Real-time intelligent crop nutrient evaluation method based on multi-mode remote sensing of unmanned aerial vehicle

Abstract

The invention relates to the technical field of data processing, in particular to a crop nutrient real-time intelligent evaluation method based on unmanned aerial vehicle multi-mode remote sensing. Comprising the following steps: performing multi-dimensional data acquisition on crops to obtain multi-dimensional crop data, and preprocessing the multi-dimensional crop data to obtain preprocessed multi-dimensional crop data; performing data enhancement fusion processing on the preprocessed multi-dimensional crop data by using a multi-scale adaptive nonlinear data fusion algorithm to obtain enhanced and fused crop data; and based on the enhanced and fused crop data, introducing a dynamic adaptive coupling evaluation algorithm to preliminarily evaluate crop nutrients to obtain a preliminary evaluation result, performing evaluation in combination with environmental data to obtain a final crop nutrient evaluation result, and comparing the final crop nutrient evaluation result with a judgment threshold. The technical problem of low accuracy and efficiency of crop nutrient real-time evaluation in a changeable agricultural environment is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method for real-time intelligent assessment of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles. Background Art

[0002] At present, the growth and nutrient management of crops in agricultural production have always been the key to improving crop yield and quality. However, most traditional crop nutrient assessment methods rely on manual detection or single chemical detection means. These methods often require a long time cycle and high costs, and it is difficult to achieve real-time monitoring and large-scale application. In large-scale agricultural production, soil nutrient detection, crop leaf analysis, and environmental factor monitoring are often carried out separately, lacking effective comprehensive data fusion, resulting in insufficient accuracy and timeliness of crop nutrient status assessment.

[0003] With the development of information technology, especially the progress of Internet of Things technology, sensor technology, and big data analysis technology, more and more agricultural fields have begun to try to use these new technologies to improve the accuracy of crop nutrient management. For example, sensors based on various environmental factors such as soil moisture, temperature, pH value, and light intensity have gradually been applied to the agricultural environment for dynamically monitoring soil conditions and crop growth environments. At the same time, through remote sensing technology, researchers can obtain the reflectance information on the surface of crops to infer the chlorophyll content and health status of crops.

[0004] However, the existing methods for real-time intelligent assessment of crop nutrients have the following technical problems: in a changing agricultural environment, the accuracy and efficiency of real-time assessment of crop nutrients are relatively low. Summary of the Invention

[0005] The present invention provides a method for real-time intelligent assessment of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles to solve the technical problem of relatively low accuracy and efficiency of real-time assessment of crop nutrients in a changing agricultural environment.

[0006] A method for real-time intelligent assessment of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles according to the present invention specifically includes the following technical solutions:

[0007] A method for real-time intelligent assessment of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles includes the following steps:

[0008] S1. Collect multi-dimensional data of crops to obtain multi-dimensional crop data, preprocess the multi-dimensional crop data to obtain preprocessed multi-dimensional crop data; use a multi-scale adaptive non-linear data fusion algorithm to perform data enhancement and fusion processing on the preprocessed multi-dimensional crop data to obtain enhanced and fused crop data;

[0009] S2. Based on the enhanced and fused crop data, introduce a dynamic adaptive coupling evaluation algorithm to preliminarily evaluate the crop nutrients, obtain the preliminary evaluation result, and then combine with the environmental data for evaluation to obtain the final crop nutrient evaluation result, and compare the final crop nutrient evaluation result with the determination threshold.

[0010] Preferably, the S1 specifically includes:

[0011] In the implementation process of the multi-scale adaptive non-linear data fusion algorithm, extract features from the preprocessed multi-dimensional crop data to obtain initial feature data; introduce an adaptive non-linear function to perform non-linear dynamic transformation on the initial feature data to obtain the data after non-linear dynamic transformation.

[0012] Preferably, the S1 specifically includes:

[0013] In the implementation process of the multi-scale adaptive non-linear data fusion algorithm, after performing non-linear dynamic transformation, perform local weighted transformation on the data after non-linear dynamic transformation to obtain local feature data, and perform weighted fusion processing on the local feature data to obtain weighted local feature data. The specific implementation formula is:

[0014]

[0015] Among them, L i (x, y) is the weighted local feature data of the i-th sensor at the position (x, y); (u, v) is the coordinate, representing the position index, corresponding to the spatial coordinate, belonging to the local area inside; α i (u, v) is the weighting coefficient of the data after non-linear dynamic transformation in the local area; θ i (x, y) is the adjustment offset term; is the local feature data of the data after non-linear dynamic transformation in the local area.

[0016] Preferably, the S1 specifically includes:

[0017] In the implementation process of the multi-scale adaptive non-linear data fusion algorithm, after local weighted transformation, introduce multi-scale processing, and decompose the weighted local feature data after local weighted transformation into different scale levels through wavelet transform to obtain the feature data at different scales after wavelet transform.

[0018] Preferably, the S1 specifically includes:

[0019] In the implementation process of the multi-scale adaptive non-linear data fusion algorithm, by combining multi-scale information, the feature data at different scales after wavelet transform are weighted and averaged to obtain the fused multi-scale crop data, and the fused multi-scale crop data corresponding to each sensor are combined to obtain the enhanced fused crop data.

[0020] Preferably, the S2 specifically includes:

[0021] In the implementation process of the dynamic adaptive coupling evaluation algorithm, the enhanced fused crop data are fused with the environmental data to obtain a preliminary evaluation result of crop nutrients.

[0022] Preferably, the S2 specifically includes:

[0023] In the implementation process of the dynamic adaptive coupling evaluation algorithm, based on the preliminary evaluation result, an environmental dynamic correction mechanism is introduced, and the environmental correction factor and the influence function based on the environmental data are combined to adjust the preliminary evaluation result in real time. The specific implementation formula is:

[0024]

[0025] where N(t) is the evaluation result of crop nutrients after environmental dynamic correction, and N initial (t) is the preliminary evaluation result, is the environmental correction factor at time point t; f 2 (E(t)) is the influence function based on the environmental data.

[0026] Preferably, the S2 specifically includes:

[0027] In the implementation process of the dynamic adaptive coupling evaluation algorithm, based on the evaluation result of crop nutrients after environmental dynamic correction, a feedback mechanism is introduced, and combined with the feedback coefficient, the preliminary evaluation result of the previous moment is introduced into the evaluation process of the current moment for dynamic sequential correction to obtain the evaluation result of crop nutrients at the current moment.

[0028] Preferably, the S2 specifically includes:

[0029] In the implementation process of the dynamic adaptive coupling evaluation algorithm, a multi-order cumulative optimization strategy is introduced to perform weighted sum and accumulation on the evaluation results of crop nutrients at different time steps to obtain the final evaluation result of crop nutrients; and a judgment threshold is set and compared with the final evaluation result of crop nutrients to judge the health status of the crop.

[0030] The beneficial effects of the technical solution of the present invention are:

[0031] 1. The multi-scale adaptive non-linear data fusion algorithm is adopted to enhance and fuse the pre-processed multi-dimensional crop data, extract features from the pre-processed multi-dimensional crop data, capture various dynamic change features of the crops, and further enhance the local features of the pre-processed multi-dimensional crop data through non-linear dynamic transformation; through weighted fusion processing, the local features of each sensor are strengthened, so that the nutrient, health and growth status information of the crops can be more comprehensively reflected in space and time; through wavelet transform, multi-scale analysis is carried out on the weighted local feature data to reveal the multi-level growth dynamics of the crops from details to the whole, ensuring the accurate retention of feature information.

[0032] 2. The dynamic adaptive coupling evaluation algorithm is adopted for crop nutrient evaluation. In the preliminary evaluation stage, the enhanced and fused crop data is combined with environmental data to conduct a preliminary evaluation of crop nutrients, and a preliminary evaluation result is obtained; the preliminary evaluation result is dynamically corrected according to environmental data and time changes, thus realizing the real-time and accurate evaluation of crop nutrient status; by introducing an environmental dynamic correction mechanism, the preliminary evaluation result can be adaptively adjusted with the time-variability of the environment, significantly improving the accuracy and real-time performance of the evaluation. Description of the Drawings

[0033] Figure 1 It is a flowchart of a method for real-time intelligent evaluation of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles according to the present invention. Detailed Embodiments

[0034] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0036] The specific scheme of a method for real-time intelligent evaluation of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles provided by the present invention will be specifically described below with reference to the drawings.

[0037] Refer to the attached Figure 1 , which shows a flowchart of a method for real-time intelligent evaluation of crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles provided by an embodiment of the present invention. The method includes the following steps:

[0038] S1. Collect multi-dimensional data of crops to obtain multi-dimensional crop data, preprocess the multi-dimensional crop data to obtain preprocessed multi-dimensional crop data; use a multi-scale adaptive non-linear data fusion algorithm to perform data enhancement fusion processing on the preprocessed multi-dimensional crop data to obtain enhanced and fused crop data;

[0039] Through a drone multi-modal remote sensing sensor, collect multi-dimensional data of crops to obtain multi-dimensional crop data; the drone multi-modal remote sensing sensor includes sensors such as an optical camera, a near-infrared sensor, a thermal infrared sensor, a lidar (LiDAR), and a high-frequency radar. High-resolution images can reflect the visual state of crops, while near-infrared and thermal infrared images help capture the physiological characteristics of plants. Lidar can provide the height and structure information of crops, and high-frequency radar can reflect the physical density and growth state of crops; the multi-dimensional crop data includes data information in multiple aspects such as the leaf color of crops, plant structure, soil humidity, chlorophyll content, and temperature distribution; perform preprocessing operations such as data denoising, calibration, registration, dimension unification processing, and standardization on the multi-dimensional crop data to obtain preprocessed multi-dimensional crop data. The preprocessing process all adopts the existing technologies well-known to those skilled in the art and will not be elaborated here.

[0040] Furthermore, use a multi-scale adaptive non-linear data fusion algorithm to perform data enhancement fusion processing on the preprocessed multi-dimensional crop data to obtain enhanced and fused crop data; the specific implementation process is as follows:

[0041] Use existing feature engineering techniques to extract features from the preprocessed multi-dimensional crop data to obtain initial feature data D = {D 1 ,D 2 ,...,D i ,...,D n}; where n represents the number of sensors; any element is represented as which represents the initial feature matrix extracted from a certain sensor or a specific feature space and has certain spatial and temporal information; m i is the spatial dimension of the i-th sensor, such as the number of pixels; n i is the time dimension extracted by the i-th sensor, such as spectral channels, time steps, etc.

[0042] Furthermore, capture the complex change trends in the preprocessed multi-dimensional crop data through non-linear dynamic transformation. To better adapt to the initial feature data D, derive the spatial gradient of each sensor data D i and introduce an adaptive non-linear function to achieve non-linear dynamic transformation and strengthen the local feature changes of the preprocessed multi-dimensional crop data. This process can be expressed by the following formula:

[0043]

[0044] Among them, is the data after non - linear dynamic transformation, representing the characteristics after non - linear dynamic transformation of the i - th sensor at the position (x, y); D i (x, y) is the initial feature data of the i - th sensor at the position (x, y), and the sensors include optical cameras, near - infrared sensors, lidars, etc.; and are the initial feature data D i (x, y) of the i - th sensor at the position (x, y) with respect to the partial derivatives of the spatial coordinates x and y, representing the rate of change of the initial feature data in the x and y directions, that is, the gradient of the initial feature data. The gradient reflects the degree and direction of the change of the initial feature data in space, can reveal the subtle changes in the initial feature data, and helps to understand the local dynamic changes of the crop at different positions or different times; β and γ are hyperparameters used to control the non - linear response of the gradient of the initial feature data in different directions, emphasizing the change sensitivity of the gradient in the x and y directions; Ω is the integration region, representing the spatial region involved in calculating the partial derivatives; dΩ is the infinitesimal element of the integration region, representing the area at a certain point in space.

[0045] After performing the non - linear dynamic transformation, local weighted transformation is performed on the data after non - linear dynamic transformation to obtain local feature data, and the expression of local feature data is strengthened through weighted fusion. First, each data after non - linear dynamic transformation is used as the center, and the size r of the local region set according to the specific scenario is used as the local region Then, the local feature data is weighted and fused through the following formula:

[0046]

[0047] Among them, L i (x, y) is the weighted local feature data of the i - th sensor at the position (x, y), representing the feature data of the local feature data at this position after weighting; (u, v) are coordinates, representing the position index, corresponding to the spatial coordinates, belonging to the local region inside, representing the spatial position of the specific sensor i in the dataset; α i (u, v) is the weighting coefficient of the data after non - linear dynamic transformation in the local region, representing the contribution degree of the data after non - linear dynamic transformation at the position (u, v) to the weighted local feature data, obtained by fitting experimental data and not limited here; is the local feature data of the data after non - linear dynamic transformation within a local area; θ i (x, y) is the adjustment offset term, which is used to refine the benchmark of the local feature data.

[0048] Through local weighted transformation, the features of each local area will be appropriately strengthened or weakened according to their performance in the entire crop area, which can better retain the local differences in crop growth status and nutrient levels, and further improve the accuracy of data fusion.

[0049] After local weighted transformation, in order to enhance the effect of data fusion, multi - scale processing is introduced. The weighted local feature data after local weighted transformation is further decomposed into multiple scale levels through wavelet transform, so as to obtain the features of the crop at different spatial scales. The change of each scale can be expressed by the following formula:

[0050]

[0051] Among them, is the feature data of the i - th sensor after wavelet transform in the area S at the k - th scale k ; represents the wavelet transform operation; S k is the area at the k - th scale.

[0052] Through wavelet transform, information at different scales from details to the whole can be obtained. The feature data after wavelet transform at these different scales can reveal various dynamic changes in crop growth, including subtle nutrient fluctuations and macroscopic crop states.

[0053] Then, combining multi - scale information, the feature data k of the i - th sensor after wavelet transform in the area S at the k - th scale is weighted and averaged to obtain the fused multi - scale crop data

[0054]

[0055] Among them, ω k is the weighting coefficient of each scale level, which is determined by the expert experience method and is not limited here; m is the total number of scale levels.

[0056] The multi - scale information provided by wavelet transform can more comprehensively understand the growth state of the crop, avoiding the omission or distortion of features at a single scale.

[0057] After completing multi - scale processing, the fused multi - scale crop data corresponding to all sensors are combined to obtain the enhanced fused crop data

[0058] S2. Based on the enhanced and fused crop data, introduce a dynamic adaptive coupling evaluation algorithm to preliminarily evaluate the crop nutrients, obtain the preliminary evaluation result, and then combine with the environmental data for evaluation to obtain the final crop nutrient evaluation result, and compare the final crop nutrient evaluation result with the determination threshold.

[0059] Based on the enhanced and fused crop data, introduce a dynamic adaptive coupling evaluation algorithm to preliminarily evaluate the crop nutrients, obtain the preliminary evaluation result, and then combine with the environmental data for optimized evaluation to obtain the optimized evaluation result; the dynamic adaptive coupling evaluation algorithm adopts the correction of environmental data, the time-series data feedback mechanism and the multi-stage cumulative optimization strategy, which can more accurately evaluate the nutrient status of crops under dynamic environmental conditions, overcome the static characteristics of traditional crop nutrient evaluation algorithms, and enable it to adaptively adjust under different times and different environmental conditions, so as to realize real-time and accurate evaluation of crop nutrients. The specific implementation process is as follows:

[0060] First, in the preliminary stage of crop nutrient evaluation, a method of fusing the enhanced and fused crop data with the environmental data is adopted to obtain the preliminary evaluation result of crop nutrients. The environmental data includes various factors such as temperature, humidity, and light intensity. The specific expression is:

[0061]

[0062] where N initial (t) represents the preliminary evaluation result of crop nutrients at time point t, is the enhanced and fused crop data at time point t; E(t) represents the environmental data at time point t; is the evaluation function, which is a function obtained through crop historical training data or experimental research and is used to map the enhanced and fused crop data and environmental data to the preliminary evaluation result. The evaluation function can be a regression model, a statistical model or a model based on physical principles, which is determined according to the specific scenario, such as where, is the th enhanced and fused crop data at time point t; is the total number of enhanced and fused crop data; is the th type of environmental data at time point t, and all environmental data are of the same dimension; is the total number of environmental data categories; is the constant term, representing the base value, which is obtained through fitting experiments and is not limited here; is the regression coefficient of the enhanced and fused crop data, representing the contribution of the enhanced and fused crop data to crop nutrients, obtained through fitting experiments, not limited herein; is the regression coefficient of environmental data, representing the degree of influence of environmental data on crop nutrients, obtained through fitting experiments, not limited herein; ∈(t) is an error term.

[0063] Furthermore, in order to consider the dynamic influence of environmental conditions on crop nutrients, an environmental dynamic correction mechanism is introduced to correct the time-variability of different environmental changes on the crop nutrient status. Therefore, an environmental correction factor in terms of time is introduced The influence of environmental data on crop nutrients is not linear but varies with time and environmental conditions. Therefore, the environmental correction factor and the influence function f 2 (E(t)) are used to adjust the preliminary evaluation results in real time, thereby enhancing the timeliness and accuracy of nutrient evaluation. The dynamic correction formula is as follows:

[0064]

[0065] where N(t) is the crop nutrient evaluation result after environmental dynamic correction, N initial (t) is the preliminary evaluation result, is the environmental correction factor at time point t, representing the correction intensity of environmental changes, obtained according to the expert experience method, such as where, is the weighting coefficient of environmental data, representing the influence intensity of different environmental factors on crop nutrients, obtained by fitting experimental data, not limited herein; T(t) is the temperature at time point t; H(t) is the humidity at time point t; L(t) is the light intensity at time point t; f 2 (E(t)) is the influence function based on environmental data, used to capture the influence of factors such as air temperature and humidity on crop nutrients, derived through experimental observations or regression analysis of an agricultural meteorological database, such as where, is the weighting coefficient of environmental data, representing the influence intensity of different environmental factors on crop nutrients, obtained by fitting experimental data, not limited herein; T opt is the optimal growth temperature of the crop, determined according to the expert experience method, not limited herein; σ T is the temperature sensitivity parameter, used to control the sensitivity of the crop to temperature changes, obtained through experiments, not limited herein; is the influence coefficient of humidity on the crop, reflecting the influence intensity of humidity on crop nutrients, determined according to the expert experience method, not limited herein; L maxis the maximum light intensity, which is used as a normalization factor to normalize the light intensity. It is determined by the expert experience method and is not limited here.

[0066] Furthermore, to improve the accuracy of the crop nutrient assessment results after environmental dynamic correction, a feedback mechanism is introduced. The change of crop nutrients is not only affected by the current environment but also by the nutrient changes at historical moments, that is, the nutrient status of crops changes continuously. Therefore, the nutrient assessment at the previous moment has a certain impact on the nutrient assessment at the current moment. Here, the preliminary assessment result N initial (t - 1) at the previous moment is introduced into the assessment process at the current moment and is weighted by the feedback coefficient to achieve dynamic temporal correction. The formula is expressed as follows:

[0067]

[0068] where, is the crop nutrient assessment result at the current time t; N initial (t - 1) is the preliminary assessment result at the previous moment; is the feedback coefficient at the time step t, which is used to adjust the impact of the preliminary assessment result at the previous moment on the crop nutrient assessment at the current moment. The feedback coefficient is derived from historical nutrient data and time series models and is not limited here.

[0069] The above process ensures the temporal consistency of the crop nutrient assessment results and prevents sudden environmental changes in the short term from having an excessive impact on the assessment results.

[0070] To further enhance the robustness of the assessment, a multi - order cumulative optimization strategy is introduced. The assessment of crop nutrients is not only a reflection of a single moment but also affected by the comprehensive influence of environmental factors over a long time period. In particular, the impact of long - term environmental changes (such as seasonal changes, climate trends, etc.) on crop nutrients is more significant. Therefore, by weighting and accumulating the crop nutrient assessment results at multiple time steps, the impact of long - term environmental changes on crop nutrients can be considered more comprehensively. The specific implementation formula is as follows:

[0071]

[0072] where, N final is the final crop nutrient assessment result; is the preliminary assessment result of crop nutrients in the time period (for example, a specific time point or within a certain time period); is the total time period, which represents the total time length of the crop growth cycle. For example, the entire season or the full growth cycle of the crop; is the cumulative correction factor, which is used to control the weighting degree of the final crop nutrient assessment results in different time periods, reflects the continuous impact of long-term climate change on crop nutrients. In some special cases, such as extreme climate changes, the weight of the cumulative correction factor will be increased, so as to consider more long-term impacts in the correction process. It is determined according to the expert experience method and is not limited here; is the environmental correction factor in the th time period, indicating the correction intensity of environmental changes. It is obtained according to the expert experience method and is not limited here; is the impact function based on environmental data, which is used to capture the impacts of factors such as temperature and humidity on crop nutrients, and is derived through experimental observations or regression analysis of agricultural meteorological databases; represents the environmental data in the th time period.

[0073] Preset the judgment thresholds ρ 1 , ρ 2 , ρ 3 and ρ 4 according to the expert experience method, and compare the set judgment thresholds with the final crop nutrient assessment results. If the final crop nutrient assessment result is greater than or equal to ρ1, the crop is in an excellent health state. Within this range, the nutrient concentration, leaf area index and photosynthetic efficiency of the crop are all at a relatively high level, growing vigorously with almost no growth obstacles, which means that the environmental conditions and nutrient supply of the crop are very ideal and the health condition is very good. At this time, no large-scale adjustment is required, and agricultural managers can maintain the current crop management strategy to ensure its continuous growth.

[0074] For crops with the final crop nutrient assessment result lower than ρ 1 and greater than or equal to ρ 2 , the crop is in a good health condition but may have slight nutrient deficiencies or environmental stresses. Although the photosynthetic efficiency and leaf area index of the crop are relatively high, they may be slightly lower than the ideal values. Therefore, the crop still maintains a healthy state, but needs to be continuously observed to ensure that no further nutrient deficiencies occur.

[0075] For crops with the final crop nutrient assessment result lower than ρ 2 and greater than or equal to ρ 3 , it indicates that the health condition of the crop is at a medium level. At this time, the growth of the crop may be affected by some environmental stresses or nutrient deficiencies, the leaf area index and photosynthesis efficiency decrease, and the growth rate slows down. At this time, agricultural managers should pay special attention to the nutrient supply of the crop, and may need to increase the application amount of some key nutrients, or promote the recovery growth of the crop by improving the environmental conditions.

[0076] For the final crop nutrient assessment result lower than ρ 3 Greater than or equal to ρ 4 When this occurs, the crop's health condition is poor, and obvious nutrient deficiencies or environmental adaptability problems have emerged. At this time, the growth of the crop has almost stagnated, the photosynthesis efficiency has decreased significantly, the leaf area has shrunk significantly, the overall growth has been hindered, and it may have entered the growth crisis stage. At this time, urgent measures must be taken, such as quickly supplementing the lacking nutrients, adjusting the irrigation volume, or strengthening pest and disease control.

[0077] If the final crop assessment result is lower than ρ4, it indicates that the crop's health condition is very poor, the growth has almost stagnated or has entered the decline state. This situation is usually caused by severe nutrient deficiencies or environmental deterioration. The crop faces great growth risks and may be approaching the death stage. In this case, agricultural managers need to immediately carry out emergency fertilization, irrigation, and pest and disease control measures to save the growth of the crop as much as possible.

[0078] In summary, a real-time intelligent assessment method for crop nutrients based on multi-modal remote sensing of unmanned aerial vehicles is completed.

[0079] The sequence of the invention embodiments is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-tasking and parallel processing are also possible or may be advantageous.

[0080] Each embodiment in this specification is described in a progressive manner. For the same or similar parts between the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

[0081] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention and should all be included in the protection scope of the present invention.

Claims

1. A real-time intelligent crop nutrient assessment method based on multimodal remote sensing by unmanned aerial vehicles, characterized in that: The following steps are involved: S1. Collect multi-dimensional crop data to obtain multi-dimensional crop data, and pre-process the multi-dimensional crop data to obtain pre-processed multi-dimensional crop data; The multi-scale adaptive nonlinear data fusion algorithm is used to perform data enhancement and fusion processing on the pre-processed multi-dimensional crop data to obtain enhanced and fused crop data; S2. Based on the enhanced fused crop data, a dynamic adaptive coupling evaluation algorithm is introduced to conduct a preliminary evaluation of crop nutrients to obtain preliminary evaluation results, which are then evaluated in combination with environmental data to obtain the final crop nutrient evaluation results, and the final crop nutrient evaluation results are compared with the judgment threshold.

2. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 1, characterized in that: The S1 specifically includes: In the process of implementing the multi-scale adaptive nonlinear data fusion algorithm, feature extraction is performed on the preprocessed multi-dimensional crop data to obtain initial feature data; an adaptive nonlinear function is introduced to perform nonlinear dynamic transformation on the initial feature data to obtain data after nonlinear dynamic transformation.

3. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 2, characterized in that: The S1 specifically includes: In the implementation process of the multi-scale adaptive nonlinear data fusion algorithm, after the nonlinear dynamic transformation, the data after the nonlinear dynamic transformation is subjected to local weighted transformation to obtain local feature data, and the local feature data is subjected to weighted fusion processing to obtain weighted local feature data. The specific implementation formula is: Among them, L i (x, y) is the weighted local feature data of the ith sensor at position (x, y); (u, v) is the coordinate, indicating the position index, corresponding to the spatial coordinate, belonging to the local area Inner; α i (u, v) is the weighting coefficient of the data after nonlinear dynamic transformation in the local area; θ i (x, y) is the adjustment offset term; It is the local feature data in the local area after nonlinear dynamic transformation.

4. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 3, characterized in that: The S1 specifically includes: In the implementation process of the multi-scale adaptive nonlinear data fusion algorithm, multi-scale processing is introduced after the local weighted transformation. The weighted local feature data after the local weighted transformation is decomposed into different scale levels through wavelet transform to obtain feature data at different scales after wavelet transform.

5. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 4, characterized in that: The S1 specifically includes: In the implementation process of the multi-scale adaptive nonlinear data fusion algorithm, the multi-scale information is combined, and the feature data at different scales after wavelet transform are weighted averaged to obtain the fused multi-scale crop data, and the fused multi-scale crop data corresponding to each sensor are merged to obtain the enhanced fused crop data.

6. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 1, characterized in that: The S2 specifically includes: In the process of implementing the dynamic adaptive coupling assessment algorithm, the enhanced fused crop data are fused with the environmental data to obtain the preliminary assessment results of crop nutrients.

7. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 6, characterized in that: The S2 specifically includes: In the implementation process of the dynamic adaptive coupling evaluation algorithm, based on the preliminary evaluation results, an environmental dynamic correction mechanism is introduced to combine the environmental correction factor with the influence function based on environmental data to make real-time adjustments to the preliminary evaluation results. The specific implementation formula is: Among them, N(t) is the crop nutrient assessment result after dynamic correction of the environment, N initial (t) is the preliminary assessment result, is the environmental correction factor at time point t; f2(E(t)) is the influence function based on environmental data.

8. The method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 7, characterized in that: The S2 specifically includes: In the implementation process of the dynamic adaptive coupling evaluation algorithm, a feedback mechanism is introduced based on the crop nutrient evaluation results after dynamic environmental correction. Combined with the feedback coefficient, the preliminary evaluation results of the previous moment are introduced into the evaluation process of the current moment, and dynamic timing correction is performed to obtain the crop nutrient evaluation results of the current moment.

9. A method for real-time intelligent assessment of crop nutrients based on multimodal remote sensing using unmanned aerial vehicles according to claim 8, characterized in that: The S2 specifically includes: In the implementation process of the dynamic adaptive coupling assessment algorithm, a multi-order cumulative optimization strategy is introduced to weight and accumulate the crop nutrient assessment results of different time steps to obtain the final crop nutrient assessment result; and a judgment threshold is set to compare with the final crop nutrient assessment result to judge the health status of the crop.

Images