A nutrient synergistic regulation system and control method in water-saving cultivation of wheat

Abstract

This invention relates to a nutrient synergistic regulation system and control method for water-saving wheat cultivation, belonging to the field of wheat cultivation technology. The system includes: mapping soil moisture and nutrient content to a first feature vector; enhancing a wheat growth state image to obtain an enhanced growth state image; extracting a second feature vector from the enhanced growth state image using an image feature extraction network; fusing the first and second feature vectors to obtain a fused feature; inputting the fused feature into a deep learning model for training to obtain a wheat nutrient demand prediction model; and using the wheat nutrient demand prediction model to regulate the nutrients of the target wheat. This invention employs a random forest algorithm, an image feature extraction network, and a deep learning model, which can fully explore and learn the complex relationship between the soil environment and wheat growth state, more accurately reflecting the actual needs of wheat, achieving precise water and fertilizer ratios, and thus improving water and fertilizer utilization efficiency.

Description

Technical Field

[0001] This invention relates to the field of wheat cultivation technology, specifically to a nutrient synergistic regulation system and control method for water-saving wheat cultivation. Background Technology

[0002] In traditional wheat cultivation, water and nutrient management are often carried out independently, neglecting their interaction and synergistic effects. This traditional approach not only fails to maximize the benefits of water and fertilizer resources but may also lead to resource waste due to over-irrigation or improper fertilization, resulting in environmental pollution and uneven crop growth.

[0003] In recent years, with the development of precision agriculture technology, the demand for intelligent control and refined management has gradually increased. In particular, the coordinated regulation of nutrients and water, as an emerging coupled management model, has gradually become an important direction for improving the efficiency of water-saving wheat cultivation. Studies have shown that the coordinated supply of water and nutrients can effectively promote crop root growth, improve water and fertilizer utilization, and thus enhance crop growth rate and stress resistance.

[0004] However, existing technologies for achieving coordinated water and fertilizer regulation often suffer from problems such as untimely response to changes in water and fertilizer demand, poor precision, and limited control methods. Some control systems can only perform simple water and fertilizer allocation under specific conditions and cannot dynamically adjust the water and fertilizer supply according to different growth stages and environmental conditions, resulting in unsatisfactory regulation effects. Summary of the Invention

[0005] To address the aforementioned problems, the purpose of this invention is to provide a nutrient synergistic regulation system and control method for water-saving wheat cultivation.

[0006] A method for synergistic regulation and control of nutrients in water-saving wheat cultivation includes:

[0007] Step 1: Collect images of soil moisture, nutrient content, and wheat growth status;

[0008] Step 2: Use the random forest algorithm to map soil moisture and nutrient content into the first feature vector;

[0009] Step 3: Enhance the wheat growth status image to obtain the enhanced growth status image;

[0010] Step 4: Use an image feature extraction network to extract the second feature vector of the enhanced growth state image;

[0011] Step 5: Fuse the first feature vector and the second feature vector to obtain the fused feature;

[0012] Step 6: Input the fused features into the deep learning model for training to obtain the wheat nutrient demand prediction model;

[0013] Step 7: Use a wheat nutrient requirement prediction model to regulate the nutrients of the target wheat.

[0014] Preferably, step 3: enhancing the wheat growth status image to obtain an enhanced growth status image includes:

[0015] Step 3.1: Perform edge enhancement processing on the wheat growth status image to obtain the preprocessed growth status image;

[0016] Step 3.2: Take a set of neighboring pixels centered on any point on the original growth state image;

[0017] Step 3.3: Use the preprocessed growth state image as the guide image to calculate the enhancement component of the center point;

[0018] Step 3.4: Enhance the wheat growth status image using the enhancement component to obtain the enhanced growth status image.

[0019] Preferably, the formula for calculating the enhancement component is:

[0020]

[0021] in, Indicates the normalized weights. This represents the augmentation component at point p. express The coordinates of point p on the image, express The coordinates of point q on the image, This represents the pixel value at point q in the preprocessed growth state image. This represents the set of neighboring pixels centered at pixel p. This represents the standard deviation of the Gaussian filter.

[0022] Preferably, step 3.4 includes:

[0023] Step 3.4.1: Iteratively compensate the preprocessed growth state image using enhancement components to obtain an iterative image; wherein the iterative compensation process is as follows:

[0024]

[0025] in, Indicates the iterative weights. This represents the enhancement component at point p after the (t+1)th iteration. This represents the enhancement component at point p after the t-th iteration. This represents the enhancement component at point q after the t-th iteration. This represents the set of neighboring pixels centered at pixel p. Indicates the preset weight;

[0026] Step 3.4.2: Convolve the iterative image to obtain the enhanced growth state image.

[0027] Preferably, in step 3.4.2, the formula is used:

[0028]

[0029] Convolutional enhancement is applied to the iterative image to obtain the enhanced growth state image; where, This represents the enhanced growth state image. This represents convolution.

[0030] Preferably, step 4: extracting the second feature vector of the enhanced growth state image using an image feature extraction network, includes:

[0031] Step 4.1: Input the enhanced grown image into the convolutional layer and pooling layer sequentially to obtain the global feature map; the formula for extracting the global feature map is:

[0032]

[0033] in, For global feature maps, To enhance the post-growth state image, , , These represent convolution operations using kernels of different sizes. This indicates concatenation along the channel dimension. This represents the max pooling operation;

[0034] Step 4.2: Use 3D convolution to extract spatial attention features and channel attention features from the enhanced grown state image;

[0035] Step 4.3: Fuse the global feature map, spatial attention features, and channel attention features to obtain the second feature vector.

[0036] Preferably, step 4.2: extracting spatial attention features and channel attention features of the enhanced grown state image using 3D convolution, including:

[0037] Formula used:

[0038]

[0039]

[0040] Spatial attention features and channel attention features are extracted from the enhanced growth state image; among them, This represents the image showing the growth state after enhancement. Representing spatial attention characteristics, Indicates channel attention features, For 3D convolution, For three-dimensional average pooling, For pooling, To modify the activation function of the linear unit, This represents the Sigmoid activation function. , These are the attention weight parameters.

[0041] Preferably, in step 4.3, the formula is used:

[0042]

[0043]

[0044] The second feature vector is obtained by fusing the global feature map, spatial attention features, and channel attention features. .

[0045] Preferably, in step 6, the training steps of the deep learning model include:

[0046] Step 6.1: Introduce a water-fertilizer coupling coefficient into the deep learning model; whereby the water-fertilizer coupling coefficient is:

[0047]

[0048] in, For soil moisture, For coupling strength parameters, The optimal soil moisture threshold;

[0049] Step 6.2: Construct the loss function of the deep learning model using the water-fertilizer coupling coefficient; wherein the loss function is:

[0050]

[0051] in, To reflect actual nutrient requirements, To predict nutrient requirements, For the sample size, For model parameters, and The regularization coefficient is . Let KL divergence be the KL divergence. and The first The empirical and prior distributions of layer features.

[0052] This invention also provides a nutrient synergistic regulation and control system for water-saving wheat cultivation, comprising:

[0053] The data acquisition module is used to collect images of soil moisture, nutrient content, and the growth status of wheat.

[0054] The feature mapping module is used to map soil moisture and nutrient content into a first feature vector using the random forest algorithm;

[0055] The enhancement module is used to enhance the wheat growth status image to obtain an enhanced growth status image;

[0056] The feature extraction module is used to extract the second feature vector of the enhanced growth state image using an image feature extraction network;

[0057] The feature fusion module is used to fuse the first feature vector and the second feature vector to obtain the fused feature;

[0058] The training module is used to input the fused features into the deep learning model for training to obtain a wheat nutrient demand prediction model;

[0059] The nutrient regulation module is used to regulate the nutrients of target wheat using a wheat nutrient demand prediction model.

[0060] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the above-described nutrient synergistic regulation system and control method for water-saving wheat cultivation.

[0061] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0062] This invention relates to a nutrient synergistic regulation system and control method for water-saving wheat cultivation. Compared with the prior art, this invention uses a random forest algorithm, an image feature extraction network, and a deep learning model, which can fully explore and learn the complex relationship between the soil environment and the wheat growth status, more accurately reflect the actual needs of wheat, achieve precise water and fertilizer ratio, and thus improve water and fertilizer utilization efficiency.

[0063] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 The flowchart of a method for synergistic nutrient regulation in water-saving wheat cultivation provided by the present invention.

[0066] Figure 2 This is a comparative diagram of dry matter accumulation in water-saving wheat cultivation provided by the present invention; where CK represents no nutrient synergistic regulation, M represents nutrient synergistic regulation completed based on experience, and MF1 represents the nutrient synergistic regulation method of the present invention. Detailed Implementation

[0067] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0068] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0069] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0070] Please see Figure 1 A method for synergistic regulation and control of nutrients in water-saving wheat cultivation, comprising:

[0071] Step 1: Use images of soil moisture, nutrient content, and wheat growth status;

[0072] Step 2: Use the random forest algorithm to map soil moisture and nutrient content into the first feature vector;

[0073] In step 2, the first step is to collect data on soil moisture and nutrient content. Each set of data should include soil moisture values, the content of different nutrients (such as nitrogen, phosphorus, and potassium), and environmental parameters that may affect these factors (such as temperature and pH). Additionally, the wheat growth status associated with this data is also required for subsequent training.

[0074] Soil moisture and nutrient content are standardized or normalized to unify the dimensions. A random forest model is used to predict each input sample (i.e., the standardized or normalized sample) and the output is the first feature vector, which is the model's mapping result of soil moisture and nutrient content.

[0075] Step 3: Enhance the wheat growth status image to obtain the enhanced growth status image;

[0076] Step 3 includes:

[0077] Step 3.1: Perform edge enhancement processing on the wheat growth status image to obtain the preprocessed growth status image;

[0078] Step 3.2: Take a set of neighboring pixels centered on any point on the original growth state image;

[0079] Step 3.3: Use the preprocessed growth state image as the guide image to calculate the enhancement component of the center point; the formula for calculating the enhancement component is:

[0080]

[0081] in, Indicates the normalized weights. This represents the augmentation component at point p. express The coordinates of point p on the image, express The coordinates of point q on the image, This represents the pixel value at point q in the preprocessed growth state image. This represents the set of neighboring pixels centered at pixel p. This represents the standard deviation of the Gaussian filter.

[0082] Step 3.4: Enhance the wheat growth status image using the enhancement component to obtain the enhanced growth status image;

[0083] Step 3.4 includes:

[0084] Step 3.4.1: Iteratively compensate the preprocessed growth state image using enhancement components to obtain an iterative image; wherein the iterative compensation process is as follows:

[0085]

[0086] in, Indicates the iterative weights. This represents the enhancement component at point p after the (t+1)th iteration. This represents the enhancement component at point p after the t-th iteration. This represents the enhancement component at point q after the t-th iteration. This represents the set of neighboring pixels centered at pixel p. Indicates the preset weight;

[0087] Step 3.4.2: Convolve the iterative image to obtain the enhanced growth state image; wherein, the enhanced growth state image is:

[0088]

[0089] Convolutional enhancement is applied to the iterative image to obtain the enhanced growth state image; where, This represents the enhanced growth state image. This represents convolution.

[0090] The guide image is the output image from the previous iteration, while the input image is the preprocessed growth status image. With each iteration, the gradually clearer large-scale structures are restored and highlighted, preserving edges and enhancing details in the growth status image, ultimately forming the final image. As the number of iterations increases, details are gradually magnified and enhanced, allowing even finer features in the image to be clearly presented. For example, subtle changes in wheat leaves, color differences, and signs of pests and diseases can all be highlighted after multiple iterations. This enhancement of image details helps identify more refined growth status features, which is crucial for predicting wheat's nutritional needs, health status, and subsequent management decisions (such as irrigation and fertilization).

[0091] Step 4: Use an image feature extraction network to extract the second feature vector of the enhanced growth state image;

[0092] Step 4 includes:

[0093] Step 4.1: Input the enhanced grown image into the convolutional layer and pooling layer sequentially to obtain the global feature map; the formula for extracting the global feature map is:

[0094]

[0095] in, For global feature maps, To enhance the post-growth state image, , , These represent convolution operations using kernels of different sizes. This indicates concatenation along the channel dimension. This represents the max pooling operation;

[0096] Step 4.2: Use 3D convolution to extract spatial attention features and channel attention features from the enhanced grown state image;

[0097] In step 4.2, the formula is used:

[0098]

[0099]

[0100] Spatial attention features and channel attention features are extracted from the enhanced growth state image; among them, This represents the image showing the growth state after enhancement. Representing spatial attention characteristics, Indicates channel attention features, For 3D convolution, For three-dimensional average pooling, For pooling, To modify the activation function of the linear unit, This represents the Sigmoid activation function. , These are the attention weight parameters.

[0101] Step 4.3: Fuse the global feature map, spatial attention features, and channel attention features to obtain the second feature vector.

[0102] In step 4.3, the formula is used:

[0103]

[0104]

[0105] The second feature vector is obtained by fusing the global feature map, spatial attention features, and channel attention features. .

[0106] Traditional 3D CNN detection networks primarily rely on stacked convolutions for feature extraction, which lacks sufficient discriminative power for small, low-contrast nodules. This invention introduces channel and spatial attention mechanisms, enabling the network to adaptively calibrate and focus on key features, significantly enhancing the feature extraction network's feature representation capabilities and detection sensitivity.

[0107] Step 5: Concatenate the first feature vector and the second feature vector to obtain the fused feature;

[0108] Step 6: Input the fused features into the deep learning model for training to obtain the wheat nutrient demand prediction model;

[0109] In step 6, the fused features are input into a deep learning model for training to obtain a wheat nutrient requirement prediction model. The deep learning model incorporates a water-fertilizer coupling coefficient.

[0110]

[0111] in, For soil moisture, The coupling strength parameter is preferably set to a value between 0.5 and 2.0; The optimal soil moisture threshold is determined based on the wheat variety and growth stage, and is typically 60% to 80% of field capacity.

[0112] Step 6.2: Construct the loss function of the deep learning model using the water-fertilizer coupling coefficient; wherein the loss function is:

[0113]

[0114] in, To reflect actual nutrient requirements, To predict nutrient requirements, For the sample size, For model parameters, and The regularization coefficient is . Let KL divergence be the KL divergence. and The first The empirical and prior distributions of layer features are analyzed. The loss function is minimized using the backpropagation algorithm, and the model parameters are updated.

[0115] Step 7: Use a wheat nutrient requirement prediction model to regulate the nutrients of the target wheat.

[0116] In step 7, the nutrient requirements of the target wheat are regulated using a wheat nutrient requirement prediction model. This is based on the predicted nutrient requirements. and current soil nutrient content Calculate the amount of fertilizer. :

[0117]

[0118] in, This is the fertilizer utilization rate coefficient, with a value range of [value range missing]. The specific amount depends on the type of fertilizer and the method of application.

[0119] According to soil moisture and crop evapotranspiration Calculate irrigation amount :

[0120]

[0121] in, Target soil moisture, Soil water holding capacity coefficient, This is to determine the effective rainfall. By accurately calculating fertilizer and irrigation amounts, synergistic regulation of water and fertilizer can be achieved, thereby improving water and fertilizer utilization efficiency.

[0122] like Figure 2 As shown, this invention enables precise control of water-saving cultivation by accurately predicting and adjusting nutrient supply, combined with comprehensive data from soil moisture, nutrient content, and growth status images. The data in the images (e.g., higher dry matter accumulation in wheat at maturity under MF1 nutrient management) demonstrate that this invention achieves precise and effective water and nutrient management by combining advanced data-driven technologies (such as deep learning and random forests) with crop physiological data.

[0123] This invention employs a random forest algorithm, an image feature extraction network, and a deep learning model to fully explore and learn the complex relationship between the soil environment and wheat growth status, more accurately reflect the actual needs of wheat, achieve precise water and fertilizer ratio, and thus improve water and fertilizer utilization efficiency.

[0124] This invention also provides a nutrient synergistic regulation and control system for water-saving wheat cultivation, comprising:

[0125] The data acquisition module is used to collect images of soil moisture, nutrient content, and the growth status of wheat.

[0126] The feature mapping module is used to map soil moisture and nutrient content into a first feature vector using the random forest algorithm;

[0127] The enhancement module is used to enhance the wheat growth status image to obtain an enhanced growth status image;

[0128] The feature extraction module is used to extract the second feature vector of the enhanced growth state image using an image feature extraction network;

[0129] The feature fusion module is used to fuse the first feature vector and the second feature vector to obtain the fused feature;

[0130] The training module is used to input the fused features into the deep learning model for training to obtain a wheat nutrient demand prediction model;

[0131] The nutrient regulation module is used to regulate the nutrients of target wheat using a wheat nutrient demand prediction model.

[0132] Compared with the prior art, the beneficial effects of the nutrient synergistic regulation and control system in water-saving wheat cultivation provided by the present invention are the same as the beneficial effects of the nutrient synergistic regulation and control method in water-saving wheat cultivation described above, and will not be repeated here.

[0133] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus. The computer program, when executed by the processor, implements the steps of the aforementioned nutrient synergistic regulation system and control method for water-saving wheat cultivation. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as those of the aforementioned nutrient synergistic regulation control method for water-saving wheat cultivation, and will not be elaborated further here.

[0134] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps in the above-mentioned nutrient synergistic regulation system and control method for water-saving wheat cultivation. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as the beneficial effects of the above-mentioned nutrient synergistic regulation control method for water-saving wheat cultivation, and will not be elaborated here.

[0135] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for synergistic regulation and control of nutrients in water-saving wheat cultivation, characterized in that, include: Step 1: Collect images of soil moisture, nutrient content, and wheat growth status; Step 2: Use the random forest algorithm to map soil moisture and nutrient content into the first feature vector; Step 3: Enhance the wheat growth status image to obtain the enhanced growth status image; Step 4: Use an image feature extraction network to extract the second feature vector of the enhanced growth state image; Step 5: Fuse the first feature vector and the second feature vector to obtain the fused feature; Step 6: Input the fused features into the deep learning model for training to obtain the wheat nutrient demand prediction model; Step 7: Use a wheat nutrient requirement prediction model to regulate the nutrients of the target wheat.

2. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 1, characterized in that, Step 3: Enhance the wheat growth status image to obtain an enhanced growth status image, including: Step 3.1: Perform edge enhancement processing on the wheat growth status image to obtain the preprocessed growth status image; Step 3.2: Take a set of neighboring pixels centered on any point on the original growth state image; Step 3.3: Use the preprocessed growth state image as the guide image to calculate the enhancement component of the center point; Step 3.4: Enhance the wheat growth status image using the enhancement component to obtain the enhanced growth status image.

3. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 2, characterized in that, The formula for calculating the enhanced component is: in, Indicates the normalized weights. This represents the augmentation component at point p. express The coordinates of point p on the image, express The coordinates of point q on the image, This represents the pixel value at point q in the preprocessed growth state image. This represents the set of neighboring pixels centered at pixel p. This represents the standard deviation of the Gaussian filter.

4. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 3, characterized in that, Step 3.4 includes: Step 3.4.1: Iteratively compensate the preprocessed growth state image using enhancement components to obtain an iterative image; wherein the iterative compensation process is as follows: in, Indicates the iterative weights. This represents the enhancement component at point p after the (t+1)th iteration. This represents the enhancement component at point p after the t-th iteration. This represents the enhancement component at point q after the t-th iteration. This represents the set of neighboring pixels centered at pixel p. Indicates the preset weight; Step 3.4.2: Convolve the iterative image to obtain the enhanced growth state image.

5. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 4, characterized in that, In step 3.4.2, the formula is used: Convolution enhancement is applied to the iterative image to obtain the enhanced growth state image; where, This represents the enhanced growth state image. This represents convolution.

6. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 5, characterized in that, Step 4: Extracting the second feature vector of the enhanced growth state image using an image feature extraction network, including: Step 4.1: Input the enhanced grown image into the convolutional layer and pooling layer sequentially to obtain the global feature map; the formula for extracting the global feature map is: in, For global feature maps, To enhance the post-growth state image, , , These represent convolution operations using kernels of different sizes. This indicates concatenation along the channel dimension. This represents the max pooling operation; Step 4.2: Use 3D convolution to extract spatial attention features and channel attention features from the enhanced grown state image; Step 4.3: Fuse the global feature map, spatial attention features, and channel attention features to obtain the second feature vector.

7. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 6, characterized in that, Step 4.2: Extract spatial attention features and channel attention features of the enhanced grown state image using 3D convolution, including: Formula used: Spatial attention features and channel attention features are extracted from the enhanced growth state image; among them, This represents the image showing the growth state after enhancement. Representing spatial attention characteristics, Indicates channel attention features, For 3D convolution, For three-dimensional average pooling, For pooling, To modify the activation function of the linear unit, This represents the Sigmoid activation function. , These are the attention weight parameters.

8. The method for controlling the synergistic regulation of nutrients in water-saving cultivation of wheat according to claim 7, characterized by, In step 4.3, the formula is used: The second feature vector is obtained by fusing the global feature map, spatial attention features, and channel attention features. .

9. The method for synergistic regulation and control of nutrients in water-saving wheat cultivation according to claim 8, characterized in that, In step 6, the training steps of the deep learning model include: Step 6.1: Introduce a water-fertilizer coupling coefficient into the deep learning model; whereby the water-fertilizer coupling coefficient is: in, For soil moisture, For coupling strength parameters, The optimal soil moisture threshold; Step 6.2: Construct the loss function of the deep learning model using the water-fertilizer coupling coefficient; wherein the loss function is: in, To reflect actual nutrient requirements, To predict nutrient requirements, For the sample size, For model parameters, and The regularization coefficient is . Let KL divergence be the KL divergence. and The first The empirical and prior distributions of layer features.

10. A nutrient synergistic regulation and control system for water-saving wheat cultivation, characterized in that, include: The data acquisition module is used to collect images of soil moisture, nutrient content, and the growth status of wheat. The feature mapping module is used to map soil moisture and nutrient content into a first feature vector using the random forest algorithm; The enhancement module is used to enhance the wheat growth status image to obtain an enhanced growth status image; The feature extraction module is used to extract the second feature vector of the enhanced growth state image using an image feature extraction network; The feature fusion module is used to fuse the first feature vector and the second feature vector to obtain the fused feature; The training module is used to input the fused features into the deep learning model for training to obtain a wheat nutrient demand prediction model; The nutrient regulation module is used to regulate the nutrients of target wheat using a wheat nutrient demand prediction model.

Images