Super-high-yield corn yield estimation method based on remote sensing time sequence curve characteristic parameters

Abstract

The invention discloses a super-high-yield corn yield estimation method based on remote sensing time sequence curve characteristic parameters, belongs to the technical field of corn yield estimation, is used for corn yield estimation, and comprises the following steps: dividing an NDVI time sequence curve of corn into a growth stage and a decline stage, fitting the two stages into two Logistic curves respectively, and selecting parameters with high correlation in the two Logistic curves as input of a machine learning regression model, and predicting the yield. According to the method, the obvious difference rule of the spectrum saturation time not reached in the corn growth period is fully utilized, the saturation limitation of a yield estimation model is effectively overcome, and a stable and reliable method is provided for accurately estimating the super-high yield part of the corn.

Description

Technical Field

[0001] The invention discloses a super-high-yield corn yield estimation method based on characteristic parameters of a remote sensing time series curve, and belongs to the technical field of corn yield estimation. Background Art

[0002] Spectral feature analysis and yield modeling of super-high-yield corn can achieve accurate monitoring of super-high-yield corn growth and field-scale yield estimation, ensuring stable grain production and increased income, which is an effective support for agricultural production management and an urgent need for the development of precision agriculture. At present, the methods of crop yield remote sensing estimation are mainly divided into three categories: 1. Empirical statistical model, which estimates the linear relationship between the predicted variable and crop yield in a given data set, with simple calculation and strong explanatory power. 2. Data assimilation model, which integrates remote sensing information into the crop growth model to minimize the difference between remote sensing observations and model variables. 3. Semi-mechanistic model, which simplifies parameter input relative to the mechanism model. Comprehensive analysis of the above studies shows that the existing technology has not solved the problem that the existing corn yield estimation model is prone to saturation in the high-yield stage, and most of the models are based on a single growth period for yield prediction. There are few studies on corn yield estimation using time series data to comprehensively extract spectral features of the entire growth period. Summary of the invention

[0003] The purpose of the present invention is to provide a super-high-yield corn yield estimation method based on characteristic parameters of remote sensing time series curves, so as to solve the problem of inaccurate corn yield estimation in the prior art.

[0004] A super-high-yield corn yield estimation method based on the characteristic parameters of remote sensing time series curves, including dividing the NDVI time series curve of corn into a growth stage and a decline stage, fitting the two stages into two Logistic curves respectively, and selecting the parameters with high correlation in the two Logistic curves as the input of the machine learning regression model to predict the yield.

[0005] The logistic curve includes: ; In the formula, is the upper limit of the Logistic curve, indicating the maximum NDVI value reached by corn during the growing season. is the logistic curve The midpoint of the corn growing season corresponds to the middle of the corn growing season. is the growth rate, which determines the Logistic curve The steepness of the surrounding terrain controls the rate of change of NDVI in the middle of the growing season. is the baseline offset value, which is used to adjust the vertical position of the Logistic curve and represents the basic NDVI value of corn in the non-growing season. is the fitted NDVI, is a natural constant, It is the cumulative day of the corn growing period.

[0006] The parameters of the two logistic curves include , , , , , , , ,in , , , The growth phase , , , , , , , The descending phase , , , .

[0007] The parameters with high correlation in the two Logistic curves are , , , , .

[0008] When using a machine learning regression model, the training set and validation set are divided by traversing k-fold cross validation to obtain the fold division.

[0009] Before dividing the NDVI time series curve, the NDVI time series curve is smoothed.

[0010] Use linear interpolation method to smooth the NDVI time series curve: ; ; In the formula, is the coordinate of the point to be interpolated on the NDVI time series curve, and are two known data points on the NDVI time series curve.

[0011] Use the exponentially weighted moving average method to smooth the NDVI time series curve: ; ; In the formula, yes The smoothed value at time, is the smoothing factor of the exponentially weighted moving average, It's time The original observation value of yes The smoothed value at time, Represents the window size of the exponentially weighted moving average.

[0012] Compared with the prior art, the present invention has the following beneficial effects: the present invention makes full use of the significant difference law of the time when the spectrum saturation is not reached during the growth period of corn, effectively overcomes the saturation limitation of the yield estimation model, and provides a stable and reliable model and method for accurate estimation of corn yield in the super-high-yield part. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a technical flow chart of the present invention; Figure 2 This is a schematic diagram of logistic curve fitting; Figure 3 Scatter plot for accuracy verification. DETAILED DESCRIPTION

[0014] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0015] A super-high-yield corn yield estimation method based on the characteristic parameters of remote sensing time series curves, including dividing the NDVI time series curve of corn into a growth stage and a decline stage, fitting the two stages into two Logistic curves respectively, and selecting the parameters with high correlation in the two Logistic curves as the input of the machine learning regression model to predict the yield.

[0016] The logistic curve includes: ; In the formula, is the upper limit of the Logistic curve, indicating the maximum NDVI value reached by corn during the growing season. is the logistic curve The midpoint of the corn growing season corresponds to the middle of the corn growing season. is the growth rate, which determines the Logistic curve The steepness of the surrounding terrain controls the rate of change of NDVI in the middle of the growing season. is the baseline offset value, which is used to adjust the vertical position of the Logistic curve and represents the basic NDVI value of corn in the non-growing season. is the fitted NDVI, is a natural constant, It is the cumulative day of the corn growing period.

[0017] The parameters of the two logistic curves include , , , , , , , ,in , , , The growth phase , , , , , , , The descending phase , , , .

[0018] The parameters with high correlation in the two Logistic curves are , , , , .

[0019] When using a machine learning regression model, the training set and validation set are divided by traversing k-fold cross validation to obtain the fold division.

[0020] Before dividing the NDVI time series curve, the NDVI time series curve is smoothed.

[0021] Use linear interpolation method to smooth the NDVI time series curve: ; ; In the formula, is the coordinate of the point to be interpolated on the NDVI time series curve, and are two known data points on the NDVI time series curve.

[0022] Use the exponentially weighted moving average method to smooth the NDVI time series curve: ; ; In the formula, yes The smoothed value at time, is the smoothing factor of the exponentially weighted moving average, It's time The original observation value of yes The smoothed value at time, Represents the window size of the exponentially weighted moving average.

[0023] The main defects of the existing technology are: the spectral reflectance saturation effect in densely vegetated areas is one of the main contemporary problems facing the remote sensing community. The existing remote sensing yield estimation model is prone to saturation in the high-yield stage and is prone to underestimate the high-yield part. Crop growth is an allometric growth process. The existing remote sensing yield estimation models are either limited to a single stage or have unstable performance across stages. They often ignore the fact that crops have significant differences in different growth stages, lack mobility throughout the growth period, and lack research on using time series data to comprehensively extract spectral features throughout the growth period to estimate crop yields.

[0024] The present invention comprehensively utilizes remote sensing image data and agricultural knowledge, constructs daily vegetation index time series data in the whole growth period of corn by linear interpolation and exponential weighted moving average smoothing methods, analyzes in detail the difference changes of vegetation index time series curves of corns with different yields, divides the time series curves into a growth period and a decline period, and explores the key parameters hidden in the vegetation index time series curves of corns with different yields that can well describe the dynamic changes of the curves by fitting them into two Logistic curves, takes the parameters with high correlation with yield as the input of the regression method, constructs a yield estimation model that can solve the problem that the common corn yield estimation model is easy to saturate in the high-yield part and considers the spectral time series characteristics of corn in the whole growth period, and realizes the spectral time series characteristic analysis and yield modeling research of super-high-yield corn.

[0025] Based on the existing remote sensing platform and sensor, the spectral information of corn during its growth period is obtained, and the spectral information is preprocessed. The presence of clouds and cloud shadows will cause the distortion of the spectral information on the ground, which seriously affects the accuracy of the reconstruction of the vegetation index time series. Therefore, most time series applications of optical satellite data require accurate cloud and shadow masks. The present invention takes the dataset products and satellite data available during the entire growth period of corn as an example. The dataset product removes clouds through its own bands, classifies clouds, cloud shadows, water and snow based on a set of top atmospheric reflectance thresholds, and defines cloud masks: pixels with median values ​​of 4 (shadow) and 5 (cloud) are clouds, and all images in the time series dataset are traversed to achieve cloud removal. The surface reflectance dataset of the acquired satellite data is declouded by combining multiple cloud detection indicators. First, the normalized water index NDWI is defined, and the cloud probability dataset is obtained. Then, the data quality assessment QA60 band is called, and the pixels with NDWI less than 0.2 and cloud probability less than 50 and QA60 marked non-cloud pixels are the final required pixels.

[0026] The estimation of vegetation characteristics is achieved by using the vegetation index VI derived from remote sensing data obtained from satellites, airborne or ground platforms. VI is calculated based on the red and near-infrared parts of the electromagnetic spectrum and has been widely used in qualitative and quantitative remote sensing monitoring of vegetation vitality and growth dynamics. The present invention takes the normalized difference vegetation index NDVI as an example. NDVI is an effective indicator for monitoring vegetation growth and changes. By analyzing the NDVI time series, the changes in the corn growing season can be observed, including the beginning of growth, peak growth and fall, which are closely related to corn yield. The NDVI time series can also be used to extract the characteristics of the growing season, such as growth rate, growth duration and seasonal changes in growth. These characteristics can be used to analyze the growth of corn in different regions or different years, and then predict the changing trend of corn yield.

[0027] When processing vegetation index time series data, cloud detection algorithms may result in some null values. In order to repair these null values, commonly used techniques include interpolation and smoothing, which are used to fill missing values, smooth noise or adjust the frequency of data, thereby generating time-continuous and spatially complete daily vegetation index time series data for the entire growth period of corn. The present invention takes the use of linear interpolation method and exponential weighted moving average smoothing method to construct vegetation index time series as an example. Linear interpolation is a simple and commonly used interpolation method, which is suitable for processing time series data due to its simplicity, computational efficiency and accuracy in processing steadily changing data. In linear interpolation, it is assumed that the data change between two adjacent data points is linear, and then this linear relationship is used to infer the value of the unknown point. Exponential weighted moving average EWMA is a commonly used time series data smoothing method, which is often used to eliminate noise and fluctuations in data in order to better observe the trend of data. EWMA estimates the mean of the data by weighted averaging the time series data, giving higher weights to recent data and lower weights to long-term data.

[0028] The present invention takes the NDVI time series curve as an example, and the overall trend of the corn NDVI time series curve of different yield levels is consistent, showing an increase before the end of the jointing stage, a decrease after the early waxy stage, and a high-level NDVI saturation platform is maintained between the end of the jointing stage and the early waxy stage. In the growth stage of the NDVI time series curve, the higher the yield level, the greater the NDVI value of the same date, the more violent the degree of increase after entering the jointing stage, and the greater the slope. In the high-level stage of the NDVI time series curve, corn is in the large trumpet stage and the silking stage, and the flowering and pollination stages are carried out, and the NDVI value reaches a saturation level at this time. In the descending stage of the NDVI time series curve, the higher the yield level, the greater the NDVI value of the same date, the more violent the degree of reduction after entering the maturity stage, and the greater the slope.

[0029] The present invention uses remote sensing technology to predict corn yield, which is of great significance for ensuring food security, promoting the development of corn yield prediction models, and guiding agricultural production. However, most of the existing models are affected by the spectral saturation effect, and cannot achieve accurate estimation in the super-high yield part, and do not consider the spectral time series characteristics of the whole growth period of corn well. The key point of the present invention is: comprehensively consider the spectral time series characteristics of the whole growth period of corn growth, analyze the dynamic difference of the corn vegetation index time series curve of different yield levels in detail, and extract the characteristic parameters describing the curve limit, slope, inflection point, and offset by fitting the Logistic curve, and establish a connection with the yield to predict the yield. The super-high-yield corn remote sensing monitoring model constructed by the present invention combines the difference law of the corn vegetation index time series curve and the spectral saturation law, has a certain mechanism, can eliminate the inaccurate estimation problem caused by not considering the difference of corn in different growth stages, can be used for the yield prediction of super-high-yield corn, and can effectively solve the problem that super-high-yield corn is easily underestimated. The method uses the characteristic parameters that can effectively describe the slope and inflection point of the time series curve as input variables, with high reliability and strong flexibility. By combining multi-source remote sensing data to achieve accurate estimation of corn yield in the super-high-yield stage, it will help promote the development of corn yield prediction models and achieve timely and accurate prediction of super-high-yield corn yields, which is of great significance for the scientific formulation of import and export decisions, grain market prices and trade, agricultural insurance assessment, and smart agriculture applications.

[0030] The technical process of the present invention is as follows Figure 1 As shown in the figure, the remote sensing data and measured yield data are collated to complete the data collection, and the data preprocessing, time series data calculation, linear interpolation and exponential weighted moving average are performed in sequence to complete the vegetation index time series construction. Then, the time series curve difference analysis, Logistic curve fitting and characteristic parameter extraction, regression analysis and k-fold cross validation are performed in sequence to complete the characteristic parameter advance and modeling.

[0031] The logistic curve fitting schematic diagram of the present invention is as follows Figure 2 As shown, Figure 2 In this paper, NDVI was cut in the middle and fitted separately to obtain 8 fitting parameters. Then, correlation analysis was performed to select 5 fitting parameters for corn yield estimation. The accuracy verification scatter plot is shown in Figure 2. Figure 3 As shown in the figure, the oblique line represents the straight line that the predicted output and the measured output are completely consistent. According to the scatter plot, it can be seen that the prediction results of the present invention are basically attached to both sides of the oblique line. After calculation, the square residual coefficient R 2 =0.72, root mean square error RMSE=1087.22kg / ha.

[0032] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features may be replaced by equivalents, and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A super high yield corn yield estimation method based on characteristic parameters of remote sensing time series curve, characterized in that: It includes dividing the NDVI time series curve of corn into a growth stage and a decline stage, fitting the two stages into two Logistic curves respectively, and selecting the parameters with high correlation in the two Logistic curves as the input of the machine learning regression model to predict the yield.

2. The method for estimating super-high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 1, characterized in that: The logistic curve includes: ； In the formula, is the upper limit of the Logistic curve, indicating the maximum NDVI value reached by corn during the growing season. is the logistic curve The midpoint of the corn growing season corresponds to the middle of the corn growing season. is the growth rate, which determines the Logistic curve The steepness of the surrounding terrain controls the rate of change of NDVI in the middle of the growing season. is the baseline offset value, which is used to adjust the vertical position of the Logistic curve and represents the basic NDVI value of corn in the non-growing season. is the fitted NDVI, is a natural constant, It is the cumulative day of the corn growing period.

3. The method for estimating super-high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 2, characterized in that: The parameters of the two logistic curves include , , , , , , , ,in , , , The growth phase , , , , , , , The descending phase , , , .

4. The method for estimating super-high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 3, characterized in that: The parameters with high correlation in the two Logistic curves are , , , , .

5. The method for estimating super-high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 4, characterized in that: When using a machine learning regression model, the training set and validation set are divided by traversing k-fold cross validation to obtain the fold division.

6. The method for estimating super-high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 5, characterized in that: Before dividing the NDVI time series curve, the NDVI time series curve is smoothed.

7. The method for estimating super-high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 6, characterized in that: Use linear interpolation method to smooth the NDVI time series curve: ； ； In the formula, is the coordinate of the point to be interpolated on the NDVI time series curve, and are two known data points on the NDVI time series curve.

8. The method for estimating super high-yield corn yield based on characteristic parameters of remote sensing time series curve according to claim 6, characterized in that: Use the exponentially weighted moving average method to smooth the NDVI time series curve: ； ； In the formula, yes The smoothed value at time, is the smoothing factor of the exponentially weighted moving average, It's time The original observation value of yes The smoothed value at time, Represents the window size of the exponentially weighted moving average.

Images