A strawberry gray mold disease early warning system and method based on multi-physiological index cooperation

Abstract

A strawberry gray mold early warning system and method based on multi-physiological index cooperation belong to the technical field of strawberry gray mold early warning, and solve the technical problem that the existing technology relies on a single data source, causing the model to be easily disturbed by environmental factors, and unable to comprehensively capture the early abnormal performance of gray mold in multiple physiological levels, thereby reducing the recognition accuracy of the asymptomatic latent period of strawberry gray mold. The data acquisition layer is used to collect fluorescence indicators, spectral indicators and photosynthetic indicators of strawberry plants; the feature processing layer is used to calculate the feature indexes corresponding to the fluorescence indicators, spectral indicators and photosynthetic indicators collected by the data acquisition layer respectively; the alarm judgment layer is used to judge the alarm levels corresponding to the feature indexes output by the feature processing layer respectively; and the decision fusion layer is used to make a fusion decision according to the alarm levels output by the alarm judgment layer to obtain a warning state. The present application is used to realize accurate and effective early warning of strawberry gray mold.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of strawberry gray mold early warning, and in particular to a strawberry gray mold early warning system and method based on multi-physiological index coordination. BACKGROUND

[0002] In strawberry production, strawberry gray mold is a common disease with the characteristics of fast onset, wide range and strong concealment. The current detection technology of strawberry gray mold has the following limitations:

[0003] 1. Single data source limitation: existing detection methods mostly rely on a single type of data, such as visible light image-based recognition methods, hyperspectral fluorescence imaging-based methods or odor volatile organic compound-based detection methods. These methods only reflect the health status of plants from a single perspective, and cannot fully capture the early abnormal performance of gray mold in multiple physiological levels;

[0004] 2. Insufficient early warning capability: traditional image recognition methods can only be recognized after visible symptoms appear, at which time the disease has already invaded the tissue and the best prevention and control opportunity has passed;

[0005] 3. Limited model generalization performance: models trained with a single data source are easily disturbed by environmental factors, such as gray mold recognition models based on temperature and humidity and image features, which have poor stability when the environment changes, while recognition methods based on knowledge distillation improve the level of model lightweight, but the feature extraction dimension is still single;

[0006] 4. Data asynchronous problem: existing multi-modal research lacks consideration of synchronous data collection, such as simply combining temperature and humidity with image features, without solving the problem of temporal and spatial inconsistency in different types of data collection, resulting in poor fusion effect;

[0007] 5. High computational complexity: some models have complex structures in pursuit of accuracy, such as the EFDDNN-EI system, which achieves high accuracy, but requires multiple levels of processing such as RBAEKF denoising, NTEWT feature extraction and CCHHO optimization, resulting in high computational cost and difficulty in deploying to field edge devices.

[0008] In the prior art, Chinese patent document CN112379711A discloses a "strawberry gray mold prediction system based on Internet of Things", which includes a strawberry gray mold germination preset unit, a greenhouse data acquisition unit, a condensation state detection unit, a data calculation unit, and a comparison and judgment unit. The strawberry gray mold germination preset unit is used to pre-set the gray mold germination probability threshold and the condensation occurrence duration threshold. The condensation state detection unit is used to detect the condensation occurrence duration. The data calculation unit is used to calculate the gray mold germination probability value according to the model formula. The comparison and judgment unit is used to compare the gray mold germination probability value with the gray mold germination probability threshold when the condensation occurrence duration value is greater than the condensation duration threshold. When the gray mold germination probability value is greater than the gray mold germination probability threshold, it is considered that the greenhouse will germinate gray mold. Through the prediction system, it can be determined that the strawberries in the greenhouse will germinate gray mold, and the temperature or humidity in the greenhouse needs to be adjusted to achieve the purpose of prevention. However, this technical solution relies on environmental factors such as temperature and humidity and condensation duration, which is an indirect prediction index, rather than direct physiological state feedback of the plant itself. Essentially, it belongs to "disease condition prediction" rather than "early disease detection", resulting in obvious lag in early warning. Usually, the alarm is given before the disease occurs on a large scale, missing the best early intervention window period. The environmental model is easily disturbed by factors such as microclimate fluctuations in the greenhouse, differences in sensor placement, and lack of specificity, which cannot distinguish between similar environmental changes caused by gray mold infection and other non-biological stress. Therefore, the false positive rate is high. In addition, this scheme lacks consideration of individual plant disease resistance differences, and cannot achieve precise diagnosis and early warning for individual plant health.

[0009] In summary, the prior art has the technical problem of reduced identification accuracy of the asymptomatic latent period of strawberry gray mold due to reliance on a single data source, which makes the model susceptible to environmental factors and unable to fully capture the early abnormal performance of gray mold at multiple physiological levels. SUMMARY

[0010] The present application solves the technical problem of reduced identification accuracy of the asymptomatic latent period of strawberry gray mold due to reliance on a single data source, which makes the model susceptible to environmental factors and unable to fully capture the early abnormal performance of gray mold at multiple physiological levels.

[0011] The strawberry gray mold early warning system based on multiple physiological indicators, according to the present application, comprises a data acquisition layer, a feature processing layer, an alarm judgment layer, a decision fusion layer, and an early warning output layer.

[0012] The data acquisition layer is used to acquire fluorescence indicators, spectral indicators, and photosynthetic indicators of strawberry plants.

[0013] The feature processing layer is configured to calculate feature indexes corresponding to the fluorescence indexes, the spectrum indexes and the photosynthesis indexes collected by the data collection layer respectively.

[0014] The alarm judgment layer is configured to judge alarm levels corresponding to the feature indexes output by the feature processing layer respectively.

[0015] The decision fusion layer is configured to make a fusion decision according to the alarm levels output by the alarm judgment layer to obtain a warning state.

[0016] Further, in an embodiment of the present application, when collecting the fluorescence indexes, the spectrum indexes and the photosynthesis indexes of the strawberry plants, a software time stamp alignment and a hardware synchronous trigger mechanism are adopted to ensure that the data collection devices are triggered at the same time, so that the data alignment is achieved.

[0017] Further, in an embodiment of the present application, the fluorescence indexes include a maximum photochemical efficiency Fv / Fm, an actual photochemical efficiency ΦPSII, a photochemical quenching coefficient qP and a non-photochemical quenching coefficient NPQ.

[0018] The spectrum indexes include a photochemical reflectance index PRI, a stress index SI and an improved red edge ratio mSR.

[0019] The photosynthesis indexes include a net photosynthetic rate Pn of the leaves, a stomatal conductance Gs, an intercellular CO2 concentration Ci and a transpiration rate Tr.

[0020] Further, in an embodiment of the present application, the feature indexes corresponding to the fluorescence indexes, the spectrum indexes and the photosynthesis indexes are a fluorescence decay index F_index, a spectrum stress index S_index and a photosynthesis inhibition index P_index respectively.

[0021] Further, in an embodiment of the present application, the alarm level corresponding to the fluorescence decay index is:

[0022] level 0 when F_index < 0.8, level 1 when 0.8 ≤ F_index < 1.0, level 2 when 1.0 ≤ F_index < 1.2, and level 3 when F_index ≥ 1.2.

[0023] The alarm level corresponding to the spectrum stress index is:

[0024] level 0 when S_index < 0.8, level 1 when 0.8 ≤ S_index < 1.0, level 2 when 1.0 ≤ S_index < 1.2, and level 3 when S_index ≥ 1.2.

[0025] ;

[0026] wherein,​​​​​​​​ a standard score of the current spectral stress index relative to the benchmark of the healthy population, a current spectral stress index, a mean value of the S_index of the healthy population, a standard deviation of the S_index of the healthy population;

[0027] the alert level corresponding to the photosynthetic inhibition index is:

[0028] level 0 when level 1 when level 2 when level 3 when .

[0029] Further, in an embodiment of the present application, the alert levels output by the alert judgment layer are fused for decision-making, specifically:

[0030] According to the fluorescence index, the spectral index and the photosynthetic index collected every day, an immune response cumulative value is obtained According to the immune response cumulative value and the alert level corresponding to the characteristic index, different early warning states are triggered.

[0031] Further, in an embodiment of the present application, the immune response cumulative value is specifically:

[0032]

[0033] wherein, is a decay factor, is the immune response cumulative value of the previous day, , and are alert levels corresponding to the characteristic indexes, , and are weights of the alert levels corresponding to the characteristic indexes, is a synergistic enhancement term.

[0034] Further, in an embodiment of the present application, when the alert levels corresponding to two or more characteristic indexes are all greater than or equal to 1, the synergistic enhancement term is triggered.

[0035] Further, in an embodiment of the present application, the triggering rule for triggering different early warning states is:

[0036] When the alert levels of the characteristic indexes corresponding to the fluorescence index, the spectral index and the photosynthetic index are all 0, the early warning state is state 0;

[0037] When the alarm level of the characteristic index corresponding to the fluorescence index is level 1 for two consecutive days, or the alarm levels of the characteristic indexes corresponding to the fluorescence index and the spectral index reach level 1 on the same day, or the cumulative value of the immune response is greater than or equal to 15, the early warning state is state 1;

[0038] When the alarm level of the characteristic index corresponding to the fluorescence index is level 2, and the alarm level of the characteristic index corresponding to the spectral index is greater than or equal to 1, or the cumulative value of the immune response is greater than or equal to 30, the early warning state is state 2;

[0039] When the alarm level of the characteristic index corresponding to the photosynthetic index is greater than or equal to 1, or the alarm level of any one of the characteristic indexes corresponding to the fluorescence index, the spectral index and the photosynthetic index is level 3, or the cumulative value of the immune response is greater than or equal to 50, the early warning state is state 3.

[0040] The strawberry gray mold early warning method based on the cooperation of multiple physiological indexes provided by the application can realize any of the systems described above, and comprises the following steps:

[0041] Step 1, pre-adapt the strawberry plant in a standard environment, acquire the multiple physiological indexes of the strawberry plant based on a software timestamp alignment and a hardware synchronization triggering mechanism;

[0042] Step 2, process the multiple physiological indexes of the strawberry plant respectively, and divide the processed multiple physiological indexes into alarm levels;

[0043] Step 3, acquire the cumulative value of the immune response;

[0044] Step 4, trigger different early warning states according to the cumulative value of the immune response and the divided alarm levels, and generate an early warning report.

[0045] The application solves the technical problem in the prior art that due to the dependence on a single data source, the model is easily disturbed by environmental factors, and the early abnormal performance of gray mold at the multiple physiological levels cannot be comprehensively captured, thereby reducing the recognition accuracy of the asymptomatic latent period of strawberry gray mold. The specific beneficial effects include:

[0046] 1. The application provides a strawberry gray mold early warning system based on the cooperation of multiple physiological indexes, which realizes a synergistic enhancement effect through the fusion of photosynthetic-spectral-fluorescence multiple physiological indexes, and the multiple modal data complement each other. When a certain data source is disturbed by the environment, the system can rely on other data sources to maintain stable performance, avoiding the problem that the model is easily disturbed by environmental factors due to the dependence on a single data source, and the early abnormal performance of gray mold at the multiple physiological levels cannot be comprehensively captured;

[0047] 2. The application provides a strawberry gray mold early warning system based on multi-physiological index cooperation, which can improve the identification accuracy of the asymptomatic latent period of strawberry gray mold by fusing decisions according to different alarm levels and immune response cumulative values, and has higher accuracy than single modal models, and can issue a warning 24-72 hours before the appearance of visible symptoms.

[0048] 3. The application provides a strawberry gray mold early warning system based on multi-physiological index cooperation, which can solve the problem of multi-source data space-time inconsistency through software timestamp alignment and hardware synchronous trigger mechanism, and the model adopts a lightweight design and can run in real time on mobile devices to meet the needs of large-scale field deployment. DETAILED DESCRIPTION

[0049] The above and / or additional aspects and advantages of the application will become apparent and be readily understood from the following description, taken in conjunction with the drawings, in which:

[0050] Figure 1 is a flow chart of the strawberry gray mold early warning system based on multi-physiological index cooperation according to embodiment one. DETAILED DESCRIPTION

[0051] Various embodiments of the application will be described below with reference to the accompanying drawings. The embodiments described with reference to the drawings are exemplary and are intended to explain the application, and cannot be understood as limiting the application.

[0052] Embodiment one: a strawberry gray mold early warning system based on multi-physiological index cooperation, comprising a data acquisition layer, a feature processing layer, an alarm judgment layer, a decision fusion layer and an early warning output layer.

[0053] The data acquisition layer is used to acquire fluorescence indexes, spectral indexes and photosynthetic indexes of strawberry plants.

[0054] The feature processing layer is used to calculate feature indexes corresponding to the fluorescence indexes, spectral indexes and photosynthetic indexes acquired by the data acquisition layer, respectively.

[0055] The alarm judgment layer is used to judge the alarm levels corresponding to the feature indexes output by the feature processing layer, respectively.

[0056] The decision fusion layer is used to make fusion decisions according to the alarm levels output by the alarm judgment layer to obtain the early warning state.

[0057] In the prior art, due to the dependence on a single data source, the model is prone to interference when the environment is affected, resulting in misjudgment, and cannot comprehensively capture the early abnormal performance of gray mold at multiple physiological levels, thereby reducing the identification accuracy of the asymptomatic latent period of strawberry gray mold, and the single data source has limitations, early warning capability is insufficient, data asynchrony, and the model has limited generalization performance.

[0058] To solve the above problems, the present embodiment provides a strawberry gray mold early warning system based on the cooperation of multiple physiological indicators. Based on the time sequence of disease development, i.e., the time sequence development of gray mold infection according to "cell level (fluorescence) -> tissue level (spectrum) -> overall function (photosynthesis)", the changes in different physiological indicators have correlation and mutual verification, and the disease stress has a cumulative effect, and the plant has self-repairing ability. Through the cooperative analysis and deep fusion of photosynthesis-spectrum-fluorescence multi-physiological indicators, the accurate early warning of strawberry gray mold is realized, and the technical problem of the prior art that the model is prone to interference due to the dependence on a single data source, which cannot comprehensively capture the early abnormal performance of gray mold at multiple physiological levels, thereby reducing the identification accuracy of the asymptomatic latent period of strawberry gray mold, is solved.

[0059] In summary, the present embodiment can capture the earliest signals of disease infection from cell function, tissue characteristics to overall physiological level by directly and synchronously monitoring three types of physiological indicators with clear biological significance, i.e., photosynthesis, spectrum and fluorescence, which realizes a fundamental change from "predicting the environment" to "diagnosing the plant itself", and effectively overcomes the defects of the prior art.

[0060] Embodiment two, the present embodiment is a further limitation of the strawberry gray mold early warning system based on the cooperation of multiple physiological indicators according to embodiment one. When collecting the fluorescence indicators, spectrum indicators and photosynthesis indicators of strawberry plants, software timestamp alignment and hardware synchronous triggering mechanism are used to ensure that the data collection equipment is triggered at the same time, and data alignment is realized.

[0061] In the prior art, there may be data asynchrony in multi-modal research, which leads to poor fusion effect of different types of data.

[0062] To solve the above technical problems, the present embodiment synchronously collects the multi-physiological indicator data of strawberry plants. The hardware synchronous triggering mechanism uses a master clock synchronization system to generate a high-precision synchronous electric pulse signal through a master control unit to simultaneously trigger three measurement devices to start collecting. After receiving the synchronous triggering signal, the chlorophyll fluorometer, hyperspectral camera and photosynthetic instrument start measuring at the same time within microseconds, thereby ensuring the time consistency of the data from the source.

[0063] The software timestamp alignment, on the basis of hardware synchronization, realizes system clock synchronization through network time protocol for each data point with millisecond-level precision. The data of photosynthetic indicators and fluorescence indicators are linearly interpolated based on the collection time of spectral indicators as the reference time sequence, to realize accurate alignment of the data.

[0064] In the embodiment, a standard operating procedure is formulated, and the strawberry plants are pre-adapted for 15-30 minutes in a standard light and temperature and humidity environment before measurement, so that the physiological state is stable. According to the standardized operation procedure, data acquisition is performed, a trigger signal is sent, and the master control unit generates a 5V TTL pulse signal. Under the synchronous trigger signal, the three devices start the measurement period at the same time, but due to the different measurement principles of each device, the chlorophyll fluorescence instrument immediately applies a 1-second saturated flash after receiving the trigger signal and immediately acquires fluorescence indicator data. The hyperspectral camera takes a one-time snapshot after receiving the trigger signal to collect spectral indicator data. The photosynthetic instrument synchronously starts a 30-second continuous measurement window after receiving the trigger signal to record the average value or dynamic curve of parameters such as Pn, Gs, Ci and Tr in the window. The measurement start time of all devices is strictly synchronized by the same trigger signal, and the entire measurement process is completed within 5 seconds to form a spatiotemporally aligned synchronous data package.

[0065] The embodiment adopts quality control measures to verify the effectiveness of the data, check whether the maximum photochemical efficiency Fv / Fm of the fluorescence indicators is within the normal range, confirm the consistency of the environment, and stabilize the CO2 concentration in the environment during measurement at 400±20 ppm. The fluctuation range of the readings of each sensor is monitored in real time, and abnormal data is automatically marked.

[0066] Embodiment three, the embodiment is a further limitation of the strawberry gray mold early warning system based on the cooperation of multiple physiological indicators according to embodiment one, wherein the fluorescence indicators include maximum photochemical efficiency Fv / Fm, actual photochemical efficiency ΦPSII, photochemical quenching coefficient qP and non-photochemical quenching coefficient NPQ.

[0067] The spectral indicators include photochemical reflectance index PRI, stress index SI and modified red edge ratio mSR.

[0068] The photosynthetic indicators include net photosynthetic rate Pn of leaves, stomatal conductance Gs, intercellular CO2 concentration Ci and transpiration rate Tr.

[0069] The selected fluorescence indicators, spectral indicators and photosynthetic indicators in the embodiment are based on their respective specificity, sensitivity and complementarity in revealing the physiological state of plants, which together form a monitoring network that can capture early signals of strawberry gray mold stress in all directions without omission.

[0070] Gray mold infection can directly or indirectly destroy the structure and function of chloroplast. Chlorophyll fluorescence is the most sensitive probe to detect the internal state of photosynthetic system. Among them, Fv / Fm is the mark of the intrinsic activity of PSII (photosynthesis system) reaction center. Any stress that affects the photosynthetic mechanism will lead to the decline of Fv / Fm. ΦPSII and qP both reflect the proportion of PSII reaction center actually used for photochemical reaction under environmental light, indicating the actual operation of photosynthetic function. NPQ reflects the ability of plants to dissipate excess light energy in the form of heat. At the early stage of stress, NPQ may increase for self-protection, and when the stress intensifies and the dissipation mechanism is damaged, NPQ will decrease, revealing the state of light protection ability of plants in response to stress.

[0071] Spectral indicators are used to capture biochemical and structural changes at the tissue level. Hyperspectral indices can reflect subtle changes in leaf internal chemical composition and cell structure, which occur earlier than visible symptoms. Among them, PRI is highly sensitive to changes in leaf carotenoids and can effectively reflect early fluctuations in photosynthetic light use efficiency. SI is sensitive to changes in leaf cell structure and water content. SI value will change significantly when gray mold causes tissue maceration and necrosis, and can be used to identify existing tissue lesions. mSR is highly correlated with chlorophyll content. Disease infection can lead to chlorophyll degradation, and mSR will decrease, directly reflecting the degree of loss of leaf photosynthetic pigments, which is an important indicator of stress severity.

[0072] Photosynthetic indicators are used to quantify the output of overall physiological function. Among them, Pn is a direct reflection of the carbon assimilation capacity of plants, and its decline indicates that the overall physiological function of the plant has been substantially damaged. Gs combined with Ci can determine the cause of photosynthetic decline. If Gs decreases and Ci decreases, it means that the main limitation comes from stomatal closure. If Gs decreases but Ci increases, it means that it is a non-stomatal factor, which is crucial for determining the specific physiological link of disease invasion. Tr reflects the water metabolism of the plant, and disease stress often disrupts water balance.

[0073] However, these indicators do not work independently, but through coordination, verification, and complementary mechanisms, they form a powerful diagnostic system, greatly improving the accuracy and reliability of early warning.

[0074] Early fluorescence indicators (especially Fv / Fm and ΦPSII) and spectral indicators (such as PRI) will first appear abnormal, which can capture subtle changes in cell function and light energy utilization, achieving early warning. In the middle stage, spectral indicators (such as SI and mSR) have obvious changes, reflecting changes in tissue structure and pigments, and photosynthetic indicators begin to fluctuate. In the later stage, photosynthetic indicators (Pn) show a significant decrease, indicating that yield has been affected. The present embodiment uses the sensitivity of early indicators to gain early warning time, and then uses the accuracy of middle and late indicators to verify the warning, avoiding false positives that may be caused by relying solely on early indicators. At the same time, there is complementarity of mechanisms and complementarity of information levels, cross-validation, reducing false positives, and panoramic diagnosis from micro to macro.

[0075] In summary, in the present embodiment, the fluorescence indicator, the spectral indicator, and the photosynthetic indicator have a multi-indicator synergistic enhancement effect. When the maximum photochemical efficiency Fv / Fm of the fluorescence indicator decreases slightly, if the net photosynthetic rate Pn of the photosynthetic indicator and the stomatal conductance Gs are abnormal and elevated at the same time, the confidence of the gray mold warning is greatly improved. This multi-angle verification mechanism significantly reduces false positives and false negatives, and improves the accuracy of the warning.

[0076] Embodiment Four, the present embodiment is a further limitation of the strawberry gray mold early warning system based on multi-physiological indicator synergy according to Embodiment One. The characteristic indexes corresponding to the fluorescence indicator, the spectral indicator, and the photosynthetic indicator are the fluorescence decay index F_index, the spectral stress index S_index, and the photosynthetic inhibition index P_index, respectively.

[0077] In the present embodiment, feature extraction is performed on the fluorescence indicator, the spectral indicator, and the photosynthetic indicator, respectively, to obtain the fluorescence decay index corresponding to the fluorescence indicator, the spectral stress index corresponding to the spectral indicator, and the photosynthetic inhibition index corresponding to the photosynthetic indicator.

[0078] The fluorescence decay index is the first line of defense, simulates the rapid stress response at the plant cell level, has the highest sensitivity, and is used to reflect whether the reaction center of the photosynthetic mechanism PSII is affected by disease. It is the most sensitive early indicator of stress, and the alarm level is divided according to its value (0-3 levels).

[0079]

[0080] The spectral stress index is the second line of defense, simulates the disease response at the plant tissue level, and has medium timeliness. The spectral stress index is a linear combination of PRI, SI, and mSR, which comprehensively characterizes the stress at the leaf tissue level. Plant disease will cause changes in chlorophyll degradation and cell structure damage in leaves, and usually appears abnormal after fluorescence and before photosynthesis, and the alarm level is divided according to the degree of deviation from the healthy baseline (0-3 levels).

[0081] S_index = 0.5 × PRI + 0.3 × SI + 0.2 × mSR;

[0082] The photosynthetic inhibition index is the third line of defense, simulates the decline of the overall physiological function of the plant, and is a marker of severe disease. It is used to reflect whether the overall carbon assimilation ability of the plant is substantially damaged, and is a marker of physiological function damage caused by disease. It is divided into warning levels (0-3 levels) according to its value.

[0083]

[0084] wherein, is the typical net photosynthetic rate value of the plant in the healthy period, is the current net photosynthetic rate value.

[0085] Embodiment five, this embodiment is a further limitation of the strawberry gray mold early warning system based on the coordination of multiple physiological indexes according to embodiment four, the warning level corresponding to the fluorescence decay index is:

[0086] is level 0 when is level 1 when is level 2 when is level 3 when

[0087] The warning level corresponding to the spectral stress index is:

[0088] is level 0 when is level 1 when is level 2 when is level 3 when

[0089] ;

[0090] wherein, is the standard score of the current spectral stress index relative to the benchmark of the healthy population, which is used to quantify the abnormal deviation degree of the spectral characteristics of the plant, is the current spectral stress index, is the mean of the S_index of the healthy population, is the standard deviation of the S_index of the healthy population;

[0091] The warning level corresponding to the photosynthetic inhibition index is:

[0092] is level 0 when is level 1 when is level 2 when is level 3 when is level 3 when​​

[0093] In this embodiment, the spectral data of at least 30 healthy strawberry plants are collected, and the mean value μ healthy of the S_index of the healthy population and the standard deviation σ healthy of the S_index of the healthy population are calculated.

[0094] In this embodiment, level 0 is normal, level 1 is a slight alarm, level 2 is a moderate alarm, and level 3 is a severe alarm.

[0095] In this embodiment, when , it indicates that the PSII is severely damaged.

[0096] In this embodiment, since Fv / Fm is a progressive decline based on physiological function, this embodiment takes 0.78 as the benchmark lower limit of the maximum photochemical efficiency of healthy plants. After dark adaptation, the Fv / Fm value of most unstressed plants stabilizes between 0.78-0.84. Taking this as the healthy baseline ensures the universality of the judgment. This embodiment takes 0.73 as the benchmark lower limit of level 1, which is about 5% lower than the healthy baseline 0.78. This degree of decline has exceeded the normal physiological fluctuation range of plants during the day, indicating that the energy conversion efficiency of the PSII reaction center has begun to appear detectable inhibition, which is the earliest and reversible signal of stress. This embodiment takes 0.65 as the benchmark lower limit of level 2, and when Fv / Fm is lower than this value, it indicates that the PSII reaction center has suffered severe and irreversible damage (inhibition degree exceeds 17%). This is usually accompanied by the destruction of photosynthetic organs and is a sign that the plant has fallen into severe physiological stress.

[0097] In this embodiment, the Z value is to compare the spectral state of the current plant with the pre-established "healthy population big data", and follow the statistical significance law under normal distribution. When the S_index value of the plant falls within the range of mean value ± 1 standard deviation (i.e. Z<1.0), this level is determined to be normal physiological fluctuation. When the deviation reaches 1-1.5 standard deviations, it has begun to deviate from the mainstream healthy population, which is defined as a slight abnormality, suggesting that attention is needed. Deviation reaches 1.5-2.0 standard deviations, indicating that the spectral characteristics of the plant have been significantly abnormal, and there is a significant difference from the healthy state. Deviation exceeds 2 standard deviations, with very high statistical significance, meaning that the spectral characteristics of the current plant is a small probability event, and it is determined to be a severe abnormality.

[0098] In this embodiment, P_index directly quantifies the degree of loss of Pn. 0.1 means that photosynthesis is inhibited by 10%, which may be due to the initial impact of slight environmental stress or disease, although it does not seriously affect growth, but it is an early signal that needs attention. 0.25 means that photosynthesis is inhibited by 25%, which indicates that the stress has had a substantial impact on the plant's carbon assimilation capacity, and the growth rate will slow down significantly, and the disease may have entered the development stage. 0.5 means that photosynthesis is inhibited by 50%, which means that the growth of the plant is almost stopped, and the physiological function is severely damaged, usually accompanied by the appearance of visible symptoms.

[0099] Embodiment six, this embodiment is a further limitation of the strawberry gray mold early warning system based on the coordination of multiple physiological indicators according to the first embodiment, the fusion decision of the alarm level output by the alarm judgment layer is specific to:

[0100] According to the immune response cumulative value obtained by the fluorescence index, the spectrum index and the photosynthetic index collected every day According to the alarm level corresponding to the immune response cumulative value and the characteristic index, different early warning states are triggered.

[0101] The immune response cumulative value is specific to:

[0102]

[0103] Among them, is the decay factor, is the immune response cumulative value of the previous day, , and are the alarm levels corresponding to the characteristic indexes, , and are the weights of the alarm levels corresponding to the characteristic indexes, is the synergistic enhancement term.

[0104] When the alarm levels corresponding to two or more characteristic indexes are greater than or equal to 1, the synergistic enhancement term is triggered.

[0105] This embodiment maintains an immune response cumulative value for each plant, introduces a decay factor and a synergistic enhancement term , that is, the current immune response cumulative value is equal to the immune response cumulative value of the previous time multiplied by the decay factor, plus the weighted sum of the alarm levels corresponding to the current defense lines and the synergistic enhancement term.

[0106] is the decay factor, representing the self-repairing or adaptive ability of the plant. The weight distribution follows the principle of early indicators first, with fluorescence indicators having the highest weight, followed by spectral indicators, and photosynthetic indicators having the lowest weight.

[0107] The synergistic enhancement term refers to an additional additive term k triggered when two or more defense lines simultaneously issue an alarm (level ≥ 1), simulating the enhanced effect of multi-system cooperative warning.

[0108] In one embodiment of the present embodiment, is 0.85, is 0.4, is 0.35, is 0.25.

[0109] In the present embodiment, when the alarm levels corresponding to the characteristic indices are all 0, the synergistic enhancement term When the alarm levels corresponding to the two characteristic indices are both greater than or equal to 1, the synergistic enhancement term When the alarm levels corresponding to the three characteristic indices are all greater than or equal to 1, the synergistic enhancement term .

[0110] In the present embodiment, λ is 0.85, indicating that the stress signal at the past time is transmitted to the current time with an intensity of 85%, simulating two physiological characteristics of plants:

[0111] 1. Stress memory effect: the response of plants to stress is cumulative and will not completely ignore the damage received before due to a short-term recovery. The decay factor ensures that persistent stress can be effectively accumulated and amplified;

[0112] 2. Self-repairing ability: 15% decay allows the immune response accumulation value of the plant to gradually fall after the stress is removed, thereby avoiding long-term false alarms due to a one-time accidental strong disturbance.

[0113] When is 0.85, the best balance between stress memory effect and self-repairing ability is achieved, allowing the model to focus on persistent threats and not overreact to temporary and recoverable environmental fluctuations (such as single-day strong light or low temperature), thereby significantly improving the stability and practicality of the warning.

[0114] In this embodiment, the fluorescence decay index weight w1 is the highest, the fluorescence index is the most sensitive and the earliest responding signal at the cell level, and it is given the highest weight to ensure the model's ability to detect the front of disease invasion. The spectral stress index weight w2 is lower than the fluorescence decay index weight, and the spectral index reflects the disease at the tissue level, which usually responds later than fluorescence but earlier than photosynthesis. As an important verification and medium-term signal, it is given a relatively low weight. The photosynthesis inhibition index weight w3 is the lowest, and the photosynthesis index is the output of the overall physiological function, which responds the most laggingly. However, its change marks substantial damage, and it is given a relatively low weight to prevent it from covering the weak signals of earlier stages, but at the same time, it is used as a key indicator of disease severity.

[0115] The weight distribution described in this embodiment makes the time focus of the early warning system move forward, and the alarm can be triggered earlier. It ensures the scientificity of the model decision, that is, the more early warning indicators, the greater the contribution to the final decision.

[0116] In this embodiment, the synergistic enhancement item simulates the phenomenon of multi-pathway synergistic amplification of alarm signals in the plant immune system. When multiple independent physiological dimensions appear abnormal at the same time, the probability caused by random noise is extremely low, which is strong evidence of real stress. And the damage of multi-system synergy to plant health is much greater than the simple addition of each system damage.

[0117] To realize the deep synergy of multi-source physiological data, this embodiment introduces a cross-modal attention mechanism. This mechanism regards photosynthesis, spectrum and fluorescence as different information modalities, and dynamically learns and assigns the importance weight of each modality at different stress stages through its internal weight calculation network. By calculating the interaction between each modality feature vector, the contribution of the most relevant modality to the current stress state is adaptively amplified, thereby realizing focused information fusion. The calculation process essentially maps the cross-modal attention weighted deep features to a scalar value that can be directly used for disease grading and early warning. This process realizes end-to-end modeling from raw multi-modal data to final decision indicators, ensuring the scientificity and accuracy of the early warning criterion.

[0118] Embodiment seven, this embodiment is a further limitation of the strawberry gray mold early warning system based on multi-physiological index synergy of embodiment six, and the trigger rules of triggering different early warning states are:

[0119] When the alarm levels of the feature indexes corresponding to the fluorescence index, the spectral index and the photosynthesis index are all 0, the early warning state is state 0;

[0120] When the alarm level of the characteristic index corresponding to the fluorescence index is level 1 for two consecutive days or more, or the alarm levels of the characteristic indexes corresponding to the fluorescence index and the spectral index reach level 1 on the same day, or the immune response cumulative value is greater than or equal to 15, the early warning state is state 1.

[0121] When the alarm level of the characteristic index corresponding to the fluorescence index is level 2, and the alarm level of the characteristic index corresponding to the spectral index is greater than or equal to 1, or the immune response cumulative value is greater than or equal to 30, the early warning state is state 2.

[0122] When the alarm level of the characteristic index corresponding to the photosynthetic index is greater than or equal to 1, or the alarm level of any one of the characteristic indexes corresponding to the fluorescence index, the spectral index and the photosynthetic index is level 3, or the immune response cumulative value is greater than or equal to 50, the early warning state is state 3.

[0123] The present embodiment is a hierarchical early warning trigger logic, which triggers early warnings of different levels according to the combination state of the immune response cumulative value and the alarm level of each defense line. The early warning trigger rule is:

[0124] (1) State 0 (healthy): all defense line alarm levels are 0;

[0125] (2) State 1 (first level early warning / potential period attention):

[0126] The first defense line is level 1 for more than or equal to 2 days, or the first and second defense lines reach level 1 at the same time, or the immune response cumulative value is greater than or equal to 15;

[0127] (3) State 2 (second level early warning / early warning before onset):

[0128] The first defense line reaches level 2, and the second defense line is greater than or equal to level 1, or the immune response cumulative value is greater than or equal to 30;

[0129] (4) State 3 (third level early warning / severe stress high risk):

[0130] The third defense line reaches level 1 or above, or any defense line reaches level 3, or the immune response cumulative value is greater than or equal to 50.

[0131] The four-stage state triggering rule designed in this embodiment is based on the interaction between the aforementioned parameter selection, alarm level and plant physiological stress process. State 1 aims to capture the earliest synergistic signal of disease infection. The triggering condition "the first line of defense is in level 1 for more than 2 days" takes advantage of the sensitivity of the fluorescence index, and the requirement of more than two days is to effectively filter out temporary environmental interference through the decay factor λ and the immune response cumulative value. The condition "the first and second lines of defense are in level 1 on the same day" directly triggers the synergistic enhancement factor k, simulating the synergistic amplification effect of early abnormalities of cell function and tissue characteristics, which is the key basis for latent period diagnosis. The setting of immune response cumulative value ≥ 15 at this time is the quantitative reflection of the early weak but persistent stress signal after attenuation and weighted fusion.

[0132] Further, the combined condition "the first line of defense is in level 2 and the second line of defense is in level 1" in state 2 requires the early signal to be upgraded while emphasizing that it must be confirmed by mid-term indicators, greatly improving the specificity of the early warning. The corresponding immune response cumulative value ≥ 30 at this stage reflects the transition from quantitative change to qualitative change of the stress signal under the joint action of weight distribution (w1 > w2) and continuous accumulation.

[0133] Further, the condition "the third line of defense is in level 1" in state 3 means that the photosynthetic function of the whole plant has been damaged, which is a decisive evidence of the degree of stress, and "any line of defense is in level 3" captures extreme abnormalities in any dimension, and a higher immune response cumulative value ≥ 50 is the final quantitative decision for persistent and strong multidimensional stress.

[0134] In summary, this embodiment completes the whole process diagnosis from early perception, mid-term confirmation to end-stage determination, ensuring the timeliness, accuracy and reliability of the early warning.

[0135] Embodiment eight, a strawberry gray mold disease early warning method based on the synergy of multiple physiological indicators, the method is used to realize the system of any one of embodiments one to seven, comprising the following steps:

[0136] Step 1, pre-adapt the strawberry plant in a standard environment, based on software timestamp alignment and hardware synchronization triggering mechanism, obtain multiple physiological indicators of the strawberry plant;

[0137] Step 2, process the multiple physiological indicators of the strawberry plant respectively, and divide the processed multiple physiological indicators into alarm levels;

[0138] Step 3, obtain the immune response cumulative value;

[0139] Step 4, according to the immune response cumulative value and the divided alarm level, trigger different early warning states, and generate an early warning report.

[0140] In the embodiment, the early warning report comprises:

[0141] 1. Current early warning level, clearly marked state 0-3;

[0142] 2. Each defense line state, display the specific value and level of F_index, S_index, P_index;

[0143] 3. Accumulation trend, show the change curve of immune response accumulation value;

[0144] 4. Management suggestions, provide targeted cultivation management and disease control suggestions;

[0145] 5. Review time, according to the early warning level to specify the next detection time (1-7 days).

[0146] The above has carried out the detailed introduction to the strawberry gray mold disease early warning system and method based on the multi-physiological index cooperation provided by the application, the principle and implementation mode of the application are described in this paper by applying specific examples, the above example is only used to help understand the method and core idea of the application; at the same time, for the general technical personnel in the art, according to the idea of the application, the specific implementation mode and application range will be changed, according to the above, the content of the specification should not be understood as the limitation of the application.

Claims

1. A strawberry gray mold early warning system based on the synergy of multiple physiological indicators, characterized in that, It includes a data acquisition layer, a feature processing layer, an alarm judgment layer, a decision fusion layer, and an early warning output layer; The data acquisition layer is used to collect fluorescence, spectral, and photosynthetic indicators of strawberry plants. The feature processing layer is used to calculate the feature indices corresponding to the fluorescence index, spectral index and photosynthetic index collected by the data acquisition layer, respectively. The alarm judgment layer is used to determine the alarm level corresponding to the feature index output by the feature processing layer. The decision fusion layer is used to make fusion decisions based on the alarm level output by the alarm judgment layer to obtain the early warning status; The fusion decision based on the alarm level output by the alarm judgment layer is specifically as follows: The cumulative value of immune response was obtained based on the daily collection of fluorescence, spectral, and photosynthetic indicators. Different warning states are triggered based on the cumulative value of the immune response and the alarm level corresponding to the characteristic index; The cumulative value of the immune response is specifically: in, As the attenuation factor, This represents the cumulative value of the immune response from the previous day. , and These are the alarm levels corresponding to the feature indices. , and These are weighted averages of the alarm levels corresponding to the feature indices. This is a synergistic enhancement item.

2. The strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to claim 1, characterized in that, When collecting fluorescence, spectral, and photosynthetic indicators of strawberry plants, a software timestamp alignment and hardware synchronization triggering mechanism are used to ensure that the data acquisition devices are triggered simultaneously, thereby achieving data alignment.

3. The strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to claim 1, characterized in that, The fluorescence indicators include the maximum photochemical efficiency Fv / Fm, the actual photochemical efficiency ΦPSⅡ, the photochemical quenching coefficient qP, and the non-photochemical quenching coefficient NPQ; The spectral parameters include the photochemical reflectance index PRI, the stress index SI, and the improved red edge ratio mSR; The photosynthetic indices include the leaf's net photosynthetic rate Pn, stomatal conductance Gs, intercellular CO2 concentration Ci, and transpiration rate Tr.

4. The strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to claim 1, characterized in that, The characteristic indices corresponding to the fluorescence index, spectral index, and photosynthetic index are the fluorescence decay index F_index, the spectral stress index S_index, and the photosynthetic inhibition index P_index, respectively.

5. A strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to claim 4, characterized in that, The alarm level corresponding to the fluorescence decay index is: when When it is level 0, when When it is level 1, when When it is level 2, when It is level 3 at the time; The alarm level corresponding to the spectral stress index is: when When it is level 0, when When it is level 1, when When it is level 2, when It is level 3 at the time; ； in, This represents the standard score of the current spectral stress index relative to a healthy population baseline. The current spectral stress index, S_index represents the mean value of the healthy group. S_index represents the standard deviation of the healthy population. The alarm level corresponding to the photosynthetic inhibition index is: when When it is level 0, when When it is level 1, when When it is level 2, when The time is level 3.

6. The strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to claim 1, characterized in that, When the alarm level corresponding to two or more feature indices is greater than or equal to 1, the collaborative enhancement item is triggered.

7. The strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to claim 1, characterized in that, The triggering rules for different warning states are as follows: When the alarm level of the characteristic index corresponding to the fluorescence index, spectral index and photosynthetic index is 0, the warning status is status 0. When the alarm level of the characteristic index corresponding to the fluorescence index is Level 1 for more than two consecutive days, or the alarm levels of the characteristic indices corresponding to the fluorescence index and the spectral index reach Level 1 on the same day, or the cumulative value of the immune response is greater than or equal to 15, the warning status is Status 1. When the alarm level of the characteristic index corresponding to the fluorescence index is level 2, and the alarm level of the characteristic index corresponding to the spectral index is greater than or equal to 1, or the cumulative value of the immune response is greater than or equal to 30, the warning state is state 2. When the alarm level of the characteristic index corresponding to the photosynthetic index is greater than or equal to 1, or the alarm level of any characteristic index among the fluorescence index, spectral index and photosynthetic index is level 3, or the cumulative value of immune response is greater than or equal to 50, the warning status is status 3.

8. A strawberry gray mold early warning system based on the synergy of multiple physiological indicators according to any one of claims 1-7, wherein the system is implemented based on a strawberry gray mold early warning method, characterized in that, The method includes the following steps: Step 1: Pre-adapt strawberry plants to a standard environment and obtain multiple physiological indicators of strawberry plants based on software timestamp alignment and hardware synchronization triggering mechanism. Step 2: Process the multiple physiological indicators of the strawberry plants separately, and classify the alarm levels of the processed multiple physiological indicators. Step 3: Obtain the cumulative value of the immune response; Step 4: Based on the cumulative value of the immune response and the classified alarm levels, trigger different warning states and generate warning reports.

Images