modeling the components by using a time sequence analysis method comprising ARIMA or seasonal decomposition, and further extracting and quantifying the main trend of each data point, wherein the main trend reflects the overall moving direction of the data, and provides a quantified basis for subsequent trend analysis.
S213: synthesizing main trend data, and adopting the formula:
；
estimating a trend in a future time period, and generating a trend analysis result;
Wherein, The results of the trend analysis are indicated,Before representationThe main trend data of the time points,Representing the weighting factor, ensures that the farther the time is from the current, the less the weight,As a total number of weights, the weight,For enhancing the impact of the long-term data.
The formula:
；
The formula has the advantages that the sensitivity and the reaction capacity of the data trend are increased by introducing square terms and adjustment coefficients, the influence of the data fluctuation on the trend can be reflected more carefully, and the prediction accuracy is enhanced.
Formula details and formula calculation derivation process:
Setting up ，，，，，First, the denominator of the weight adjustment is calculated:
；
Then calculate the weighted trend value molecules:
；
so that the number of the parts to be processed, ；
The results show that the trend analysis results are 89 with reference to the weight and data volatility between the data points according to the current and past data points, which means that in the current time series analysis, the trend value is 89, which is a comprehensive trend index obtained based on past data and adjustment coefficients, reflects the main trend of the data points on the time series, and provides a quantitative basis for data analysis and decision.
Referring to fig. 4, the variable point detection result is obtained by the following steps:
s311: applying a trend analysis result, deploying a variable point detection algorithm to identify the change in the time sequence, calibrating potential variable points by comparing the data fluctuation of continuous time points, and generating a preliminary variable point list;
The method comprises the steps of deploying a variable point detection algorithm to identify changes in a time sequence, calibrating variable points according to the fact that the difference is larger than a preset sensitive threshold value through comparing data fluctuation between continuous time points, determining the threshold value based on a data change average value of a previous continuous period, and matching with reference environmental noise or potential data errors, so that the sensitivity and accuracy of the algorithm to real variable points are ensured, and the generated preliminary variable point list not only reflects mutation of the time sequence, but also eliminates influence caused by random fluctuation or data acquisition occasional errors.
S312: determining the time position of each variable point from the preliminary variable point list, recording the data difference before and after each variable point, drawing the time and the associated influence of the variable point, and generating a variable point time position record;
Determining the time position of each variable point from the generated preliminary variable point list, recording the data difference before and after the variable point, recording the time and data change of each variable point, comprising the comparison analysis of the data before and after the variable point, ensuring the importance and the influence range of each variable point to be accurately identified through quantitative data comparison comprising the change amplitude and the change rate, and providing accurate input for the subsequent data processing and analysis through the screening of truly critical variable points.
S313: integrating the time and position records of the variable points, calculating the significance of each variable point, and adopting the formula:
；
Generating a change point detection result;
Wherein, Indicating a point in timeIs used for the data value of (a),As weight, highlight time pointAndThe significance of the difference between them,Is a constant, is used for balancing denominator,And (5) representing the significance scores of the variable points, and ensuring that the scores reflect the significance of the variable points.
The formula:
；
The formula has the advantages that the significance of the variable points can be quantified through the weighted difference, the sensitivity and the response speed of the model to local mutation are enhanced by referring to the local characteristics of the time series data, and meanwhile, the sensitivity and the response speed of the model to local mutation are improved through parameters And (3) adjusting, and enhancing the adaptability and stability of the model in a differential data environment.
Formula details and formula calculation derivation process:
in the example of change point detection, a time series is set up in To the point ofThe period is obviously changed, and the data difference before and after the change point is calculated according to the values of 120-100, 130-80 and 140-70, provided thatAll of them are 1, and are made into the invented product,。
；
The results show that inTo the point ofThe score of the significant change was 0.9333, indicating that this is a highly significant change point with critical reference value for subsequent data analysis and decision making.
Referring to fig. 5, the steps for obtaining the land coverage prediction model specifically include:
S411: sorting the variable point detection results, extracting spectrum difference data of each pixel point, including spectrum reflectivity differences before and after the variable point, and generating spectrum difference data;
In the process of sorting the change point detection result, firstly, the spectrum difference data of each pixel point needs to be extracted, the key of the step is that the change of the spectrum reflectivity data before and after the change point is compared to determine which changes are obvious and which are only noise, then the data is recorded, necessary input information is provided for the subsequent pixel classification and feature extraction, the real-time property and the correlation of the data are ensured through monitoring the change in real time, and the subsequent step can be carried out according to the actually observed spectrum change when the data are processed.
S412: based on spectrum difference data, a time sequence decomposition algorithm is adopted to analyze trend slope and seasonal parameters, pixels are classified, each class represents trend and periodic behavior of a land coverage type, and classification and feature data are generated;
According to spectrum difference data, analyzing trend slope and seasonal parameters of each pixel point is a core task of the section, the analysis comprises combining trend slope and seasonal variation data through a mathematical model, a classification algorithm is used for classifying pixels into different land coverage types according to comprehensive data, each type represents a target land use mode, and in the classification process, accurate input and correct processing of the data are particularly paid attention to so as to ensure reliability and accuracy of classification results, so that a stable foundation is provided for a land coverage prediction model.
S413: processing the classification and feature data, predicting land coverage type change in a future time period, and adopting the formula:
；
generating a land cover prediction model;
Wherein, The trend slope data is represented as such,Representing the parameter of the season,Features representing neighboring pixels, enhancing the predictive ability of the model to future changes,Representing the output of the predictive model,、、The weights of the multiple factors in the prediction are adjusted.
The formula:
；
the formula is beneficial in that the trend slope is calculated ) Seasonal parameter) Data of adjacent pixels) In combination, the instantaneous change and the long-term periodic characteristic of the reference time series can be integrated, so that the land coverage change in the future time period can be predicted more accurately.
Formula details and formula calculation derivation process:
Setting up ，For the following，Weight coefficient，，，。
；
；
；
The results show that the model predicts a land cover change index of 0.77 for a given weight and parameter setting, which means that there is a high likelihood that the land cover type will change in the future based on the current trend slope, seasonal parameters, and data for neighboring pixels.
The time sequence decomposition algorithm decomposes the spectrum difference data into three parts of long-term trend, seasonal period and noise according to the formula:
；
calculating trend slope and seasonal parameters of the land cover type;
Wherein, Representative time pointIs used for the measurement of the spectrum data of the (a),Representative time pointIs used for the long-term trend of (a),Representative time pointIs used as a component of the composition,Representative time pointIs a function of the noise of the (c),Is a coefficient for adjusting the influence degree of noise, is used for scaling and adjusting the noise,Is and period ofAn associated adjustment factor for controlling the intensity of the influence of the seasonal period,Is a noise smoothing coefficient for attenuating the effect of noise when its influence is small,Representing the length of the seasonal period.
The formula:
；
formula details and formula calculation derivation process:
Parameter acquisition mode and numerical value:
Indicating a point in time Spectrum data of (2) is obtained by remote sensing image spectrum analysis, and a pixel point in a certain region is taken at a time pointIs assumed to be the point in timeSpectral data of (2) are(In reflectance);
Indicating a point in time Can be obtained by moving average of long-term history spectrum data, and obtained by using 5 years moving average；
Indicating a point in timeIs obtained by periodically decomposing the spectral data, periodically decomposing the data using a fourier transform, assuming that the fourier transform is obtained at a point in timeThe seasonal component of (2) is；
Indicating a point in timeCan be determined by residual analysis, using the residual of spectral difference data subtracted with trend and seasonal components as noise, assuming that the noise is；
To adjust the coefficient of noise influence degree, the method is set according to the stability of the land coverage type and the noise level, and selectsThis value is set according to historical empirical data, and is increased when the noise is large to reduce the influence;
for adjusting the coefficient related to the seasonal period, for controlling the period influence, setting This coefficient is set based on the degree of importance of the seasonal impact to the overall trend impact;
For seasonal period length, the period is determined to be one year by analyzing the period change characteristics in the data, namely ；
For smoothing noise, for reducing the influence of noise at a small timeBased on an analysis of the effect of noise on the overall calculation.
Calculation and deduction process of formula: first, all known parameters are put into the formula:
；
substituting a known value:
；
Step-by-step calculation is performed:
first, the square root term is calculated:
；
Then calculate denominator part:
；
Calculating the components of noise:
；
Calculate the final ：
；
Analysis of results: the results indicate the time pointIs of (3)For 0.6377, this value contains the effects of long-term trends, seasonal components, and noise. By calculation, it can be found that the effect of noise on the spectral data is properly tuned, smoother, and long-term trends and seasonal components are reasonably separated and weighted integrated. This value can be used to further analyze the varying characteristics of the land cover type, helping to accurately classify and model the land cover type.
A remote sensing-based land degradation trend analysis system for executing the above-mentioned remote sensing-based land degradation trend analysis method, the system comprising:
the multi-time phase data processing module acquires multi-time phase image data of a target area from a remote sensing satellite, reads spectral reflectance data of each pixel point in the image, records spectral reflectance changes of a plurality of time points, calculates initial time sequence difference data, and generates a time sequence spectrum difference record;
The trend analysis module is used for executing moving average calculation based on the time sequence spectrum difference record, executing main trend extraction on each data point, and applying linear regression analysis on the time sequence data to obtain a main trend analysis result;
The variable point detection module uses a main trend analysis result to detect variable points, identifies significant change points in the time sequence of each pixel point through a statistical test, records the time position of each variable point, compares the data before and after the variable points, analyzes the intensity and direction of data fluctuation, and generates variable point time records and variable intensity analysis;
The land coverage prediction module integrates seasonal change parameters of the land by utilizing the change point time record and the change intensity analysis, classifies pixels by cluster analysis, extracts characteristic data, establishes a prediction model of a land coverage type and generates a land coverage prediction result.
The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.
Claims (8)
1. The land degradation trend analysis method based on remote sensing is characterized by comprising the following steps of:
Performing spectral analysis on each pixel point based on multi-temporal remote sensing data of the target area, calculating spectral reflectivity changes of the pixel point at a plurality of time points, recording corresponding time sequence data, and generating initial time sequence difference data;
Applying time sequence analysis to the initial time sequence difference data, calculating a moving average value of each pixel point, extracting trend components of each data point, and generating a trend analysis result;
Detecting the change points by using the trend analysis result, identifying the change points in the time sequence of each pixel point, recording the time position of each change point, comparing the data difference before and after the change points, and generating a change point detection result;
And sorting spectrum difference data before and after the variable points by utilizing the variable point detection result, performing pixel classification and feature extraction by combining trend slope and seasonal parameters, and predicting future land coverage type change to generate a land coverage prediction model.
2. The remote sensing-based land degradation trend analysis method of claim 1, wherein the initial time series difference data comprises a time point mark and a spectrum reflectivity difference record, the trend analysis result comprises a moving average value and a main trend, the change point detection result specifically refers to change point time position and data difference comparison, and the land coverage prediction model comprises pixel classification, feature extraction, trend slope and seasonal parameters.
3. The method for analyzing land degradation trend based on remote sensing according to claim 1, wherein the step of obtaining the initial time-series difference data comprises the steps of:
Acquiring multi-temporal remote sensing data in a target area, performing spectrometer measurement, acquiring spectral reflectivity data of each pixel at a plurality of time points, and generating a multi-temporal spectral data set;
analyzing the multi-temporal spectrum data set, calculating spectrum change between any two continuous time points, including spectrum reflectivity difference of each pixel point, and recording spectrum change values;
integrating the spectrum variation value, and adopting the formula:
；
calculating to obtain the time sequence spectrum difference of each pixel point, and generating initial time sequence difference data;
Wherein, The time-series difference data is represented,Representative time pointIs used for the measurement of the spectral reflectance data of (a),The difference in absolute value is represented by a difference in absolute value,In order to participate in the number of points in time of the calculation,The representation takes the square root to smooth the difference.
4. The method for analyzing land degradation trend based on remote sensing according to claim 3, wherein the step of obtaining the trend analysis result comprises the steps of:
selecting each pixel point data extracted from the initial time sequence difference data, applying a time sequence analysis method, executing moving average calculation, and generating moving average data by weighting and smoothing data points;
extracting key trend changes from the moving average data by calculating the long-term trend of each time point, and generating main trend data;
And synthesizing the main trend data, and adopting the formula:
；
estimating a trend in a future time period, and generating a trend analysis result;
Wherein, The results of the trend analysis are indicated,Before representationThe main trend data of the time points,Representing the weighting factor, ensures that the farther the time is from the current, the less the weight,As a total number of weights, the weight,For enhancing the impact of the long-term data.
5. The remote sensing-based land degradation trend analysis method according to claim 4, wherein the obtaining step of the change point detection result specifically comprises:
Applying the trend analysis result, deploying a variable point detection algorithm to identify the change in the time sequence, calibrating potential variable points by comparing the data fluctuation of continuous time points, and generating a preliminary variable point list;
Determining the time position of each variable point from the preliminary variable point list, recording the data difference before and after each variable point, drawing the time and the associated influence of the variable point, and generating a variable point time position record;
integrating the time and position records of the variable points, calculating the significance of each variable point, and adopting the formula:
；
Generating a change point detection result;
Wherein, Indicating a point in timeIs used for the data value of (a),As weight, highlight time pointAndThe significance of the difference between them,Is expressed at the calculation time pointThe number of surrounding time points considered when the variability of (a) is significant,Is a constant, is used for balancing denominator,And (5) representing the significance scores of the variable points, and ensuring that the scores reflect the significance of the variable points.
6. The remote sensing-based land degradation trend analysis method according to claim 5, wherein the obtaining step of the land coverage prediction model specifically comprises:
Sorting the variable point detection results, extracting spectrum difference data of each pixel point, including spectrum reflectivity differences before and after the variable point, and generating spectrum difference data;
Based on the spectrum difference data, a time sequence decomposition algorithm is adopted to analyze trend slope and seasonal parameters, pixels are classified, each class represents trend and periodic behavior of a land coverage type, and classification and feature data are generated;
and processing the classification and characteristic data, predicting land coverage type change in a future time period, and adopting the formula:
；
generating a land cover prediction model;
Wherein, The trend slope data is represented as such,Representing the parameter of the season,Features representing neighboring pixels, enhancing the predictive ability of the model to future changes,Representing the output of the predictive model,、、The weights of the multiple factors in the prediction are adjusted.
7. The remote sensing-based land degradation trend analysis method of claim 6, wherein the time series decomposition algorithm decomposes the spectrum difference data into three parts of long-term trend, seasonal period and noise according to the formula:
；
calculating trend slope and seasonal parameters of the land cover type;
Wherein, Representative time pointIs used for the measurement of the spectrum data of the (a),Representative time pointIs used for the long-term trend of (a),Representative time pointIs used as a component of the composition,Representative time pointIs a function of the noise of the (c),Is a coefficient for adjusting the influence degree of noise, is used for scaling and adjusting the noise,Is and period ofAn associated adjustment factor for controlling the intensity of the influence of the seasonal period,Is a noise smoothing coefficient for attenuating the effect of noise when its influence is small,Representing the length of the seasonal period.
8. A remote sensing-based land degradation trend analysis system, characterized in that the remote sensing-based land degradation trend analysis method according to any one of claims 1-5, the system comprising:
the multi-time phase data processing module acquires multi-time phase image data of a target area from a remote sensing satellite, reads spectral reflectance data of each pixel point in the image, records spectral reflectance changes of a plurality of time points, calculates initial time sequence difference data, and generates a time sequence spectrum difference record;
The trend analysis module executes moving average calculation based on the time sequence spectrum difference record, executes main trend extraction on each data point, and applies linear regression analysis on the time sequence data to obtain a main trend analysis result;
The variable point detection module uses the main trend analysis result to detect variable points, identifies significant change points in the time sequence of each pixel point through statistical tests, records the time position of each variable point, compares the data before and after the variable points, analyzes the intensity and direction of data fluctuation, and generates variable point time records and variable intensity analysis;
And the land coverage prediction module integrates seasonal change parameters of the land by utilizing the change point time record and the change intensity analysis, classifies pixels by cluster analysis, extracts characteristic data, establishes a prediction model of the land coverage type and generates a land coverage prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411488480.0A CN119007023A (en) | 2024-10-24 | 2024-10-24 | Land degradation trend analysis method and system based on remote sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411488480.0A CN119007023A (en) | 2024-10-24 | 2024-10-24 | Land degradation trend analysis method and system based on remote sensing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119007023A true CN119007023A (en) | 2024-11-22 |
Family
ID=93484964
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411488480.0A Withdrawn CN119007023A (en) | 2024-10-24 | 2024-10-24 | Land degradation trend analysis method and system based on remote sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119007023A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119374566A (en) * | 2024-12-31 | 2025-01-28 | 山东省鲁岳资源勘查开发有限公司 | A method and system for engineering surveying and mapping based on unmanned aerial vehicle remote sensing |
| CN119380204A (en) * | 2024-12-26 | 2025-01-28 | 中国地质大学（武汉） | Land cover change monitoring method, model training method and related device |
| CN119785211A (en) * | 2024-12-18 | 2025-04-08 | 北京民生智库科技信息咨询有限公司 | A system and method for identifying urban green space encroachment and degradation based on computer vision recognition |
| CN120071163A (en) * | 2025-04-29 | 2025-05-30 | 山东省地质科学研究院 | Terrain identification and three-dimensional reconstruction method based on remote sensing mapping image |
| CN120580419A (en) * | 2025-08-01 | 2025-09-02 | 中国科学院空天信息创新研究院 | Remote sensing image target trend detection method, device, equipment and medium |
-
2024
- 2024-10-24 CN CN202411488480.0A patent/CN119007023A/en not_active Withdrawn
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119785211A (en) * | 2024-12-18 | 2025-04-08 | 北京民生智库科技信息咨询有限公司 | A system and method for identifying urban green space encroachment and degradation based on computer vision recognition |
| CN119380204A (en) * | 2024-12-26 | 2025-01-28 | 中国地质大学（武汉） | Land cover change monitoring method, model training method and related device |
| CN119380204B (en) * | 2024-12-26 | 2025-04-08 | 中国地质大学（武汉） | Land cover change monitoring method, model training method and related device |
| CN119374566A (en) * | 2024-12-31 | 2025-01-28 | 山东省鲁岳资源勘查开发有限公司 | A method and system for engineering surveying and mapping based on unmanned aerial vehicle remote sensing |
| CN120071163A (en) * | 2025-04-29 | 2025-05-30 | 山东省地质科学研究院 | Terrain identification and three-dimensional reconstruction method based on remote sensing mapping image |
| CN120580419A (en) * | 2025-08-01 | 2025-09-02 | 中国科学院空天信息创新研究院 | Remote sensing image target trend detection method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN119007023A (en) | Land degradation trend analysis method and system based on remote sensing | |
| CN119810100B (en) | A persimmon quality detection method and system based on machine vision | |
| CN117576093B (en) | EL image-based solar cell piece hidden crack detection method | |
| CN119915837B (en) | PVC floor surface coating defect detection method and system | |
| Giménez-Gallego et al. | On-tree fruit image segmentation comparing Mask R-CNN and Vision Transformer models. Application in a novel algorithm for pixel-based fruit size estimation | |
| CN118781122B (en) | Industrial image real-time analysis system and method | |
| CN114136920A (en) | Hyperspectrum-based single-grain hybrid rice seed variety identification method | |
| CN119992233A (en) | Leaf pest and disease recognition method and system based on image processing | |
| CN119086464A (en) | Environmental monitoring method and device based on multispectral image processing | |
| US12488875B1 (en) | Targeted control method and system for blackhead removal | |
| CN120009209B (en) | A food contamination detection method and system based on multi-wavelength spectrophotometry | |
| CN118607346B (en) | Printing process parameter optimization method and system | |
| Gu et al. | Quantitative analysis of watermelon fruit skin phenotypic traits via image processing and their potential in maturity and quality detection | |
| Yang et al. | Machine learning based framework for the detection of mushroom browning using a portable hyperspectral imaging system | |
| CN119067982A (en) | A Fault Diagnosis System for Chemical Equipment | |
| CN119202563A (en) | Prediction method, system and storage medium for soil organic carbon content of invasive plants | |
| CN117893537A (en) | Method and system for detecting discoloration of pallet surface material | |
| Irvine et al. | Image quality and performance modeling for automated target detection | |
| CN118628923B (en) | Tree age identification system and method based on image identification | |
| CN119624958B (en) | Intelligent recognition analysis method for colloidal gold test strip result based on machine learning | |
| CN119289908A (en) | Aluminum profile section thickness consistency detection method based on X-ray scanning | |
| CN119688643A (en) | Pine nut pest monitoring system and method based on intelligent sensor | |
| CN117523013B (en) | Color change characteristic analysis method and system for high-moisture-conductivity fabric | |
| CN118014443A (en) | A drug monitoring method and device | |
| CN119411452A (en) | A method for detecting construction of asphalt concrete pavement of municipal roads |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20241122 |