A wheat intelligent water-saving irrigation method

Abstract

The application provides a wheat intelligent water-saving irrigation method, comprising: acquiring visible light images and hyperspectral images of each growth stage of wheat; selecting an image of an optional wave band from the hyperspectral images as an enhanced image; fusing the enhanced image and the visible light image to form a fused image; labeling the water requirement of the wheat based on the fused image of each growth stage of the wheat to form a training sample; inputting the training sample into a neural network for training to obtain a wheat water requirement prediction model; and calculating the water requirement of the wheat in a target wheat field using the wheat water requirement prediction model and irrigating the target wheat field according to the water requirement. The application can more accurately analyze and predict the water requirement of the wheat by acquiring image data of each growth stage of the wheat, thereby not only ensuring that the irrigation amount meets the actual demand and effectively reducing the waste of water resources, but also helping to improve the growth environment of crops and thereby improving the yield and quality of the wheat.

Description

Technical Field

[0001] This invention belongs to the field of smart irrigation technology, and more specifically, relates to a smart water-saving irrigation method for wheat. Background Technology

[0002] With the intensification of global climate change and the increasing scarcity of water resources, agricultural irrigation faces severe challenges. As one of the world's most important food crops, wheat production is not only directly related to food security but also closely linked to farmers' livelihoods.

[0003] In recent years, with the rapid development of smart agriculture technology, irrigation using advanced sensors, the Internet of Things, and other technologies has gradually become a research hotspot. For example, soil moisture sensors can monitor soil moisture in real time, allowing workers to adjust irrigation timing and volume based on practical experience. However, this method is susceptible to the influence of workers' subjective judgment, easily leading to water waste and low irrigation efficiency, negatively impacting wheat growth and yield. Summary of the Invention

[0004] To address the aforementioned problems, the present invention aims to provide a smart water-saving irrigation method for wheat.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A smart water-saving irrigation method for wheat includes:

[0007] Step 1: Acquire visible light and hyperspectral images of wheat at various growth stages;

[0008] Step 2: Select any band from the hyperspectral images as the enhanced image;

[0009] Step 3: Fuse the enhanced image and the visible light image to form a fused image;

[0010] Step 4: Label the water requirement of wheat based on the fused images of wheat at various growth stages to form training samples;

[0011] Step 5: Input the training samples into the neural network for training to obtain the wheat water requirement prediction model; Step 6: Evaluate the accuracy of the wheat water requirement prediction model. If the accuracy is not within the preset range, select an image of a different band from the hyperspectral image as the enhancement image and return to Step 3.

[0012] Step 7: When the accuracy is within the preset range, use the wheat water requirement prediction model to calculate the water requirement of wheat in the target wheat field, and irrigate the target wheat field according to the water requirement.

[0013] Preferably, step 3: fusing the enhanced image and the visible light image to form a fused image includes:

[0014] Step 3.1: Use non-subsampled contourlet transform to decompose the enhanced image and the visible light image respectively to obtain the enhanced image transform coefficients and the visible light image transform coefficients;

[0015] Step 3.2: Extract the feature map of the enhanced image;

[0016] Step 3.3: Normalize the feature map of the enhanced image to obtain a normalized feature map;

[0017] Step 3.4: Calculate the weight coefficients based on the normalized feature map;

[0018] Step 3.5: Use the weighting coefficients to weight the enhanced image transformation coefficients and the visible light image transformation coefficients to form a fusion coefficient;

[0019] Step 3.6: Perform non-subsampled contour wave inverse transform on the fusion coefficients to obtain the fused image.

[0020] Preferably, step 3.2: extracting the feature map of the enhanced image includes:

[0021] Step 3.2.1: Construct a window with a point on the enhanced image as the center point;

[0022] Step 3.2.2: Process the pixels within the window using a nonlinear filter, and continuously slide the window to form a feature map of the enhanced image; wherein, the expression formula for the nonlinear filter is:

[0023]

[0024] Where g(i,j) represents the pixel value of the feature map at point (i,j), S(i,j) is the set of pixels within a window centered at point (i,j), f(k,l) represents the pixel value of the enhanced image at point (i,j), w(i,j,k,l) ​​represents the filter weights, and σ d σ represents the standard deviation of the distance. r This represents the standard deviation of pixels.

[0025] Preferably, step 3.3: normalizing the feature map of the enhanced image to obtain a normalized feature map includes:

[0026] Obtain the maximum and minimum values ​​of all pixels in the enhanced image feature map;

[0027] The feature map is normalized based on the maximum and minimum values ​​of the pixels to obtain a normalized feature map; the normalization formula is as follows:

[0028]

[0029] In the formula, S norm (x,y) represents the normalized feature map, S BF (x,y) represents the original feature map, S max S represents the maximum value among all pixels in the feature map. min This represents the minimum value among all pixels in the feature map.

[0030] Preferably, step 3.4: calculating the weight coefficients based on the normalized feature map includes:

[0031] Step 3.4.1: Calculate the information content of each block based on the normalized feature map and the pixel distribution on the visible light image; the formula for calculating the information content is:

[0032]

[0033] Among them, E IR (x,y) represents the amount of information in the visible light image of the block centered at (x,y), E ID (x,y) represents the amount of information in the normalized feature map of the block centered at (x,y), ID(x+i,y+j) represents the pixel value at point (x+i,y+j) in the normalized feature map, IR(x+i,y+j) represents the pixel value at point (x+i,y+j) in the visible light image, and w(i,j) represents the weight matrix.

[0034] Step 3.4.2: Calculate the information weight based on the amount of information in each block;

[0035] Step 3.4.3: Introduce normalized feature maps and obtain weight coefficients based on information content weights.

[0036] Preferably, in step 3.4.2, the information weight is calculated based on the proportion of information in each block; wherein the formula for calculating the information weight is:

[0037]

[0038] Where α1(x,y) represents the first information weight and α2(x,y) represents the second information weight.

[0039] Preferably, step 3.4.3: introducing normalized feature maps to obtain weight coefficients based on information content weights, includes:

[0040] Formula used:

[0041]

[0042] Calculate the weighting coefficients; where w IR (x,y) represents the first weighting coefficient, w ID (x,y) represents the second weighting coefficient, α1′(x,y)=S norm (x,y)+α1(x,y), α1′(x,y) min Let α1′(x,y) represent the minimum value in α1′(x,y). max Let α1′(x,y) represent the maximum value in α2′(x,y), and α2′(x,y) = 1 - S norm (x,y)+α2(x,y), α2′(x,y) min Let α2′(x,y) represent the minimum value in α2′(x,y). max This represents the maximum value in α2′(x,y).

[0043] Preferably, in step 3.5, the visible light image transformation coefficient and the enhanced image transformation coefficient are weighted and summed based on the first weighting coefficient and the second weighting coefficient respectively to form a fusion coefficient; wherein the weighted summation formula is:

[0044] IF coef (x,y)=w IR (x,y)IR coef (x,y)+w ID (x,y)ID coef (x,y)

[0045] Among them, IR coef (x,y) represents the transform coefficient corresponding to the point (x,y) in the visible light image, ID coef (x,y) represents the transform coefficient corresponding to the point (x,y) on the enhanced image, IF coef (x,y) represents the fusion coefficient.

[0046] 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 above-described intelligent water-saving irrigation method for wheat.

[0047] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the above-described intelligent water-saving irrigation method for wheat.

[0048] The beneficial effects of the intelligent water-saving irrigation method for wheat provided by this invention are as follows: Compared with the prior art, this invention can more accurately analyze and predict the water demand of wheat by acquiring image data of wheat at various growth stages. This not only ensures that the irrigation amount meets the actual needs and effectively reduces the waste of water resources, but also helps to improve the crop growth environment, thereby improving the yield and quality of wheat. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention, 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.

[0050] Figure 1 A flowchart of a smart water-saving irrigation method for wheat provided in an embodiment of the present invention;

[0051] Figure 2 The relationship between wheat hyperspectral data at the jointing stage and wheat leaf water content is provided in this embodiment of the invention; the horizontal axis represents wavelength and the vertical axis represents correlation coefficient.

[0052] Figure 3 The relationship between hyperspectral data and wheat leaf water content during the grain-filling stage is provided in this embodiment of the invention. Detailed Implementation

[0053] To make the technical problems, solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0054] Please see Figure 1 A smart water-saving irrigation method for wheat, comprising:

[0055] Step 1: Acquire visible light and hyperspectral images of wheat at various growth stages;

[0056] Step 2: Select any band from the hyperspectral images as the enhanced image;

[0057] Step 3: Fuse the enhanced image and the visible light image to form a fused image;

[0058] Furthermore, step 3 includes:

[0059] Step 3.1: Use non-subsampled contourlet transform to decompose the enhanced image and the visible light image respectively to obtain the enhanced image transform coefficients and the visible light image transform coefficients;

[0060] Step 3.2: Extract the feature map of the enhanced image;

[0061] Step 3.2 includes:

[0062] Step 3.2.1: Construct a window with a point on the enhanced image as the center point;

[0063] Step 3.2.2: Process the pixels within the window using a nonlinear filter, and continuously slide the window to form a feature map of the enhanced image; wherein, the expression formula for the nonlinear filter is:

[0064]

[0065] Where g(i,j) represents the pixel value of the feature map at point (i,j), S(i,j) is the set of pixels within a window centered at point (i,j), f(k,l) represents the pixel value of the enhanced image at point (i,j), w(i,j,k,l) ​​represents the filter weights, and σ d σ represents the standard deviation of the distance. r This represents the standard deviation of pixels.

[0066] The nonlinear filter of this invention takes into account the spatial location and pixel value of pixels when processing images. Based on this, the enhanced image processing can remove noise or background interference, and enhance the biological characteristics of wheat, such as plant texture, shape, and color variations. This has a positive effect on analyzing the health status and maturity of plants.

[0067] Step 3.3: Normalize the feature map of the enhanced image to obtain a normalized feature map;

[0068] In step 3.3, the present invention can complete the normalization process based on the max-min normalization algorithm, and the specific process is as follows:

[0069] Step 3.3.1: Obtain the maximum and minimum values ​​of all pixels in the enhanced image feature map;

[0070] Step 3.3.2: Normalize the feature map based on the maximum and minimum values ​​of the pixels to obtain a normalized feature map; the normalization formula is as follows:

[0071]

[0072] In the formula, S norm (x,y) represents the normalized feature map, SBF (x,y) represents the original feature map, S max S represents the maximum value among all pixels in the feature map. min This represents the minimum value among all pixels in the feature map.

[0073] Step 3.4: Calculate the weight coefficients based on the normalized feature map;

[0074] Furthermore, step 3.4 includes:

[0075] Step 3.4.1: Calculate the information content of each block based on the normalized feature map and the pixel distribution on the visible light image; the formula for calculating the information content is:

[0076]

[0077] Among them, E IR (x,y) represents the amount of information in the visible light image of the block centered at (x,y), E ID (x,y) represents the amount of information in the normalized feature map of the block centered at (x,y), ID(x+i,y+j) represents the pixel value at point (x+i,y+j) in the normalized feature map, IR(x+i,y+j) represents the pixel value at point (x+i,y+j) in the visible light image, and w(i,j) represents the weight matrix.

[0078] Step 3.4.2: Calculate the information weight based on the amount of information in each block;

[0079] In step 3.4.2, the present invention calculates the information weight based on the proportion of information in each block; wherein, the formula for calculating the information weight is:

[0080]

[0081] Where α1(x,y) represents the first information weight and α2(x,y) represents the second information weight.

[0082] Step 3.4.3: Introduce normalized feature maps and obtain weight coefficients based on information content weights. The formula for calculating the weight coefficients is as follows:

[0083]

[0084] Among them, w IR (x,y) represents the first weighting coefficient, w ID (x,y) represents the second weighting coefficient, α1′(x,y)=S norm (x,y)+α1(x,y), α1′(x,y) minLet α1′(x,y) represent the minimum value in α1′(x,y). max Let α1′(x,y) represent the maximum value in α2′(x,y), and α2′(x,y) = 1 - S norm (x,y)+α2(x,y), α2′(x,y) min Let α2′(x,y) represent the minimum value in α2′(x,y). max This represents the maximum value in α2′(x,y).

[0085] Step 3.5: Use the weighting coefficients to weight the enhanced image transformation coefficients and the visible light image transformation coefficients to form a fusion coefficient;

[0086] In step 3.5, the visible light image transformation coefficient and the enhanced image transformation coefficient are weighted and summed based on the first weighting coefficient and the second weighting coefficient, respectively, to form a fusion coefficient; wherein the weighted summation formula is:

[0087] IF coef (x,y)=w IR (x,y)IR coef (x,y)+w ID (x,y)ID coef (x,y)

[0088] Among them, IR coef (x,y) represents the transform coefficient corresponding to the point (x,y) in the visible light image, ID coef (x,y) represents the transform coefficient corresponding to the point (x,y) on the enhanced image, IF coef (x,y) represents the fusion coefficient.

[0089] Step 3.6: Perform non-subsampled contour wave inverse transform on the fusion coefficients to obtain the fused image.

[0090] Since each block in the image contains different amounts of information, this invention needs to minimize the influence of background information as much as possible, and use areas with high information content as the areas where the hyperspectral image needs to be fused. This can improve the feature representation ability of the subsequent fused image in the neural network.

[0091] Step 4: Label the water requirement of wheat based on the fused images of wheat at various growth stages to form training samples;

[0092] In practical applications, the humidity values ​​required for each stage of wheat growth can be labeled.

[0093] Step 5: Input the training samples into the neural network for training to obtain the wheat water requirement prediction model;

[0094] In step 5, the training samples need to be divided into training set, validation set and test set. A common ratio is 70% for training, 15% for validation and 15% for testing. Then, they are input into the CNN model for training, and the training effect is evaluated in real time using a loss function. When the value of the loss function falls within the preset range, the training is complete and the wheat water requirement prediction model is obtained.

[0095] Step 6: Evaluate the accuracy of the wheat water requirement prediction model. If the accuracy is not within the preset range, select another band of image from the hyperspectral image as the enhancement image and return to step 3.

[0096] Step 7: When the accuracy is within the preset range, use the wheat water requirement prediction model to calculate the water requirement of wheat in the target wheat field, and irrigate the target wheat field according to the water requirement.

[0097] like Figure 2-3 As shown, the correlation between wheat leaf water content and the hyperspectral acquisition band is parabolic. Therefore, by fusing the hyperspectral image as an enhanced image into the visible light image, the water supply required by wheat at each growth stage can be reflected.

[0098] Visible light images and hyperspectral images provide different types of information. Visible light images can reflect the growth status and health of wheat, while hyperspectral images can provide detailed information about the water content of wheat leaves. Fusing these two types of images can yield more comprehensive features, allowing neural networks to learn features at different levels and scales. This enables them to better capture the subtle differences in the water status changes of wheat at various growth stages. Real-time analysis of this type of fused data can provide timely information support for agricultural management, helping farmers make scientific and rational irrigation decisions and optimize water resource utilization.

[0099] 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. When the computer program is executed by the processor, it implements the steps in the above-described intelligent water-saving irrigation method for wheat and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0100] 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 intelligent water-saving irrigation method for wheat and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart water-saving irrigation method for wheat, characterized in that, include: Step 1: Acquire visible light and hyperspectral images of wheat at various growth stages; Step 2: Select any band from the hyperspectral images as the enhanced image; Step 3: Fuse the enhanced image and the visible light image to form a fused image; Step 3: Fusing the enhanced image and the visible light image to form a fused image includes: Step 3.1: Use non-subsampled contourlet transform to decompose the enhanced image and the visible light image respectively to obtain the enhanced image transform coefficients and the visible light image transform coefficients; Step 3.2: Extract the feature map of the enhanced image; Step 3.3: Normalize the feature map of the enhanced image to obtain a normalized feature map; Step 3.4: Calculate the weight coefficients based on the normalized feature map; Step 3.4: Calculate the weight coefficients based on the normalized feature map, including: Step 3.4.1: Calculate the information content of each block based on the normalized feature map and the pixel distribution on the visible light image; the formula for calculating the information content is: Among them, E IR (x,y) represents the amount of information in the visible light image of the block centered at (x,y), E ID (x,y) represents the amount of information in the normalized feature map of the block centered at (x,y), ID(x+i,y+j) represents the pixel value at point (x+i,y+j) in the normalized feature map, IR(x+i,y+j) represents the pixel value at point (x+i,y+j) in the visible light image, and w(i,j) represents the weight matrix. Step 3.4.2: Calculate the information weight based on the amount of information in each block; In step 3.4.2, the information weight is calculated based on the proportion of information in each block; the formula for calculating the information weight is as follows: Where α1(x,y) represents the first information weight and α2(x,y) represents the second information weight; Step 3.4.3: Introduce normalized feature maps and obtain weight coefficients based on information content weights; Step 3.4.3: Introducing normalized feature maps and obtaining weight coefficients based on information content weights, including: Formula used: Calculate the weighting coefficients; where w IR (x,y) represents the first weighting coefficient, w ID (x,y) represents the second weighting coefficient, α1′(x,y)=S norm (x,y)+α1(x,y), α1′(x,y) min Let α1′(x,y) represent the minimum value in α1′(x,y). max Let α1′(x,y) represent the maximum value in α2′(x,y), and α2′(x,y) = 1 - S norm (x,y)+α2(x,y), α2′(x,y) min Let α2′(x,y) represent the minimum value in α2′(x,y). max This represents the maximum value in α2′(x,y); Step 3.5: Use the weighting coefficients to weight the enhanced image transformation coefficients and the visible light image transformation coefficients to form a fusion coefficient; In step 3.5, the visible light image transformation coefficient and the enhanced image transformation coefficient are weighted and summed based on the first weighting coefficient and the second weighting coefficient, respectively, to form a fusion coefficient; wherein the weighted summation formula is: IF coef (x,y)=w IR (x,y)IR coef (x,y)+w ID (x,y)ID coef (x,y) Among them, IR coef (x,y) represents the transform coefficient corresponding to the point (x,y) in the visible light image, ID coef (x,y) represents the transform coefficient corresponding to the point (x,y) on the enhanced image, IF coef (x,y) represents the fusion coefficient; Step 3.6: Perform non-subsampled contourlet inverse transform on the fusion coefficients to obtain the fused image; Step 4: Label the water requirement of wheat based on the fused images of wheat at various growth stages to form training samples; Step 5: Input the training samples into the neural network for training to obtain the wheat water requirement prediction model; Step 6: Evaluate the accuracy of the wheat water requirement prediction model. If the accuracy is not within the preset range, select another band of image from the hyperspectral image as the enhancement image and return to step 3. Step 7: When the accuracy is within the preset range, use the wheat water requirement prediction model to calculate the water requirement of wheat in the target wheat field, and irrigate the target wheat field according to the water requirement.

2. The intelligent water-saving irrigation method for wheat as described in claim 1, characterized in that, Step 3.2: Extracting the feature map of the enhanced image, including: Step 3.2.1: Construct a window with a point on the enhanced image as the center point; Step 3.2.2: Process the pixels within the window using a nonlinear filter, and continuously slide the window to form a feature map of the enhanced image; wherein, the expression formula for the nonlinear filter is: Where g(i,j) represents the pixel value of the feature map at point (i,j), S(i,j) is the set of pixels within a window centered at point (i,j), f(k,l) represents the pixel value of the enhanced image at point (i,j), w(i,j,k,l) ​​represents the filter weights, and σ d σ represents the standard deviation of the distance. r This represents the standard deviation of pixels.

3. The intelligent water-saving irrigation method for wheat as described in claim 2, characterized in that, Step 3.3: Normalize the feature map of the enhanced image to obtain a normalized feature map, including: Obtain the maximum and minimum values ​​of all pixels in the enhanced image feature map; The feature map is normalized based on the maximum and minimum values ​​of the pixels to obtain a normalized feature map; the normalization formula is as follows: In the formula, S norm (x,y) represents the normalized feature map, S BF (x,y) represents the original feature map, S max S represents the maximum value among all pixels in the feature map. min This represents the minimum value among all pixels in the feature map.

4. An electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, characterized in that, When the computer program is executed by the processor, it implements the steps of the smart water-saving irrigation method for wheat as described in any one of claims 1-3.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent water-saving irrigation method for wheat as described in any one of claims 1-3.

Images