Ground hyperspectral field greenhouse gas N2O monitoring method
By preprocessing ground hyperspectral data and constructing dual-band spectral indices, combined with the random forest (RF) machine learning algorithm, the problems of high-cost equipment and insufficient satellite monitoring are solved, enabling low-cost, high-precision N2O gas monitoring and accurate capture of emission peaks.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-26
- Publication Date
- 2026-04-03
AI Technical Summary
In the existing technology, the high cost of imported equipment and insufficient satellite monitoring capabilities lead to discontinuous and low-precision N2O gas monitoring. Traditional methods are unable to capture emission peaks, resulting in underestimated emissions. Furthermore, N2O emissions are affected by a variety of factors and vary drastically.
By preprocessing and denoising ground hyperspectral data, a dual-band spectral index is constructed, and an N2O gas monitoring model is established using the random forest (RF) machine learning algorithm. The denoised spectral data is then used for monitoring.
It achieves low-cost, high-precision N2O gas monitoring, accurately captures emission peaks, reduces equipment costs, and improves the continuity and accuracy of monitoring.
Smart Images
Figure CN121783867A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of hyperspectral remote sensing technology, and in particular relates to a ground-based hyperspectral method for monitoring the greenhouse gas N2O in open fields. Background Technology
[0002] Farmland ecosystems are the primary source of nitrous oxide (N2O) emissions, contributing approximately 60%-70% of global anthropogenic N2O emissions. Although the concentration of N2O in the atmosphere is far lower than that of CO2, its global warming potential per molecule is 298 times that of CO2 (on a 100-year timescale), and it is a major depleter of stratospheric ozone. Accurate monitoring of N2O is of paramount importance for accurately assessing national greenhouse gas inventories, achieving "dual-carbon" strategic goals, and optimizing agricultural management to ensure food security. Currently, atmospheric greenhouse gas monitoring methods mainly include spectroscopic methods, satellite remote sensing, and gas chromatography. However, spectroscopic and gas chromatography methods largely rely on imported equipment, which is expensive and has high maintenance costs. Furthermore, greenhouse gas satellites have insufficient detection capabilities, and their short observation duration and uncertainties pose challenges to the long-term stable capture of greenhouse gas concentrations. Moreover, N2O emissions are complexly influenced by factors such as fertilization, irrigation, rainfall, and temperature, resulting in extremely rapid fluctuations in emission fluxes. Traditional chamber method intermittent manual sampling is difficult to capture emission peaks, which may lead to a significant underestimation of emissions. Therefore, research on high-precision, low-cost N2O gas monitoring has become particularly important. Summary of the Invention
[0003] Technical Solution: To address the aforementioned technical problems, this invention preprocesses and denoises the collected ground hyperspectral data, and then constructs spectral indices using the denoised spectral data in dual bands. The calculated spectral indices are then combined with indoor measured N2O emission fluxes for modeling and analysis. Specifically, a field greenhouse gas N2O monitoring method based on ground hyperspectral data is proposed, the steps of which are as follows: Step 1: Data Acquisition and Preprocessing Hyperspectral data and corresponding gas samples were collected from field sampling points. The hyperspectral data were preprocessed and noise was reduced to obtain the measured spectral data. The gas samples were analyzed and calculated in the laboratory to obtain the measured N2O gas emission flux data. Step 2: Constructing Spectral Indices The spectral index is constructed using the spectral data after preprocessing and noise reduction in step one. The spectral index is constructed using a dual-band method, which involves using any two spectral bands to construct the spectral index. Step 3: Construction of the N2O Gas Monitoring Model Using the dual-band spectral index as the independent variable and the N2O gas emission flux corresponding to each sample as the dependent variable, the training set and validation set were divided according to the data ratio of 25%-45%. The N2O gas monitoring model was established by the random forest (RF) machine learning algorithm, and the results of N2O gas monitoring were output by the spectral indices obtained by different spectral index construction methods.
[0004] As an improvement, the spectral indices in step two include the given wavelength R, simple ratio SR, wavelength difference D, normalized difference ND, inverse difference ID, double difference DDn, corrected simple ratio mSR1, corrected simple ratio mSR2, corrected normalized difference mND, and corrected inverse difference mID. The formulas for calculating the indices are as follows: ; (1) (2) = (3) (4) (5) in, For the spectral band, The spectral data step size is 10nm. It is a band spectral reflectance, It is a band Spectral reflectance.
[0005] Beneficial effects: The present invention proposes a ground-based hyperspectral method for monitoring N2O greenhouse gas in the field. The method preprocesses and performs SG noise reduction and smoothing on the collected ground-based hyperspectral data. The noise-reduced spectral data is used to construct spectral indices, namely 9 spectral indices constructed by dual bands. These indices are then used to establish an N2O monitoring model, ultimately achieving accurate monitoring of N2O gas at the field scale.
[0006] This invention predicts N2O emissions by modeling the emission flux of N2O gas through outdoor hyperspectral data collection and indoor measurement. This can reduce the problems of high cost, discontinuous monitoring time, and low accuracy of previous N2O monitoring equipment, and achieve accurate results through a simple operation. Attached Figure Description
[0007] Figure 1 This is a flowchart of the present invention.
[0008] Figure 2The graphs shown are the spectral index modeling results constructed in Embodiment 1 of the present invention. Each graph is divided into two parts, left and right. The left part represents the fitting accuracy R of the modeling validation set. 2, The right part represents the normalized root mean square error (NRMSE) of the validation set fitting, and the horizontal and vertical axes represent the two bands corresponding to the constructed spectral indices. (a) shows the modeling results of the spectral indices SR. (b) shows the modeling results of the spectral indices D. (c) shows the modeling results of the spectral indices ID. (d) shows the modeling results of the spectral indices mSR2. (e) shows the modeling results of the spectral indices ND.
[0009] Figure 3 The following are scatter plots of the optimal dual-band index modeling extracted in Embodiment 1 of the present invention: (a) shows the scatter plot of the validation set fitting of the SR index constructed by bands 940nm and 1240nm; (b) shows the scatter plot of the validation set fitting of the D index constructed by bands 530nm and 2140nm; (c) shows the scatter plot of the validation set fitting of the ID index constructed by bands 400nm and 680nm; (d) shows the scatter plot of the validation set fitting of the mSR2 index constructed by bands 680nm and 1980nm; and (e) shows the scatter plot of the validation set fitting of the ND index constructed by bands 1120nm and 1770nm. Detailed Implementation
[0010] The technical solutions in the embodiments of the present invention will be clearly and completely described below, so that those skilled in the art can better understand the advantages and features of the present invention, thereby making a clearer definition of the scope of protection of the present invention. The embodiments described in this invention are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0011] This invention provides a method for monitoring N2O, a greenhouse gas in the field, based on ground-based hyperspectral imaging. The method comprises the following steps: Step 1: Data acquisition and preprocessing. Hyperspectral data and corresponding gas samples were collected from field sampling points. The hyperspectral data were preprocessed and noise was reduced to obtain the measured spectral data. The gas samples were analyzed and calculated in the laboratory to obtain the measured N2O gas emission flux data. Step 2: Constructing Spectral Indices The spectral index is constructed using the spectral data after preprocessing and noise reduction in step one. The spectral index is constructed using a dual-band method, which involves using any two spectral bands to construct the spectral index. Step 3: Construction of the N2O Gas Monitoring Model Using the spectral index constructed from two bands as the independent variable and the N2O gas emission flux corresponding to each sample as the dependent variable, the training and validation sets were divided according to a data ratio of 25%-45%, and a random forest was applied. RF Machine learning algorithms are used to build an N2O gas monitoring model and output the results of N2O gas monitoring based on spectral indices obtained by different spectral index construction methods.
[0012] As a specific embodiment of the present invention, the specific steps in step one are as follows: (1.1) Hyperspectral data: Spectral data acquisition and usage ASD FieldSpec 4. On-site gas spectroscopy analysis was conducted outdoors before gas collection at the sampling points. The fiber optic probe was placed vertically 50 cm above the sampling points. Five spectral curves were collected at each sampling point, with a spectral range of 350 nm to 2500 nm and an interval of 1 nm. The spectral data were then imported. Python In the process SG Smooth noise reduction.
[0013] (1.2) N2O gas data: After the spectral data acquisition was completed, a transparent acquisition device was placed above each sampling point. Four air gas samples were collected at each time point from 0 to 30 minutes using a syringe and placed in an opaque gas bag. The collected gas samples were analyzed and measured in a gas chromatograph in the laboratory, and the emission flux of N2O gas was calculated.
[0014] In this invention, the spectral index construction method in step two is a dual-band spectral index construction, that is, using any two spectral bands to construct the spectral index, including... R , SR , D , ND , ID The index, and the one introduced through the formula Δλ To construct the index DDn , mSR 1. mSR 2. mND , mID .
[0015] The constructed spectral index formulas are shown in Table 1. This section uses common index types, including simple ratios (...). SR ), wavelength difference ( D ), normalized difference ( ND ) and deficit ( ID ), to develop new indices. In addition, it includes dual differences ( DDn ), Corrected simple ratio 1 ( mSR 1) Corrected simplex ratio 2 ( mSR2 ), Corrected normalization differences ( mND ) and modified inverse difference ( mID ).
[0016] Table 1 Spectral Index Formula .
[0017] This invention also includes a method for evaluating the accuracy of the estimation model, specifically: calculating the commonly used coefficient of determination when constructing the crop phenotypic parameter estimation model. R 2 Normalized root mean square error NRMSE Used to evaluate model performance, RMSE The influence of dimensions was eliminated on this basis. When R 2 The value is close to 1 and NRMSE The smaller the value, the better the model performance.
[0018] (6) (7) (8) in n It is the number of samples. yi It is the first i The true value of each sample It is the first i Estimates for each sample It is the average of the actual observed values. It is the maximum value of the observed values. This is the minimum value of the observed values.
[0019] Example 1 This invention provides a method for monitoring N2O, a greenhouse gas in the field, based on ground-based hyperspectral imaging. The method comprises the following steps: Step 1: Data acquisition and preprocessing. use ASD FieldSpec 4. On-site gas spectrometer analysis was conducted outdoors before gas collection at the sampling points. The fiber optic probe was placed vertically 50 cm above the sampling points. Five spectral curves were collected at each sampling point, with a spectral range of 350 nm to 2500 nm and an interval of 1 nm. ViewSpecPro The average value of the five spectral curves at each sampling point was processed, and the spectral bands with absorption peaks caused by atmospheric moisture absorption in the range of 300-399nm and 2401nm-2500nm were deleted.
[0020] After spectral data acquisition, a transparent acquisition device was placed above each sampling point. Four air gas samples were collected at each time point from 0 to 30 minutes using a syringe and placed in an opaque gas bag. The collected gas samples were analyzed and measured using a gas chromatograph in the laboratory, and the emission flux of N2O gas was calculated. The emission flux calculation formula is as follows.
[0021] (9) In the formula: F The emission flux of N2O ( ug / m 2 / h ); ρ The density of CO2 and N2O gases under standard conditions; dc / dt The slope of the regression curve of gas concentration versus time in the static chamber; P 0 The air pressure under standard conditions ( kPa ); T The average temperature (°C) of the air inside the static chamber at the time of sampling. P atmospheric pressure ( kPa ); H The height of the air chamber inside the box ( cm The calculation of N2O gas emission flux was performed using Excel, and the process is shown in Table 1 below.
[0022] Table 1 Calculation data of N2O gas emission flux
[0023] Note: The slope of the regression curve of gas concentration c in the static chamber versus time t. dc / dt It is -0.1604.
[0024] Table x The time points are 0-30 minutes. y The N2O gas concentrations measured at each time point are as follows. ρ , H , T The values represent the density of N2O gas under standard conditions, the average sampling temperature, and the height of the sampling device. F Import the emission flux formula into Excel to calculate the N2O emission flux for the corresponding sample.
[0025] Step 2: Constructing Spectral Indices In this invention, the spectral index construction method in step two is a dual-band spectral index construction, that is, using any two spectral bands to construct the spectral index, including... R , SR , D , ND , ID Index, Index DDn , mSR1 , mSR2 , mND , mID .
[0026] Step 3: Construction of the N2O Gas Monitoring Model Using spectral indices constructed through pairwise random combinations of dual-band spectra as independent variables and the N2O gas emission flux corresponding to each sample as the dependent variable, the training and validation sets were divided according to a 30% data ratio. An N2O gas monitoring model was established using the Random Forest (RF) machine learning algorithm. Five-fold cross-validation was incorporated during the modeling process to find the optimal modeling parameters. The final output shows the N2O gas monitoring results for both spectral index construction methods. The optimal index is shown below. Figure 3 As shown.
[0027] Combination Figure 2 and Figure 3 For SR spectral indices, the optimal band combination for inversion is to use λ1 near 940 nm and λ2 near 1420 nm. 2 The highest value is 0.645. For the D-type spectral index, the optimal band combination for inversion is to use λ1 near 550 nm and λ2 near 2150 nm, R 2 The highest value is 0.696. For the ID spectral index type, the optimal band combination for inversion is to use λ1 near 400 nm and λ2 near 550 nm, R 2 The highest value can reach 0.701.
[0028] For the mSR2 spectral index, the optimal band combination for inversion is to use λ1 near 900 nm and λ2 near 1210 nm, R 2 The highest value is 0.643. For the ND spectral index type, the optimal band combination for inversion is to use λ1 near 1300 nm and λ2 near 1700 nm, R 2 The highest value is 0.603.
[0029] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.
Claims
1. A ground-based hyperspectral method for monitoring the greenhouse gas N2O in open fields, characterized in that: The steps of the method are as follows: Step 1: Data Acquisition and Preprocessing Hyperspectral data and corresponding gas samples were collected from field sampling points. The hyperspectral data were preprocessed and noise was reduced to obtain the measured spectral data. The gas samples were analyzed and calculated in the laboratory to obtain the measured N2O gas emission flux data. Step 2: Constructing Spectral Indices The spectral index is constructed using the spectral data after preprocessing and noise reduction in step one. The spectral index is constructed using a dual-band method, which involves using any two spectral bands to construct the spectral index. Step 3: Construction of the N2O Gas Monitoring Model Using the dual-band spectral index as the independent variable and the N2O gas emission flux corresponding to each sample as the dependent variable, the training set and validation set were divided according to the data ratio of 25%-45%. The N2O gas monitoring model was established by the random forest (RF) machine learning algorithm, and the results of N2O gas monitoring were output by the spectral indices obtained by different spectral index construction methods.
2. The ground-based hyperspectral method for monitoring greenhouse gas N2O in open fields according to claim 1, characterized in that: In step two, the spectral index includes a given wavelength. R Simple comparison SR Wavelength difference D Normalized difference ND Deficit ID Index, dual differences DDn Correcting simple ratios mSR 1. Correction for simple ratio mSR; 2. Correction for normalized differences. mND Correcting inverse differences mID The formula for calculating the index is as follows: ; （1） ; （2） ; （3） ; （4） ； （5） in, For the spectral band, The spectral data step size is 10nm. It is a band spectral reflectance, It is a band Spectral reflectance.