A rice leaf roller early warning method based on SIF and CPD-Logistic

Abstract

The application discloses a rice leaf roller early warning method based on SIF and CPD-Logistic, comprising the following steps: continuously monitoring the rice canopy spectrum by using a fluorescence observation system, acquiring SIF data and performing abnormal detection and correction processing; selecting SIF data of a healthy growth stage of rice, and establishing a healthy physiological benchmark; time integrating the difference between the smoothed SIF data of the stress period and the benchmark period, and constructing a cumulative physiological deficit; constructing a cumulative physiological deficit-Logistic coupling model, fitting the degree of deviation of the observed value from the benchmark value, and calculating a stress index; according to the stress index and the growth trend of the cumulative physiological deficit, a rice pest and disease grading early warning mechanism is constructed and early warning information is output. The application can effectively identify the early physiological abnormalities of the rice leaf roller by combining the SIF data and the cumulative physiological deficit index, and can depict the stress evolution process through the Logistic model, thereby improving the accuracy and stability of the early warning.

Description

Technical Field

[0001] This invention relates to the field of agricultural remote sensing and plant pathology monitoring technology, and in particular to a method for early warning of rice leaf roller based on SIF and CPD-Logistic. Background Technology

[0002] Rice leaf rollers are one of the major insect pests in rice production. The larval stage is the most damaging, as they feed on the leaf mesophyll tissue, causing leaf curling, chlorosis, and even drying, significantly reducing the effective photosynthetic area and weakening photosynthetic capacity. In severe infestations, they also damage the canopy structure, affecting the efficiency of light capture and utilization, thus having a sustained negative impact on rice growth and development. Due to their high concealment and rapid spread, once obvious symptoms appear in the field, the pest has usually entered the middle to late stages, significantly increasing the difficulty of control. Currently, monitoring of rice leaf rollers mainly relies on manual field surveys, including observing the degree of leaf curling, counting insect population density, and judging the pest severity based on experience. These methods suffer from low efficiency, high subjectivity, and insufficient spatial representativeness. Furthermore, they often only identify pests after symptoms appear, lacking the ability to detect early stages of the infestation, resulting in a significant lag in control measures and failing to meet the demands of modern agriculture for real-time and precise monitoring.

[0003] With the development of remote sensing technology, using crop spectral information to invert physiological states has gradually become an important research direction. Among them, solar-induced chlorophyll fluorescence (SIF), as an important signal reflecting the photosynthetic process of plants, can directly characterize the energy distribution state of photosystem II (PSII) and is considered a key indicator connecting the light reaction and carbon assimilation processes. Compared with traditional vegetation indices (such as NDVI, EVI, etc.), which mainly reflect changes in leaf structure or chlorophyll content, SIF originates from the fluorescence release process generated by energy redistribution during photosynthesis and is more sensitive to plant physiological stress, especially in the early stages of stress, often responding before changes in leaf structure. However, existing SIF-based crop stress monitoring methods mostly rely on single-moment or daily average data for analysis, focusing on the comparison of instantaneous differences and ignoring the cumulative effect of stress intensity and duration. This makes it difficult to comprehensively depict the true physiological state of crops under continuous stress. At the same time, under natural observation conditions, SIF signals are easily affected by factors such as changes in solar radiation, cloud cover fluctuations, and observation geometry, exhibiting strong short-term fluctuation characteristics. Without effective data processing and feature extraction mechanisms, environmental noise can easily be misjudged as physiological abnormalities, reducing the reliability of monitoring results.

[0004] Furthermore, the occurrence and development of rice leaf rollers exhibit distinct phases and nonlinear characteristics. From initial slight feeding to rapid spread in the middle stage and concentrated outbreak in the later stage, the impact of rice on physiological functions shows a trend of accumulation and accelerated evolution over time. However, existing technologies mostly employ linear analysis or empirical threshold determination methods, lacking the ability to dynamically model the stress evolution process and making it difficult to identify key inflection points in stress occurrence, thus limiting the improvement of early warning capabilities.

[0005] In summary, existing technologies still have significant shortcomings in comprehensively characterizing stress intensity and duration, effectively utilizing high-frequency time series data, modeling nonlinear evolution processes, and resisting environmental interference. There is an urgent need for a new technical solution that can construct an indicator system reflecting cumulative physiological damage based on high temporal resolution SIF data, and combine it with a nonlinear model suitable for stress evolution processes to dynamically characterize the rice leaf roller, thereby achieving a highly sensitive and reliable early warning before the appearance of rice symptoms. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a simple, safe, and reliable early warning method for rice leaf rollers based on SIF and CPD-Logistics.

[0007] The technical solution of this invention to solve the above-mentioned technical problems is: a method for early warning of rice leaf roller based on SIF and CPD-Logistic, comprising the following steps:

[0008] Step 1: Continuously monitor the rice canopy spectrum using a fluorescence observation system to obtain time-series SIF data including the rice leaf roller stress period. The SIF data is sunlight-induced chlorophyll fluorescence data, and anomaly detection and correction processing are performed on the SIF data.

[0009] Step 2: Select SIF data from the healthy growth stage of rice to establish a healthy physiological baseline;

[0010] Step 3: Integrate the smoothed SIF data between the stress period and the baseline period over time to construct the cumulative physiological deficit (CPD);

[0011] Step 4: Construct a cumulative physiological deficit-logistic coupled model CPD-Logistic based on the cumulative physiological deficit (CPD) and logistic model, fit the degree of deviation of the observed values ​​from the baseline value, and calculate the stress index SI;

[0012] Step 5: Based on the growth trend of the stress index SI and the cumulative physiological deficit CPD, construct a graded early warning mechanism for rice diseases and pests and output early warning information.

[0013] The above-mentioned early warning method for rice leaf rollers based on SIF and CPD-Logistics, step 1 is as follows:

[0014] Step 11: Install a spectral sensing device above the rice canopy and adjust the probe of the spectral sensing device to be vertically pointed to the central canopy area of ​​the rice.

[0015] Step 12: Use a spectral sensing device to continuously collect daily downlink solar irradiance data and uplink ground object irradiance data throughout the entire rice growing season;

[0016] Step 13: Record and store the collected downlink irradiance data and uplink ground object irradiance data to the local storage unit;

[0017] Step 14: Use the spectral fitting and inversion algorithm based on the Fraunhofer dark line to calculate the downlink irradiance data and the uplink ground object irradiance data to obtain SIF data;

[0018] Step 15: Use the IQR interquartile range method to remove outliers from the SIF data, use the Loess local weighting method to smooth the valid SIF data, and use local quadratic polynomial weighted fitting to eliminate noise and retain the true trend of SIF changes over time.

[0019] The above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic, in step 14, the process of the spectral fitting inversion algorithm is expressed as follows:

[0020] ;

[0021] in, The uplink ground object radiance received by the spectral sensing device; Downward solar irradiance; Surface reflectance; Pi; To induce chlorophyll fluorescence under sunlight; Wavelength;

[0022] SIF data were obtained by fitting high-resolution spectra near the Fraunhofer dark line and separating them from the collected downlink irradiance data and uplink ground object irradiance data.

[0023] In the above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic, the formula for the interquartile range method in step 15 is as follows:

[0024] ;

[0025] in, Interquartile range; It is the first quartile; It is the third quartile;

[0026] The range for outlier determination is: ].

[0027] In the above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic, the formula for smoothing the effective SIF data using the Loess local weighting method in step 15 is as follows:

[0028] ;

[0029] in, for Smoothed SIF values ​​at time points; For the first The weights of the effective SIF data for each data point; For a specific moment; For the first Valid SIF data for each data point; This represents the number of data points participating in the local fitting.

[0030] ;

[0031] in, For the first The observation time of each data point; Indicates the width of a local window; when hour, .

[0032] In the above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic, step 2 involves extracting daily diurnal variation characteristic curves and fitting them to construct a SIF benchmark model representing the normal photosynthetic level, using the baseline period as the healthy growth stage of rice.

[0033] ;

[0034] in, for Reference SIF data at any given time; For a specific moment; For the base period in The average effective SIF data at time point; The base period; This is a locally weighted regression smoothing function.

[0035] The specific process of step 3 in the above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic is as follows:

[0036] Step 31: Match the smoothed SIF data of each day during the stress period with the baseline SIF data according to time period to obtain paired data under a unified time series.

[0037] Step 32: Calculate the difference between the smoothed SIF data and the baseline SIF data at each moment within a fixed time period each day during the stress period, and retain only the negative deviation, that is, the decrease caused by stress, to exclude the influence of enhanced light. Integrate the negative deviation over time to obtain the daily cumulative physiological deficit.

[0038] ;

[0039] in, To accumulate physiological deficit; SIF for the base period; The stress period SIF; For a moment.

[0040] The specific process of step 4 in the above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic is as follows:

[0041] Step 41: Normalize the daily cumulative physiological deficit;

[0042] ;

[0043] in, For the first The normalized stress index of the sky; For the coercion period The accumulated physiological deficit of the day; This represents the maximum cumulative physiological deficit. Index of the number of days in the period of coercion;

[0044] Step 42: Couple the normalized result of cumulative physiological deficit (CPD) with the logistic model to obtain the cumulative physiological deficit-logistic coupled model CPD-Logistic, thus realizing the coupled modeling of cumulative damage and development stage.

[0045] ;

[0046] in, This is the point in time when the stress index grows the fastest, which is the turning point when the stress effect accelerates. The rate of change;

[0047] Step 43: Evaluate the CPD-Logistic fit and select the coefficient of determination. The goodness or badness of model fitting is evaluated; simultaneously, a significance test (p-test) is used to verify the statistical reliability of the correlation between measured and predicted values; when When the value is >0.7 and p<0.05, the CPD-Logistic model is considered to have a good fit, and the time variation of the stress index is effectively characterized by CPD-Logistic.

[0048] ;

[0049] in, For the first The actual stress index value of the day; For the first The model predicts the stress index value for the day; The mean of all actual stress index values; n represents the total number of days.

[0050] In the above-mentioned rice leaf roller early warning method based on SIF and CPD-Logistic, step 5 involves the comprehensive stress index. Based on the CPD growth trend, a tiered early warning rule is constructed, preferably divided into:

[0051] when When the CPD value is greater than 0 and continues to increase, it is determined to be an outbreak of leafroller moth, and an early warning signal is generated;

[0052] When 0 < When the value is less than 0.3, it is considered mild stress.

[0053] When 0.3≤ When the stress level is less than 0.7, it is considered moderate stress.

[0054] when A value ≥0.7 is considered severe stress.

[0055] Based on the judgment results, the corresponding early warning level is output to realize the early identification, development monitoring and severity assessment of pests and diseases.

[0056] The beneficial effects of this invention are as follows: By combining high temporal resolution SIF data with cumulative physiological deficit indicators, this invention can effectively identify early physiological abnormalities in rice leaf rollers and improve the accuracy and stability of early warning by characterizing the stress evolution process using a Logistic model. This method is not only applicable to rice leaf rollers but can also be extended to monitoring various crop stress scenarios such as drought, high temperature, and nutrient stress by adjusting parameter ranges and threshold settings, thus broadening its applicability. Attached Figure Description

[0057] Figure 1 This is the overall flowchart of the present invention.

[0058] Figure 2 This is a time-series variation diagram of SIF obtained in an embodiment of the present invention.

[0059] Figure 3 This is a schematic diagram of the CPD change curve during the stress period in an embodiment of the present invention.

[0060] Figure 4 This is a graph showing the CPD-Logistic fitting evaluation results in an embodiment of the present invention. Detailed Implementation

[0061] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0062] like Figure 1 As shown, a method for early warning of rice leaf roller based on SIF and CPD-Logistic includes the following steps:

[0063] Step 1: Continuously monitor the rice canopy spectrum using a fluorescence observation system to obtain time-series SIF data including the rice leaf roller stress period. The SIF data is sunlight-induced chlorophyll fluorescence data, and anomaly detection and correction processing are performed on the SIF data.

[0064] Step 1 is as follows:

[0065] Step 11: Install a spectral sensing device 1.5-3m above the rice canopy and adjust the probe of the spectral sensing device to be vertically pointed to the central canopy area of ​​the rice.

[0066] Step 12: Use a spectral sensing device to continuously collect daily downlink solar irradiance data and uplink ground object irradiance data throughout the entire rice growing season.

[0067] The observation period will cover the rice leaf roller stress stage during the growing season, from July 15, 2024 to September 3, 2024, from 6:00 to 18:00 daily.

[0068] Step 13: Record and store the collected downlink irradiance data and uplink ground object irradiance data to the local storage unit.

[0069] Step 14: Use the spectral fitting and inversion algorithm based on the Fraunhofer dark line to calculate the downlink irradiance data and the uplink ground object irradiance data to obtain SIF data;

[0070] The process of the spectral fitting and inversion algorithm is expressed as follows:

[0071] ;

[0072] in, The uplink ground object radiance received by the spectral sensing device; Downward solar irradiance; Surface reflectance; Pi; To induce chlorophyll fluorescence under sunlight; Wavelength;

[0073] SIF data were obtained by fitting high-resolution spectra near the Fraunhofer dark line and separating them from the collected downlink irradiance data and uplink ground object irradiance data.

[0074] Step 15: Use the IQR interquartile range method to remove outliers from the SIF data, use the Loess local weighting method to smooth the valid SIF data, and use local quadratic polynomial weighted fitting to eliminate noise and retain the true trend of SIF changes over time.

[0075] The formula for the interquartile range method is:

[0076] ;

[0077] in, Interquartile range; It is the first quartile; It is the third quartile;

[0078] The range for outlier determination is: ].

[0079] The formula for smoothing effective SIF data using the Loess local weighting method is as follows:

[0080] ;

[0081] in, for Smoothed SIF values ​​at time points; For the first The weights of the effective SIF data for each data point; For a specific moment; For the first Valid SIF data for each data point; This represents the number of data points participating in the local fitting.

[0082] ;

[0083] in, For the first The observation time of each data point; Indicates the width of a local window; when hour, .

[0084] Step 2: Select SIF data from the healthy growth stage of rice to establish a healthy physiological baseline.

[0085] In step 2, taking the baseline period as the healthy growth stage of rice, the daily variation characteristic curves are extracted and fitted to construct a SIF baseline model representing the normal photosynthetic level:

[0086] ;

[0087] in, for Reference SIF data at any given time; For a specific moment; For the base period in The average effective SIF data at time point; The base period; This is a locally weighted regression smoothing function.

[0088] In this embodiment of the invention, step 2 involves using the SIF data processing method from step 1 to further obtain a time-series SIF variation diagram of rice from August 20, 2024 to August 28, 2024, as shown below. Figure 2 As shown, August 28, 2024, is the date on which the rice leaf roller pest was identified. The curve from August 20 to August 24, 2024, is stable and shows no obvious abnormalities. However, the curve after that shows significant stress characteristics. The signal is stable and has strong anti-interference ability during the 10:00-14:00 period. Therefore, the SIF observation data from 10:00 to 14:00 every day during the period from August 20 to August 24, 2024, is selected as the healthy growth stage of rice, and preprocessed and curve fitted.

[0089] Step 3: Integrate the smoothed SIF data between the stress period and the baseline period over time to construct the cumulative physiological deficit (CPD).

[0090] The specific process of step 3 is as follows:

[0091] Step 31: Select SIF observation data from August 25, 2024 to August 28, 2024 during the stress period, and match the smoothed SIF data of each day with the baseline SIF data according to time to obtain paired data under a unified time series.

[0092] Step 32: Calculate the difference between the smoothed SIF data and the baseline SIF data at each time point during the period from 10:00 to 14:00 every day during the stress period, and retain only the negative deviation, that is, the decrease caused by stress, to exclude the influence of light enhancement. Integrate the negative deviation in the time dimension to obtain the daily cumulative physiological deficit.

[0093] ;

[0094] in, To accumulate physiological deficit; SIF for the base period; The stress period SIF; For a moment.

[0095] A schematic diagram of CPD changes during the stress period is shown below. Figure 3 As shown, the dashed line " The figure shows the theoretical baseline value under no-stress conditions, while the four solid lines of different colors represent the diurnal variation curves of SIF for four consecutive days from August 25th to 28th, 2024. The shaded areas in the figure correspond one-to-one with the colors of the diurnal variation curves of SIF, and their areas represent the CPD value of that day, representing the negative deviation integral of the SIF curve from the baseline curve during the stress period. As the number of stress days increases, the shaded areas of each day overlap, and the color of the overlapping areas deepens, characterizing the cumulative effect of CPD over multiple days. The CPD value increases significantly with the passage of time, intuitively reflecting that as the stress period lengthens, the deviation between the SIF value and the baseline value during the stress period becomes larger and larger, especially the severe "sinking" that occurs at noon, indicating that the photosynthetic capacity of the plant is continuously and severely inhibited.

[0096] Step 4: Construct a cumulative physiological deficit-logistic coupled model CPD-Logistic based on the cumulative physiological deficit (CPD) and logistic model, fit the degree of deviation of the observed values ​​from the baseline value, and calculate the stress index SI.

[0097] The specific process of step 4 is as follows:

[0098] Step 41: Normalize the daily cumulative physiological deficit;

[0099] ;

[0100] in, For the first The normalized stress index of the sky ; For the coercion period The accumulated physiological deficit of the day; This represents the maximum cumulative physiological deficit. Index of the number of days in the period of coercion;

[0101] Step 42: Couple the normalized result of cumulative physiological deficit (CPD) with the logistic model to obtain the cumulative physiological deficit-logistic coupled model CPD-Logistic, thus realizing the coupled modeling of cumulative damage and development stage.

[0102] ;

[0103] in, This is the point in time when the stress index grows the fastest, which is the turning point when the stress effect accelerates. The rate of change;

[0104] Step 43: Evaluate the CPD-Logistic fit and select the coefficient of determination. The goodness or badness of model fitting is evaluated; simultaneously, a significance test (p-value) is used to verify the statistical reliability of the correlation between measured and predicted values; when When the value is >0.7 and p<0.05, the CPD-Logistic model is considered to have a good fit, and the time variation of the stress index is effectively characterized by CPD-Logistic.

[0105] ;

[0106] in, For the first The actual stress index value of the day; For the first The model predicts the stress index value for the day; The mean of all actual stress index values; n represents the total number of days.

[0107] The CPD-Logistic model fit evaluation results are as follows: Figure 4 As shown, the green curve is the CPD-Logistic model fitting curve, reflecting the "S"-shaped growth trend of the stress index SI with the increase of stress days; the green circle is the actual measured stress index value, obtained by normalization based on the daily cumulative physiological deficit CPD during the stress period. The results show that with the increase of stress days, the stress index SI of rice exhibits a typical "S"-shaped growth trend, that is, slow growth in the early stage, rapid increase in the middle stage, and saturation in the later stage, with extremely high model fit. =0.966, p<0.05), proving that the model can accurately describe and predict the deterioration of the physiological state of plants under continuous stress, providing a reliable mathematical basis for disaster early warning in precision agriculture.

[0108] Step 5: Based on the growth trend of the stress index SI and the cumulative physiological deficit CPD, construct a graded early warning mechanism for rice diseases and pests and output early warning information.

[0109] Comprehensive Stress Index Based on the CPD growth trend, a tiered early warning rule is constructed, preferably divided into:

[0110] when When the CPD value is greater than 0 and continues to increase, it is determined to be an outbreak of leafroller moth, and an early warning signal is generated;

[0111] When 0 < When the value is less than 0.3, it is considered mild stress.

[0112] When 0.3≤ When the stress level is less than 0.7, it is considered moderate stress.

[0113] when A value ≥0.7 is considered severe stress.

[0114] Based on the judgment results, the corresponding early warning level is output to realize the early identification, development monitoring and severity assessment of pests and diseases.

[0115] This invention combines high temporal resolution SIF data with cumulative physiological deficit indicators to effectively identify early physiological abnormalities in rice leaf rollers. Furthermore, it uses a logistic model to characterize the stress evolution process, improving the accuracy and stability of early warning systems. This method is not only applicable to rice leaf rollers but can also be extended to monitoring various crop stresses such as drought, high temperature, and nutrient stress by adjusting parameter ranges and threshold settings.

Claims

1. A method for early warning of rice leaf roller based on SIF and CPD-Logistic, characterized in that, Includes the following steps: Step 1: Continuously monitor the rice canopy spectrum using a fluorescence observation system to obtain time-series SIF data including the rice leaf roller stress period. The SIF data is sunlight-induced chlorophyll fluorescence data, and anomaly detection and correction processing are performed on the SIF data. Step 2: Select SIF data from the healthy growth stage of rice to establish a healthy physiological baseline; Step 3: Integrate the smoothed SIF data between the stress period and the baseline period over time to construct the cumulative physiological deficit (CPD); Step 4: Construct a cumulative physiological deficit-logistic coupled model CPD-Logistic based on the cumulative physiological deficit (CPD) and logistic model, fit the degree of deviation of the observed values ​​from the baseline value, and calculate the stress index SI; Step 5: Based on the growth trend of the stress index SI and the cumulative physiological deficit CPD, construct a graded early warning mechanism for rice diseases and pests and output early warning information.

2. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 1, characterized in that, Step 1 is as follows: Step 11: Install a spectral sensing device above the rice canopy and adjust the probe of the spectral sensing device to be vertically pointed to the central canopy area of ​​the rice. Step 12: Use a spectral sensing device to continuously collect daily downlink solar irradiance data and uplink ground object irradiance data throughout the entire rice growing season; Step 13: Record and store the collected downlink irradiance data and uplink ground object irradiance data to the local storage unit; Step 14: Use the spectral fitting and inversion algorithm based on the Fraunhofer dark line to calculate the downlink irradiance data and the uplink ground object irradiance data to obtain SIF data; Step 15: Use the IQR interquartile range method to remove outliers from the SIF data, use the Loess local weighting method to smooth the valid SIF data, and use local quadratic polynomial weighted fitting to eliminate noise and retain the true trend of SIF changes over time.

3. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 2, characterized in that, In step 14, the process of the spectral fitting and inversion algorithm is expressed as follows: ； in, The uplink ground object radiance received by the spectral sensing device; Downward solar irradiance; Surface reflectance; Pi; To induce chlorophyll fluorescence under sunlight; Wavelength; SIF data were obtained by fitting high-resolution spectra near the Fraunhofer dark line and separating them from the collected downlink irradiance data and uplink ground object irradiance data.

4. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 2, characterized in that, In step 15, the formula for the interquartile range method is: ； in, Interquartile range; It is the first quartile; It is the third quartile; The range for outlier determination is: ].

5. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 2, characterized in that, In step 15, the formula for smoothing the effective SIF data using the Loess local weighting method is as follows: ； in, for Smoothed SIF values ​​at time points; For the first The weights of the effective SIF data for each data point; For a specific moment; For the first Valid SIF data for each data point; This represents the number of data points participating in the local fitting. ； in, For the first The observation time of each data point; Indicates the width of a local window; when hour, .

6. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 1, characterized in that, In step 2, taking the baseline period as the healthy growth stage of rice, the daily variation characteristic curves are extracted and fitted to construct a SIF baseline model representing the normal photosynthetic level: ； in, for Reference SIF data at any given time; For a specific moment; For the base period The average value of valid SIF data at any given time; The base period; This is a locally weighted regression smoothing function.

7. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 1, characterized in that, The specific process of step 3 is as follows: Step 31: Match the smoothed SIF data of each day during the stress period with the baseline SIF data according to time period to obtain paired data under a unified time series. Step 32: Calculate the difference between the smoothed SIF data and the baseline SIF data at each moment within a fixed time period each day during the stress period, and retain only the negative deviation, that is, the decrease caused by stress, to exclude the influence of enhanced light. Integrate the negative deviation over time to obtain the daily cumulative physiological deficit. ； in, To accumulate physiological deficit; SIF for the base period; The stress period SIF; For a moment.

8. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 1, characterized in that, The specific process of step 4 is as follows: Step 41: Normalize the daily cumulative physiological deficit; ； in, For the first The normalized stress index of the sky; For the coercion period The accumulated physiological deficit of the day; This represents the maximum cumulative physiological deficit. Index of the number of days in the period of coercion; Step 42: Couple the normalized result of cumulative physiological deficit (CPD) with the logistic model to obtain the cumulative physiological deficit-logistic coupled model CPD-Logistic, thus realizing the coupled modeling of cumulative damage and development stage. ； in, This is the point in time when the stress index grows the fastest, which is the turning point when the stress effect accelerates. The rate of change; Step 43: Evaluate the CPD-Logistic fit and select the coefficient of determination. The goodness or badness of model fitting is evaluated; simultaneously, a significance test (p-test) is used to verify the statistical reliability of the correlation between measured and predicted values; when When the value is >0.7 and p<0.05, the CPD-Logistic model is considered to have a good fit, and the time variation of the stress index is effectively characterized by CPD-Logistic. ； in, For the first The actual stress index value of the day; For the first The model predicts the stress index value for the day; The mean of all actual stress index values; n represents the total number of days.

9. The rice leaf roller early warning method based on SIF and CPD-Logistic as described in claim 1, characterized in that, In step 5, the comprehensive stress index Based on the CPD growth trend, a tiered early warning rule is constructed, preferably divided into: when When the CPD value is greater than 0 and continues to increase, it is determined to be an outbreak of leafroller moth, and an early warning signal is generated; When 0 < When the value is less than 0.3, it is considered mild stress. When 0.3≤ When the stress level is less than 0.7, it is considered moderate stress. when A value ≥0.7 is considered severe stress. Based on the judgment results, the corresponding early warning level is output to realize the early identification, development monitoring and severity assessment of pests and diseases.

Images