A seawall front vegetation intelligent irrigation control method and system

Abstract

The application provides a seawall front vegetation intelligent irrigation control method and system, and relates to the technical field of vegetation watering.The method comprises the following steps: constructing a seawall front irrigation system; acquiring a vegetation image; identifying a vegetation growth stage through a convolutional neural network according to the vegetation image; dividing vegetation regions according to vegetation types and the identified vegetation growth stage; predicting the required unit area water supply of each vegetation region through an artificial neural network according to soil granularity, average air temperature, precipitation, air humidity, a normalized water body index, vegetation types and the vegetation growth stage; and irrigating the vegetation in each vegetation region through the seawall front irrigation system according to the required unit area water supply of each vegetation region.The application can provide the best growth conditions for seawall front vegetation, promote the health and growth of the vegetation, help maintain and improve the stability of the ecosystem in the seawall front region, and support biodiversity.

Description

Technical Field

[0001] This invention relates to the field of vegetation irrigation technology, and in particular to an intelligent irrigation control method and system for vegetation in front of seawalls. Background Technology

[0002] Vegetation measures in front of seawalls are an important way to ecologically transform seawalls. They are an effective measure to address the current situation of "smooth surfaces on three sides" in my country's seawalls, increase the porosity of the revetment structure, thereby improving coastal habitats and enriching biodiversity. The stems and leaves of plants can also increase the roughness of the revetment structure, improve the flood control standard by reducing wave run-up, and thus achieve ecological disaster reduction functions.

[0003] To provide suitable moisture conditions for the vegetation in front of the seawall, irrigation is necessary. Currently, flood irrigation and manual irrigation are the main methods used to irrigate the vegetation in front of the seawall.

[0004] However, different types of vegetation have different water requirements, and vegetation at different growth stages also has different water requirements. Flood irrigation cannot distinguish between different types of vegetation in different areas, which can easily lead to excessive or insufficient water supply, affecting the healthy growth of vegetation. On the other hand, artificial irrigation is time-consuming and labor-intensive, and it is difficult to accurately control the amount of water to be given to different types of vegetation at different growth stages, which also affects the healthy growth of vegetation. Summary of the Invention

[0005] To address the technical problem that current technologies struggle to accurately control the amount of water needed for different types and growth stages of vegetation, thus affecting the healthy growth of vegetation, this invention provides an intelligent irrigation control method and system for vegetation in front of seawalls.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] First aspect:

[0008] This invention provides an intelligent irrigation control method for vegetation in front of a seawall, comprising:

[0009] S1: Construct an irrigation system in front of the seawall;

[0010] S2: Acquire vegetation images;

[0011] S3: Based on the vegetation image, identify the vegetation growth stage using a convolutional neural network;

[0012] S4: Divide vegetation areas according to vegetation types and identified vegetation growth stages;

[0013] S5: Based on soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type and vegetation growth stage, an artificial neural network is used to predict the water supply required per unit area for each vegetation region.

[0014] S6: Irrigate the vegetation in each vegetation area according to the water supply required per unit area for each vegetation area through the seawall irrigation system.

[0015] The second aspect:

[0016] An embodiment of the present invention provides an intelligent irrigation control system for vegetation in front of a seawall, comprising:

[0017] processor;

[0018] A memory storing computer-readable instructions, which, when executed by the processor, implement the intelligent irrigation control method for vegetation in front of the seawall as described in the first aspect.

[0019] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0020] In this invention, based on soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type, and vegetation growth stage, an artificial neural network is used to predict the required water supply per unit area for each vegetation zone. Through the seawall irrigation system, intelligent irrigation is carried out on the vegetation in each zone, reducing the cost of manual operation and management, providing optimal growth conditions for the vegetation in front of the seawall, promoting the health and growth of the vegetation, helping to maintain and improve the ecosystem stability of the area in front of the seawall, and supporting biodiversity. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0022] Figure 1 A flowchart illustrating an intelligent irrigation control method for vegetation in front of a seawall, provided as an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of a method for intelligent irrigation control of vegetation in front of a seawall, provided in an embodiment of the present invention.

[0024] Figure 3 This is a schematic diagram of a smart irrigation control system for vegetation in front of a seawall, provided as an embodiment of the present invention. Detailed Implementation

[0025] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0026] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0027] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0028] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0029] Reference manual attached Figure 1 The diagram shows a flowchart of an intelligent irrigation control method for vegetation in front of a seawall, provided by an embodiment of the present invention.

[0030] Reference manual attached Figure 2 The diagram shows a structural schematic of an intelligent irrigation control method for vegetation in front of a seawall, provided by an embodiment of the present invention.

[0031] This invention provides a method for intelligent irrigation control of vegetation in front of a seawall. This method can be implemented using an intelligent irrigation control device for vegetation in front of the seawall, which can be a terminal or a server. The processing flow of the intelligent irrigation control method for vegetation in front of the seawall may include the following steps:

[0032] S1: Construct an irrigation system in front of the seawall.

[0033] In one possible implementation, the seawall irrigation system specifically includes: a water intake channel, a sedimentation tank, a storage tank, hoses, and valves. The water intake channel outside the seawall serves as the water source. Seawater, after sedimentation in the sedimentation tank, enters the storage tank. The storage tank is connected to the vegetation planting area via hoses. Irrigation holes are made in the hoses, and valves are installed at these holes. Drip irrigation is performed on the vegetation by opening the valves.

[0034] It should be noted that the irrigation system utilizes the water diversion channel outside the seawall as a water source, directly drawing seawater from the ocean. This design takes advantage of the abundant seawater resources in coastal areas, reducing reliance on freshwater resources. Especially in areas with scarce freshwater resources, this method can significantly alleviate water shortage problems.

[0035] Furthermore, before entering the reservoir, the seawater passes through a sedimentation tank to remove silt, impurities, and suspended solids, preventing these impurities from entering the irrigation system and thus preventing blockages in irrigation pipes and holes, ensuring the long-term stable operation of the irrigation system.

[0036] Furthermore, the reservoir can store settled seawater, regulate water volume, and ensure a stable water supply for the irrigation system at different times. Especially during tidal changes or unstable water diversion, the reservoir can act as a buffer to guarantee a continuous water supply.

[0037] Furthermore, the design of the hoses and irrigation orifices allows the drip irrigation system to deliver water directly to the roots of plants, reducing water evaporation and runoff loss. Drip irrigation offers better control over water volume, meeting the needs of different vegetation areas and is particularly suitable for intensive management.

[0038] Furthermore, the valves can flexibly adjust the irrigation volume in different areas according to actual needs, ensuring that different plants or vegetation at different growth stages receive adequate water supply. Through an automated control system, the valves can also be adjusted in real time, further improving the precision of irrigation.

[0039] S2: Obtain vegetation images.

[0040] Alternatively, vegetation images can be acquired via satellite or via a wide-angle camera. This invention does not limit the specific method of acquiring vegetation images.

[0041] S3: Based on the vegetation images, identify the vegetation growth stages using a convolutional neural network.

[0042] Convolutional Neural Networks (CNNs) are powerful and widely used deep learning models, particularly adept at processing image data. In vegetation growth stage identification tasks, CNNs can effectively extract image features and automatically identify the growth stage of vegetation, providing technical support for precision agriculture and ecological management.

[0043] In one possible implementation, the convolutional neural network includes: a backbone feature extraction network, a feature fusion network, and a vegetation growth stage detection head. S3 specifically includes sub-steps S301 to S303:

[0044] S301: Extract multi-scale feature maps from vegetation images using a backbone feature extraction network.

[0045] It should be noted that the backbone feature extraction network can extract multi-scale feature maps from vegetation images. Feature maps of different scales can capture information at different levels in the image. Small-scale feature maps capture details, such as the edges and textures of leaves, while large-scale feature maps capture global features, such as the morphology and structure of vegetation. Multi-scale feature extraction can preserve the multi-level information of the image, allowing the model to make full use of this information in subsequent processing, thereby improving its ability to understand and process complex scenes.

[0046] Optionally, the backbone feature extraction network includes concatenated convolutional units, namely a first convolutional unit, a second convolutional unit, and a third convolutional unit. S301 specifically includes:

[0047] S3011: Using the vegetation image as the input to the first convolutional unit, the small-scale feature map in the vegetation image is output through the first convolutional unit.

[0048] Optionally, S3011 specifically includes: performing convolution processing on the fabric image through a 3×3 first convolutional layer to obtain an initial feature map.

[0049] The initial feature map is copied to obtain the first initial feature map and the second initial feature map.

[0050] The initial feature map is separated by channels to obtain a first branch feature map and a second branch feature map. The second branch feature map is then convolved sequentially through a first grouped convolutional layer, a depthwise separable convolutional layer, and a second grouped convolutional layer to obtain a second branch separated feature map. Grouped convolution and depthwise separable convolution can significantly reduce model parameters and computational cost while retaining sufficient feature extraction capability. Especially in deep networks, this design can alleviate computational burden and improve model efficiency.

[0051] The first branch feature map is concatenated with the second branch separation feature map to obtain a connected feature map. Channel shuffling is then performed on the connected feature map to obtain a shuffled feature map. Through operations such as separation, concatenation, and shuffling, information between channels is fully exchanged and fused, enabling the network to capture the complex relationships between features in the image and improving the expressive power of the feature maps.

[0052] The second initial feature map is symmetrically flipped to obtain a flipped feature map. This symmetrical flipping operation enables the model to process images from different directions or angles. Even when an image is flipped or rotated in a real-world scene, the model can still recognize its features. This improves the model's performance in various situations, especially when the input image may have multiple viewpoints.

[0053] The shuffled feature map and the flipped feature map are multiplied element-wise, and then convolved through a 3×3 second convolutional layer to obtain the enhanced feature map. Performing operations such as symmetrical flipping and element-wise multiplication on the feature map enhances the model's robustness to image transformations. This allows the model to accurately identify and classify input images from different angles or with deformations.

[0054] The enhanced feature map is added to the initial feature map, and then convolved through a 3×3 third convolutional layer to obtain a small-scale feature map. By adding the processed feature map to the initial feature map, the original key information can be preserved while incorporating the enhanced new features, avoiding information loss and enhancing the representativeness of the features.

[0055] S3012: Using the small-scale feature map as the input to the second convolutional unit, the second convolutional unit outputs the medium-scale feature map in the vegetation image.

[0056] Optionally, S3012 specifically includes: separating the small-scale feature map according to channels to obtain a first small-scale branch feature map and a second small-scale branch feature map. The second small-scale branch feature map is then convolved sequentially through a first grouped convolutional layer, a depthwise separable convolutional layer, and a second grouped convolutional layer to obtain a second small-scale branch separation feature map. The first small-scale branch feature map and the second small-scale branch separation feature map are then convolved to obtain a connected feature map. The connected feature map is then channel-shuffled to obtain a shuffled feature map. The shuffled feature map is added to the small-scale feature map and then convolved through a 3×3 third convolutional layer to obtain a medium-scale feature map.

[0057] S3013: Using the mesoscale feature map as input to the third convolutional unit, the third convolutional unit outputs the large-scale feature map of the vegetation image.

[0058] Optionally, S3013 specifically includes: separating the mesoscale feature map according to channels to obtain a first mesoscale branch feature map and a second mesoscale branch feature map. The second mesoscale branch feature map is then convolved sequentially through a first grouped convolutional layer, a depthwise separable convolutional layer, and a second grouped convolutional layer to obtain a second mesoscale branch separation feature map. The second mesoscale branch feature map and the second mesoscale branch separation feature map are then convolved to obtain a connected feature map. The connected feature map is then channel-washed to obtain a shuffled feature map. The shuffled feature map is added to the mesoscale feature map and then convolved through a 3×3 third convolutional layer to obtain a large-scale feature map.

[0059] In this invention, each level of feature map represents different layers of information in the image. Small-scale feature maps may contain more details, medium-scale feature maps integrate more local context, while large-scale feature maps capture global features. Through progressive convolution operations, starting with small-scale feature maps, the network gradually extracts medium-scale and large-scale feature maps. This design can capture detailed features (such as texture and edges) to more global features (such as shape and structure) from the image. This local-to-global feature extraction approach enables the network to progressively understand the overall semantics and details of the image.

[0060] S302: Multi-scale feature maps are fused using a feature fusion network to obtain a fused feature map.

[0061] It should be noted that fusing multi-scale feature maps through a feature fusion network can effectively combine detailed and global information to form a more comprehensive and robust feature representation. This fusion helps the model to more accurately understand the overall growth status of vegetation, rather than just local features.

[0062] In one possible implementation, the feature fusion network is specifically a feature pyramid network, and step S302 specifically includes: upsampling the large-scale feature map to match its size with that of the medium-scale feature map; performing weighted fusion of the upsampled large-scale and medium-scale feature maps to obtain a preliminary fused feature map; upsampling the preliminary fused feature map to match its size with that of the small-scale feature map; and performing weighted fusion of the upsampled preliminary fused feature map and the small-scale feature map to obtain a fused feature map.

[0063] In this invention, by fusing large-scale, medium-scale, and small-scale feature maps, global information and local detail information of an image can be combined. Large-scale feature maps capture the overall structure of the image, while small-scale feature maps retain more detailed information. By fusing these features, the feature pyramid network enables the model to utilize information at different scales simultaneously, thereby better understanding the content of the image. Furthermore, during multi-scale fusion, due to the hierarchical structure of the feature pyramid network, gradients propagate more smoothly, reducing the risk of gradient vanishing or exploding, resulting in more stable model training and faster convergence.

[0064] S303: Based on the fused feature map, the vegetation growth stage is determined using the vegetation growth stage detection head.

[0065] The vegetation growth stages include: seedling stage, growth stage, reproductive stage, fruit development stage, and decline stage.

[0066] It should be noted that, based on the fused feature map, the vegetation growth stage detection head can more accurately identify and classify the growth stages of vegetation. Because the fused feature map integrates information from different scales, the detection head can better distinguish the subtle differences between different growth stages of vegetation.

[0067] In one possible implementation, S303 specifically involves: calculating the probability that the vegetation image belongs to each growth stage based on the fused feature map, and using the growth stage with the highest probability value as the detection result.

[0068]

[0069] in, P This represents the probability that a vegetation image belongs to each growth stage. p i Indicates that the vegetation image belongs to the first i The probability of each growth stage, where Softmax represents the Softmax activation function. W Represents the weight matrix. V r Represents the fused feature map. b This indicates the bias term.

[0070] It should be noted that by calculating the probability of each growth stage using the Softmax function, the detection head can quantify the likelihood of an image belonging to each growth stage.

[0071] In this invention, by fusing information at different scales, the detection head can better process complex features in images, especially when the differences in certain growth stages are small or the features are not obvious, it can still make accurate judgments.

[0072] S4: Divide vegetation areas according to vegetation types and identified vegetation growth stages.

[0073] Among the vegetation commonly planted in front of the seawall are seaside paspalum, Spartina alterniflora, Casuarina equisetifolia, Verbena officinalis, Reed, Suaeda salsa, and Robinia pseudoacacia, etc.

[0074] Specifically, we can consider vegetation A in its seedling stage as one region, vegetation A in its growth stage as another region, and vegetation B in its seedling stage as yet another region.

[0075] It should be noted that different types and growth stages of vegetation have different requirements for water, nutrients, and light. Dividing vegetation into different zones according to species and growth stage allows for precise irrigation tailored to the specific needs of each zone, providing the most suitable growing conditions for each stage of vegetation and promoting healthy growth and development.

[0076] S5: Based on soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type, and vegetation growth stage, an artificial neural network is used to predict the required water supply per unit area for each vegetation region.

[0077] Artificial Neural Networks (ANNs) are computational models inspired by biological neural networks, widely used in fields such as pattern recognition, classification, regression, image processing, and speech recognition. ANNs process complex input data by simulating the connections and signal transmission between neurons in the brain, and learn to extract features from the data and build models to perform prediction or decision-making tasks.

[0078] It should be noted that different vegetation regions have different water requirements due to variations in soil conditions, climate, vegetation species, and growth stages. ANN can combine this multi-dimensional information to calculate the precise water supply per unit area for each region, enabling personalized precision irrigation.

[0079] In one possible implementation, the artificial neural network includes an input layer, a hidden layer, an output layer, and a prediction layer. S5 specifically includes sub-steps S501 to S503:

[0080] S501: In the input layer, the input state vector includes multiple state parameters, specifically including: soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type, and vegetation growth stage.

[0081] S502: In each neuron of the hidden layer, the input state vector is weighted and summed to obtain the hidden state vector.

[0082]

[0083] in, y j Indicates the first j The hidden state output by each hidden layer neuron. s 1 represents the hidden layer activation function. W j Indicates the first j The weight vector of each hidden layer neuron. T This indicates the transpose operation. X Represents the input state vector. b j Indicates the first j The bias term of each hidden layer neuron. oh ij Indicates the first j The th hidden layer neuron i The weights of each state parameter,x i Indicates the first i Each state parameter value, n This indicates the total number of state parameters.

[0084] S503: Each neuron in the output layer outputs a predicted value for water supply per unit area.

[0085]

[0086] in, a This represents the predicted water supply per unit area. s 2 represents the output layer activation function. oh j Indicates the first j The connection weights between hidden layer neurons and the output layer b f This represents the bias term of the output layer. m This represents the total number of neurons in the hidden layer.

[0087] This invention utilizes artificial neural networks to predict water supply per unit area in vegetated regions. By combining multi-dimensional environmental data and vegetation information, precise, dynamic, and automated irrigation management can be achieved. This not only improves water resource utilization efficiency and promotes healthy vegetation growth but also enhances overall management efficiency and ecological benefits, thus providing strong support for sustainable development.

[0088] Gradient descent can be used to train artificial neural networks.

[0089] Specifically, a mean squared error loss function for the predicted water supply can be constructed, and then the network parameters of the artificial neural network can be updated according to the following formula:

[0090]

[0091] in, i t+1 Indicates the first t Model parameters at +1 iteration i t Indicates the first t Model parameters at the next iteration i t-1 Indicates the first t Model parameters at -1st iteration or t Indicates the first t Adaptive learning rate at the next iteration L Represents the loss function. Represents the loss function L For network parameters it gradient, β This represents the momentum coefficient.

[0092] It should be noted that the introduction of the momentum term helps to reduce the oscillations in parameter updates during training, especially when facing noisy gradients or complex loss surfaces, and can provide a smoother optimization path.

[0093] Optionally, the adaptive learning rate is calculated as follows:

[0094]

[0095] in, or min This represents the minimum learning rate. or max This represents the maximum learning rate. T max denoted by , where represents the maximum number of iterations, and cos represents the cosine function.

[0096] In this invention, a higher learning rate in the early stages of training helps to quickly find the convergence direction; a lower learning rate in the later stages of training helps to finely optimize the model parameters. By dynamically adjusting the learning rate, the model can be exposed to different learning rate ranges during training, thereby helping to improve the model's generalization ability and reduce the risk of overfitting.

[0097] S6: Irrigate the vegetation in each vegetation area according to the water supply required per unit area in each vegetation area through the seawall irrigation system.

[0098] In one possible implementation, S6 specifically includes sub-steps S601 to S607:

[0099] S601: Calculate the total water supply based on the water supply per unit area required by each vegetation zone:

[0100]

[0101] in, A Indicates the total water supply. a k Indicates the first k The water supply required per unit area for each vegetated region S k Indicates the first k The area of ​​each vegetation zone.

[0102] S602: Determine the initial opening parameters of the reservoir gate based on the total water supply:

[0103]

[0104] in,β This indicates the initial opening parameter of the reservoir gate. b This represents the opening coefficient.

[0105] S603: Control the gate of the water storage tank according to the initial opening parameters of the gate.

[0106] In this invention, the initial opening parameters of the gate are determined by calculating the total water supply. This enables the system to automatically adjust the gate opening, avoiding the errors and complexity of manual adjustment and improving the overall intelligence level of the irrigation system. Reasonably determining the gate opening ensures a stable and efficient water supply to each vegetated area, reducing potential resistance and waste during water supply, thereby lowering the energy consumption of pumps and other equipment.

[0107] S604: During the irrigation process, based on the PID controller, the opening parameters of the reservoir gate are adjusted in real time to achieve the target total water supply.

[0108] Specifically, for a PID controller, the difference between the real-time water supply and the total water supply is calculated. e .

[0109] Calculate the proportional term, integral term, and differential term based on the difference between the hourly water supply and the total water supply:

[0110]

[0111] in,[ P ] indicates the proportion term, [ I ] represents the integral term, [ D ] represents the differential term. K p Represents the proportional gain coefficient. K i Represents the integral gain coefficient. K d Represents the differential gain coefficient. e This represents the difference between the real-time water supply and the total water supply. t Indicates time, dt Indicates to t Perform differentiation;

[0112] Calculate the PID control parameters based on the proportional, integral, and derivative terms:

[0113]

[0114] in, PID Indicates PID control parameters;

[0115] The opening parameters of the reservoir gate are adjusted in real time based on the PID control parameters.

[0116] In this invention, the PID controller can dynamically adjust based on the difference between the real-time water supply and the target total water supply, ensuring that the actual water supply closely follows demand. This real-time adjustment mechanism enables the system to respond quickly to any changes in water supply demand, guaranteeing the accuracy of water supply.

[0117] S605: Determine the initial valve opening parameters based on the required water supply per unit area and the number of irrigation holes in each vegetation zone:

[0118]

[0119] in, l k Indicates the first k The initial opening parameters of the valves within a vegetated area. c k Indicates the first k The number of irrigation holes contained in a vegetated area.

[0120] S606: Control the valves in each vegetation zone according to the initial valve opening parameters.

[0121] In this invention, by incorporating the number of irrigation holes into the opening parameter calculation, the system can evenly distribute water resources according to the actual distribution of irrigation holes, ensuring that the water output of each irrigation hole is consistent and avoiding uneven water distribution caused by differences in the number of irrigation holes.

[0122] S607: During the irrigation process, based on the PID controller, the valve opening parameters of each vegetation area are adjusted in real time to achieve the required water supply per unit area for each vegetation area.

[0123] The PID controller that controls the opening parameters of valves in each vegetation zone operates on the same principle as the PID controller that controls the opening parameters of the reservoir gate. To avoid repetition, this invention will not elaborate further.

[0124] In this invention, the PID controller can respond in real time to changes in the actual water supply demand of each vegetation area during irrigation, ensuring that each area always receives a precise water supply by adjusting the valve opening. This ensures that the vegetation in each area receives appropriate moisture, preventing over- or under-irrigation.

[0125] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0126] In this invention, based on soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type, and vegetation growth stage, an artificial neural network is used to predict the required water supply per unit area for each vegetation zone. Through the seawall irrigation system, intelligent irrigation is carried out on the vegetation in each zone, reducing the cost of manual operation and management, providing optimal growth conditions for the vegetation in front of the seawall, promoting the health and growth of the vegetation, helping to maintain and improve the ecosystem stability of the area in front of the seawall, and supporting biodiversity.

[0127] Reference manual attached Figure 3 The diagram shows a structural schematic of an intelligent irrigation control system for vegetation in front of a seawall provided by the present invention.

[0128] The present invention also provides an intelligent irrigation control system 20 for vegetation in front of a seawall, comprising:

[0129] Processor 201;

[0130] The memory 202 stores computer-readable instructions, which, when executed by the processor 201, implement the intelligent irrigation control method for vegetation in front of the seawall as described in the method embodiment.

[0131] The intelligent irrigation control system 20 for vegetation in front of the seawall provided by the present invention can execute the above-mentioned intelligent irrigation control method for vegetation in front of the seawall and achieve the same or similar technical effects. To avoid repetition, the present invention will not elaborate further.

[0132] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0133] In this invention, based on soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type, and vegetation growth stage, an artificial neural network is used to predict the required water supply per unit area for each vegetation zone. Through the seawall irrigation system, intelligent irrigation is carried out on the vegetation in each zone, reducing the cost of manual operation and management, providing optimal growth conditions for the vegetation in front of the seawall, promoting the health and growth of the vegetation, helping to maintain and improve the ecosystem stability of the area in front of the seawall, and supporting biodiversity.

[0134] 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 technical scope 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.

[0135] The following points need to be explained:

[0136] (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0137] (2) For clarity, the thickness of a region or area is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0138] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0139] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for intelligent irrigation control of vegetation in front of a seawall, characterized in that, include: S1: Construct an irrigation system in front of the seawall; S2: Acquire vegetation images; S3: Based on the vegetation image, identify the vegetation growth stage using a convolutional neural network; S4: Divide vegetation areas according to vegetation types and identified vegetation growth stages; S5: Based on soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type and vegetation growth stage, an artificial neural network is used to predict the water supply required per unit area for each vegetation region. S6: Irrigate the vegetation in each vegetation area according to the water supply required per unit area for each vegetation area through the seawall irrigation system. The artificial neural network includes an input layer, a hidden layer, an output layer, and a prediction layer; S5 specifically includes: S501: In the input layer, input a state vector, which includes multiple state parameters, specifically including: soil particle size, average temperature, precipitation, air humidity, normalized water index, vegetation type, and vegetation growth stage. S502: For each neuron in the hidden layer, the input state vector is weighted and summed to obtain the hidden state vector: ； in, y j Indicates the first j The hidden state output by each hidden layer neuron. σ 1 represents the hidden layer activation function. W j Indicates the first j The weight vector of each hidden layer neuron. T This indicates the transpose operation. X Represents the input state vector. b j Indicates the first j The bias term of each hidden layer neuron. ω ij Indicates the first j The th hidden layer neuron i The weights of each state parameter, x i Indicates the first i Each state parameter value, n Indicates the total number of state parameters; S503: Each neuron in the output layer outputs a predicted value for water supply per unit area. ； in, This represents the predicted water supply per unit area. σ 2 represents the output layer activation function. ω j Indicates the first j The connection weights between hidden layer neurons and the output layer b f This represents the bias term of the output layer. m This represents the total number of neurons in the hidden layer.

2. The intelligent irrigation control method for vegetation in front of a seawall according to claim 1, characterized in that, The seawall irrigation system specifically includes: a water diversion channel, a sedimentation tank, a water storage tank, hoses, and valves; Using the water diversion channel outside the seawall as the water source, seawater enters the water storage tank after sedimentation in the sedimentation tank. The water storage tank is connected to the vegetation planting area through the hose. Irrigation holes are opened on the hose, and valves are installed at the irrigation holes. The vegetation is drip-irrigated by opening the valves.

3. The intelligent irrigation control method for vegetation in front of a seawall according to claim 1, characterized in that, The convolutional neural network includes: a backbone feature extraction network, a feature fusion network, and a vegetation growth stage detection head; S3 specifically includes: S301: Extract multi-scale feature maps from the vegetation image using a backbone feature extraction network; S302: Multi-scale feature maps are fused using a feature fusion network to obtain a fused feature map; S303: Based on the fused feature map, the vegetation growth stage is determined using the vegetation growth stage detection head.

4. The intelligent irrigation control method for vegetation in front of a seawall according to claim 3, characterized in that, The backbone feature extraction network includes convolutional units connected in series, namely a first convolutional unit, a second convolutional unit, and a third convolutional unit; S301 specifically includes: S3011: Using the vegetation image as the input to the first convolutional unit, the small-scale feature map in the vegetation image is output through the first convolutional unit. S3012: Using the small-scale feature map as the input of the second convolutional unit, the medium-scale feature map in the vegetation image is output through the second convolutional unit; S3013: Using the mesoscale feature map as the input to the third convolutional unit, the large-scale feature map in the vegetation image is output through the third convolutional unit.

5. The intelligent irrigation control method for vegetation in front of a seawall according to claim 4, characterized in that, S3011 specifically includes: The vegetation image is convolved using a 3×3 first convolutional unit to obtain an initial feature map; The initial feature map is copied to obtain a first initial feature map and a second initial feature map; The first initial feature map is separated according to channels to obtain a first branch feature map and a second branch feature map; The second branch feature map is processed by convolution through the first group convolution unit, the depthwise separable convolution unit, and the second group convolution unit in sequence to obtain the second branch separated feature map; The first branch feature map is connected to the second branch separation feature map to obtain a connection feature map; The connection feature map is subjected to channel shuffling to obtain a shuffling feature map; The second initial feature map is symmetrically flipped to obtain a flipped feature map; The shuffled feature map and the flipped feature map are multiplied element-wise and then convolved through a 3×3 second convolutional unit to obtain the enhanced feature map. The enhanced feature map is added to the initial feature map, and then convolved using a 3×3 third convolutional unit to obtain the small-scale feature map.

6. The intelligent irrigation control method for vegetation in front of a seawall according to claim 4, characterized in that, The feature fusion network is specifically a feature pyramid network, and S302 specifically includes: The large-scale feature map is upsampled to match the size of the medium-scale feature map. The upsampled large-scale feature map is weighted and fused with the mesoscale feature map to obtain a preliminary fused feature map. The preliminary fused feature map is upsampled to match the size of the small-scale feature map. The pre-fusion feature map after upsampling is weighted and fused with the small-scale feature map to obtain the fusion feature map.

7. The intelligent irrigation control method for vegetation in front of a seawall according to claim 4, characterized in that, Specifically, S303 is: Based on the fused feature map, the probability of the vegetation image belonging to each growth stage is calculated, and the growth stage with the highest probability value is taken as the detection result: ； in, P This indicates the probability that the vegetation image belongs to each growth stage. p i This indicates that the vegetation image belongs to the first... i The probability of each growth stage, where Softmax represents the Softmax activation function. W Represents the weight matrix. V r Represents the fused feature map. b This indicates the bias term.

8. The intelligent irrigation control method for vegetation in front of a seawall according to claim 2, characterized in that, S6 specifically includes: S601: Calculate the total water supply based on the water supply per unit area required by each vegetation zone: ； in, A Indicates the total water supply. Indicates the first k The water supply required per unit area for each vegetated region S k Indicates the first k The area of ​​each vegetation zone; S602: Determine the initial opening parameters of the reservoir gate based on the total water supply: ； in, β This indicates the initial opening parameter of the reservoir gate. b Indicates the opening coefficient; S603: Control the gate of the water storage tank according to the initial opening parameters of the gate; S604: During the irrigation process, based on the PID controller, the opening parameters of the reservoir gate are adjusted in real time to achieve the total water supply target; S605: Determine the initial valve opening parameters based on the required water supply per unit area and the number of irrigation holes in each vegetation zone: ； in, λ k Indicates the first k The initial opening parameters of the valves within each vegetated area. c k Indicates the first k The number of irrigation holes contained in a vegetated area; S606: Control the valves in each vegetation zone according to the initial opening parameters of the valves; S607: During the irrigation process, based on the PID controller, the valve opening parameters of each vegetation area are adjusted in real time to achieve the required water supply per unit area for each vegetation area.

9. A smart irrigation control system for vegetation in front of a seawall, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the intelligent irrigation control method for vegetation in front of the seawall as described in any one of claims 1 to 8.

Images