A farmland plot construction and acreage measurement method based on remote sensing images

Abstract

The present application relates to a kind of farmland plot construction and measure acre method based on remote sensing image, the method comprises:U1.Multiple unmanned aerial vehicles enter the farmland area to be measured, based on the image data information of real-time acquisition farmland area of airborne camera, and the pre-processing of image is carried out, obtain the pre-processed multiple groups of image data information of farmland area;U2.Based on the pre-processed multiple groups of image data information of farmland area, using improved multi-focal point image fusion algorithm is fused to the image of farmland area, obtain the image data information of fused farmland area.The present application not only can quickly, accurately collect farmland image data, realize the efficient coverage and data update of large area land, but also can accurately identify farmland area, be applicable to irregular block and complex terrain, ensure the accuracy and reliability of area calculation result.

Description

Technical Field

[0001] The present invention relates to the technical field of farmland acreage measurement, and in particular to a farmland plot construction and acreage measurement method based on remote sensing images. Background Art

[0002] With the continuous development of drone technology, it is affecting the continuous progress of agricultural technology. Among them, the continuous updating and improvement of technologies such as drone spraying pesticides on farmland and drone sowing are changing the rapid development of agriculture.

[0003] Agricultural success often depends on factors farmers have little or no control over: weather and soil conditions, temperature, precipitation, and more. Key to efficiency lies in their ability to adapt, which is largely influenced by accurate, near-real-time information. For example, when measuring acres of farmland, traditional ground-based surveying tools require manual fieldwork, which is time-consuming, labor-intensive, and inefficient, especially for large farmland areas. Furthermore, measurement accuracy is difficult to guarantee due to factors such as topography. Furthermore, the time required prevents frequent updates of map information. Summary of the Invention

[0004] In view of the above problems, the present invention provides a method for constructing and measuring farmland plots based on remote sensing images, which can not only quickly and accurately collect farmland image data, achieve efficient coverage and data update of large areas of land, but also accurately identify farmland areas. It is suitable for irregular plots and complex terrain, ensuring the accuracy and reliability of area calculation results.

[0005] In order to achieve the above-mentioned and other related purposes, the present invention provides the following technical solutions:

[0006] A method for constructing and measuring farmland plots based on remote sensing images, the method comprising:

[0007] U1. Multiple drones enter the farmland area to be measured, acquire real-time image data of the farmland area using onboard cameras, and perform image preprocessing to obtain multiple sets of preprocessed image data of the farmland area.

[0008] U2. Based on the preprocessed image data information of the plurality of groups of farmland areas, the images of the farmland area are fused using an improved multi-focus image fusion algorithm to obtain fused image data information of the farmland area;

[0009] U3. Based on the fused image data information of the farmland area, a deep learning convolutional neural network model integrating the OSTU algorithm is constructed to identify the image of the farmland area boundary points and construct a pixel matrix of the farmland area boundary to obtain pixel matrix data information of the farmland area boundary;

[0010] U4. Based on the pixel point matrix data information of the farmland area boundary, the pixel points of the farmland area boundary are optimized by using the whale optimization algorithm based on the golden sine factor, and the pixel point matrix data information of the optimized farmland area boundary is obtained;

[0011] U5. Based on the pixel point matrix data information of the optimized farmland area boundary, the area of the farmland area is calculated by using the multivariate integral algorithm of farmland area boundary fitting, and the area data information of the farmland area is obtained.

[0012] Further, in step U1, the preprocessing of the image includes image denoising and artifact removal, image correction, and image enhancement.

[0013] Further, the image enhancement is to enhance the image of the farmland area by using the improved Retinex algorithm, which includes:

[0014] U11. Based on the image data information of the farmland area, the pixel point matrix of the image of the farmland area is constructed, and the image pixel point matrix data information of the farmland area is obtained;

[0015] U12. Based on the image pixel point matrix data information of the farmland area, the pixel point enhancement function G of the image is constructed,

[0016]

[0017] Wherein, x is the image pixel point matrix data information of the farmland area, α is the Gaussian surrounding space constant, * is the convolution operation, e is the natural constant;

[0018] U13. Based on the image pixel point enhancement function G, the image of the farmland area is enhanced, and the enhanced image data information of the farmland area is obtained.

[0019] Further, in step U2, the image of the farmland area is fused by using the improved multifocal image fusion algorithm, which includes:

[0020] U21. Based on the preprocessed multi-group farmland area image data information, the Laplacian transform function F of each group of farmland area image is constructed,

[0021]

[0022] Wherein, y i is the preprocessed i-th group of farmland area image data information, β i is the gradient factor of the preprocessed i-th group of farmland area image, which characterizes the definition of each group of farmland area image, and the definition data information of each group of farmland area image is obtained;

[0023] U22. Based on the image clarity data information of each group of farmland areas, a feature extraction function H of the farmland area is established.

[0024]

[0025] Among them, r i is the image clarity data information of the i-th group of farmland areas, λ1, λ2 and λ3 are the feature extraction factors of the farmland area images, and feature extraction is performed on the farmland area images to obtain the feature matrix data information of each group of farmland areas;

[0026] U23. Based on the characteristic matrix data information of each group of farmland areas, construct a fusion function R of the farmland area image,

[0027]

[0028] Among them, z i is the characteristic matrix data information of the i-th group of farmland areas, ω i is the weight coefficient of the feature matrix of the i-th group of farmland areas, n is the sample capacity, and the images of the farmland areas are fused to obtain the image data information of the fused farmland areas.

[0029] Furthermore, the constraints of the feature extraction factors λ1, λ2 and λ3 of the farmland area image are as follows:

[0030]

[0031] The weight coefficient ω of the characteristic matrix of the i-th group of farmland areas i for,

[0032]

[0033] Among them, z i is the characteristic matrix data information of the i-th group of farmland areas.

[0034] Furthermore, in step U3, the construction of a deep learning convolutional neural network model integrating the OSTU algorithm to identify images of farmland area boundary points includes:

[0035] U31. Based on the image data information of the fused farmland area, a grayscale histogram of the farmland area is constructed, and a cumulative distribution function Q of the grayscale histogram of the farmland area is established.

[0036]

[0037] Wherein, a is the image data information of the fused farmland area, η1, η2 and η3 are the gain constant parameters of the farmland area image, and the boundary threshold of the farmland area image is calculated to obtain the data information of the boundary threshold of the farmland area image;

[0038] U32. Input the data information of the boundary threshold of the farmland area image into the deep learning convolutional neural network model for training and learning, and determine the kernel function P of the convolutional neural network model.

[0039]

[0040] Among them, b is the data information of the boundary threshold of the farmland area image, ρ1, ρ2 and ρ3 are the boundary recognition factors of the farmland area, and the trained deep learning convolutional neural network model is obtained;

[0041] U33. Based on the trained deep learning convolutional neural network model, the fused image data information of the farmland area is input, the image of the boundary points of the farmland area is identified, and the pixel point matrix of the boundary of the farmland area is constructed to obtain the pixel point matrix data information of the boundary of the farmland area.

[0042] Furthermore, in step U4, the optimization of pixel points at the boundary of the farmland area using the whale optimization algorithm based on the golden sine factor includes:

[0043] U41. Based on the pixel matrix data information of the farmland area boundary, the whale population is initialized, the maximum number of iterations L is determined, and the data information of the initialized whale population is obtained;

[0044] U42. Based on the data information of the initialized whale population, establish the fitness function S of the individual whale population,

[0045]

[0046] Among them, c is the data information of the initialized whale population, σ1, σ2 and σ3 are the fitness determining factors of the individual whale population, and the fitness values ​​of the individual whale population are calculated to obtain the fitness value data information of the individual whale population;

[0047] U43. Based on the fitness value data information of the individual whale population, establish the objective function V,

[0048]

[0049] Among them, h is the fitness value data information of the whale population individuals, δ1 and δ2 are golden sine factors, and the pixel points at the boundary of the farmland area are optimized to obtain the optimized pixel point matrix data information of the boundary of the farmland area.

[0050] Further, in step U5, the multi-element integral algorithm using the farmland region boundary fitting is used to calculate the area of the farmland region, including:

[0051] U51. Based on the pixel matrix data information of the optimized farmland region boundary, a multi-point fitting function O of the farmland region is established,

[0052]

[0053] Wherein g is the pixel matrix data information of the optimized farmland region boundary, θ1, θ2, θ3, θ4 and θ5 are fitting constant parameters, the farmland region boundary pixel points are fitted, and the curve data information of the upper half and the lower half of the fitted farmland region boundary is obtained;

[0054] U52. Based on the curve data information of the upper half and the lower half of the fitted farmland region boundary, a multi-element integral function M of the farmland region is established,

[0055]

[0056] Wherein f1 is the curve of the upper half of the fitted farmland region boundary, f2 is the curve of the lower half of the fitted farmland region boundary, γ1, γ2 and γ3 are integral constant factors of the farmland region area;

[0057] U53. Based on the multi-element integral function M of the farmland region, the area of the farmland region is calculated to obtain the area data information of the farmland region.

[0058] In order to achieve the above-mentioned purpose and other related purposes, the application further provides a farmland plot construction and mu measurement system based on remote sensing images, comprising a computer device programmed or configured to execute the steps of any one of the farmland plot construction and mu measurement methods based on remote sensing images.

[0059] In order to achieve the above-mentioned purpose and other related purposes, the application further provides a computer readable storage medium, which stores a computer program programmed or configured to execute any one of the farmland plot construction and mu measurement methods based on remote sensing images.

[0060] The application has the following positive effects:

[0061] 1. The present invention fuses images of farmland areas by adopting an improved multi-focus image fusion algorithm, and combines it with a deep learning convolutional neural network model that integrates the OSTU algorithm to identify images of farmland area boundary points and construct a pixel point matrix of the farmland area boundary. This method can not only quickly and accurately collect farmland image data, achieve efficient coverage and data update of large areas of land, but also accurately identify farmland areas. It is suitable for irregular plots and complex terrains, ensuring the accuracy and reliability of area calculation results.

[0062] 2. This invention optimizes the pixel points at the boundaries of farmland areas by using a whale optimization algorithm based on the golden sine factor, and estimates the area of ​​farmland areas by combining it with a multivariate integration algorithm for fitting the boundaries of farmland areas. Compared with traditional manual measurement, this method not only significantly reduces the cost of farmland management and improves the efficiency of data acquisition and analysis, but also does not require human intervention throughout the process, reducing labor intensity and improving measurement efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the method flow of the present invention;

[0064] Figure 2 Schematic diagram of the improved Retinex algorithm of the present invention;

[0065] Figure 3 Schematic diagram of the process of the improved multi-focus image fusion algorithm of the present invention;

[0066] Figure 4 This is a flow chart of the construction of a deep learning convolutional neural network model integrating the OSTU algorithm of the present invention;

[0067] Figure 5 Schematic diagram of the flow of the whale optimization algorithm based on the golden sine factor of the present invention;

[0068] Figure 6 Schematic diagram of the flow of the multivariate integration algorithm for farmland area boundary fitting of the present invention;

[0069] Figure 7 Schematic diagram of the farmland area image of the present invention. DETAILED DESCRIPTION

[0070] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0071] Example 1: Figure 1 or Figure 7 As shown, a method for constructing and measuring farmland plots based on remote sensing images includes:

[0072] U1. Multiple drones enter the farmland area to be measured, acquire real-time image data of the farmland area using onboard cameras, and perform image preprocessing to obtain multiple sets of preprocessed image data of the farmland area.

[0073] U2. Based on the preprocessed image data information of the plurality of groups of farmland areas, the images of the farmland area are fused using an improved multi-focus image fusion algorithm to obtain fused image data information of the farmland area;

[0074] U3. Based on the fused image data information of the farmland area, a deep learning convolutional neural network model integrating the OSTU algorithm is constructed to identify the image of the farmland area boundary points and construct a pixel matrix of the farmland area boundary to obtain pixel matrix data information of the farmland area boundary;

[0075] U4. Based on the pixel matrix data information of the farmland area boundary, the pixel points of the farmland area boundary are optimized using the whale optimization algorithm based on the golden sine factor to obtain the optimized pixel matrix data information of the farmland area boundary;

[0076] U5. Based on the optimized pixel matrix data information of the farmland area boundary, the area of ​​the farmland area is estimated using a multivariate integration algorithm for farmland area boundary fitting to obtain area data information of the farmland area.

[0077] Furthermore, in step U1, the image preprocessing includes image denoising and artifact removal, image correction and image enhancement.

[0078] In this embodiment, if Figure 2 As shown, the image enhancement is to enhance the image of the farmland area using the improved Retinex algorithm, including:

[0079] U11. Based on the image data information of the farmland area, construct a pixel matrix of the image of the farmland area to obtain the image pixel matrix data information of the farmland area;

[0080] U12. Based on the image pixel matrix data information of the farmland area, construct an image pixel enhancement function G,

[0081]

[0082] Where x is the image pixel matrix data information of the farmland area, α is the Gaussian surrounding space constant, * is the convolution operation, and e is a natural constant;

[0083] U13. Based on the image pixel enhancement function G, the image of the farmland area is enhanced to obtain enhanced image data information of the farmland area.

[0084] In this embodiment, if Figure 3 As shown, in step U2, the use of the improved multi-focus image fusion algorithm to fuse images of the farmland area includes:

[0085] U21. Based on the pre-processed image data information of the plurality of groups of farmland areas, constructing a Laplace transform function F of each group of farmland area images,

[0086]

[0087] Among them, y i is the image data information of the i-th group of farmland area after preprocessing, β i is the gradient factor of the pre-processed image of the i-th group of farmland areas, and the clarity of the image of each group of farmland areas is characterized to obtain the clarity data information of the image of each group of farmland areas;

[0088] U22. Based on the image clarity data information of each group of farmland areas, a feature extraction function H of the farmland area is established.

[0089]

[0090] Among them, r i is the image clarity data information of the i-th group of farmland areas, λ1, λ2 and λ3 are the feature extraction factors of the farmland area images, and feature extraction is performed on the farmland area images to obtain the feature matrix data information of each group of farmland areas;

[0091] U23. Based on the characteristic matrix data information of each group of farmland areas, construct a fusion function R of the farmland area image,

[0092]

[0093] Among them, z i is the characteristic matrix data information of the i-th group of farmland areas, ω i is the weight coefficient of the feature matrix of the i-th group of farmland areas, n is the sample capacity, and the images of the farmland areas are fused to obtain the image data information of the fused farmland areas.

[0094] In this embodiment, the constraints of the feature extraction factors λ1, λ2 and λ3 of the farmland area image are:

[0095]

[0096] The weight coefficient ω of the characteristic matrix of the i-th group of farmland areas i for,

[0097]

[0098] Among them, z i is the characteristic matrix data information of the i-th group of farmland areas.

[0099] Example 2: Based on the farmland plot construction and acreage measurement method based on remote sensing images in Example 1, the present invention is further illustrated and described below.

[0100] like Figure 1 or Figure 7 As shown, a method for constructing and measuring farmland plots based on remote sensing images includes:

[0101] U1. Multiple drones enter the farmland area to be measured, acquire real-time image data of the farmland area using onboard cameras, and perform image preprocessing to obtain multiple sets of preprocessed image data of the farmland area.

[0102] U2. Based on the preprocessed image data information of the plurality of groups of farmland areas, the images of the farmland area are fused using an improved multi-focus image fusion algorithm to obtain fused image data information of the farmland area;

[0103] U3. Based on the fused image data information of the farmland area, a deep learning convolutional neural network model integrating the OSTU algorithm is constructed to identify the image of the farmland area boundary points and construct a pixel matrix of the farmland area boundary to obtain pixel matrix data information of the farmland area boundary;

[0104] U4. Based on the pixel matrix data information of the farmland area boundary, the pixel points of the farmland area boundary are optimized using the whale optimization algorithm based on the golden sine factor to obtain the optimized pixel matrix data information of the farmland area boundary;

[0105] U5. Based on the optimized pixel matrix data information of the farmland area boundary, the area of ​​the farmland area is estimated using a multivariate integration algorithm for farmland area boundary fitting to obtain area data information of the farmland area.

[0106] In this embodiment, if Figure 4 As shown, in step U3, the construction of a deep learning convolutional neural network model integrating the OSTU algorithm to identify images of farmland area boundary points includes:

[0107] U31. Based on the image data information of the fused farmland area, a grayscale histogram of the farmland area is constructed, and a cumulative distribution function Q of the grayscale histogram of the farmland area is established.

[0108]

[0109] Wherein, a is the image data information of the fused farmland area, η1, η2 and η3 are the gain constant parameters of the farmland area image, and the boundary threshold of the farmland area image is calculated to obtain the data information of the boundary threshold of the farmland area image;

[0110] U32. Input the data information of the boundary threshold of the farmland area image into the deep learning convolutional neural network model for training and learning, and determine the kernel function P of the convolutional neural network model.

[0111]

[0112] Among them, b is the data information of the boundary threshold of the farmland area image, ρ1, ρ2 and ρ3 are the boundary recognition factors of the farmland area, and the trained deep learning convolutional neural network model is obtained;

[0113] U33. Based on the trained deep learning convolutional neural network model, the fused image data information of the farmland area is input, the image of the boundary points of the farmland area is identified, and the pixel point matrix of the boundary of the farmland area is constructed to obtain the pixel point matrix data information of the boundary of the farmland area.

[0114] In this embodiment, if Figure 5 As shown, in step U4, the optimization of pixel points at the boundary of the farmland area using the whale optimization algorithm based on the golden sine factor includes:

[0115] U41. Based on the pixel matrix data information of the farmland area boundary, the whale population is initialized, the maximum number of iterations L is determined, and the data information of the initialized whale population is obtained;

[0116] U42. Based on the data information of the initialized whale population, establish the fitness function S of the individual whale population,

[0117]

[0118] Among them, c is the data information of the initialized whale population, σ1, σ2 and σ3 are the fitness determining factors of the individual whale population, and the fitness values ​​of the individual whale population are calculated to obtain the fitness value data information of the individual whale population;

[0119] U43. Based on the fitness value data information of the individual whale population, establish the objective function V,

[0120]

[0121] wherein h is the fitness value data information of the individual of the whale population, δ1 and δ2 are golden sine factors, the pixel points of the farmland region boundary are optimized to obtain the pixel point matrix data information of the optimized farmland region boundary.

[0122] In the present embodiment, as shown in FIG. 8, in step U5, the multi-element integral algorithm using farmland region boundary fitting is used to calculate the area of the farmland region, which includes: Figure 6

[0123] U51. Based on the pixel point matrix data information of the optimized farmland region boundary, a multi-point fitting function O of the farmland region is established,

[0124]

[0125] wherein g is the pixel point matrix data information of the optimized farmland region boundary, θ1, θ2, θ3, θ4 and θ5 are fitting constant parameters, the farmland region boundary pixel points are fitted to obtain the curve data information of the upper half and the lower half of the fitted farmland region boundary;

[0126] U52. Based on the curve data information of the upper half and the lower half of the fitted farmland region boundary, a multi-element integral function M of the farmland region is established,

[0127]

[0128] wherein f1 is the curve of the upper half of the fitted farmland region boundary, f2 is the curve of the lower half of the fitted farmland region boundary, γ1, γ2 and γ3 are integral constant factors of the area of the farmland region;

[0129] U53. Based on the multi-element integral function M of the farmland region, the area of the farmland region is calculated to obtain the area data information of the farmland region.

[0130] In the present embodiment, the present application provides a farmland plot construction and mu measurement system based on remote sensing images, which comprises a computer device programmed or configured to perform the steps of any one of the farmland plot construction and mu measurement methods based on remote sensing images.

[0131] In the present embodiment, the present application provides a computer readable storage medium having stored thereon a computer program programmed or configured to perform any one of the farmland plot construction and mu measurement methods based on remote sensing images.

[0132] ​Any reference to memory, storage, database or other media used in the embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0133] In summary, the present invention can not only quickly and accurately collect farmland image data, achieve efficient coverage and data update of large areas of land, but also accurately identify farmland areas, is suitable for irregular plots and complex terrain, and ensure the accuracy and reliability of area calculation results.

[0134] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A method for constructing and measuring farmland plots based on remote sensing images, characterized in that: The method comprises: U1. Multiple drones enter the farmland area to be measured, acquire real-time image data of the farmland area using onboard cameras, and perform image preprocessing to obtain multiple sets of preprocessed image data of the farmland area. U2. Based on the preprocessed image data information of the plurality of groups of farmland areas, the images of the farmland area are fused using an improved multi-focus image fusion algorithm to obtain fused image data information of the farmland area; U3. Based on the fused image data information of the farmland area, a deep learning convolutional neural network model integrating the OSTU algorithm is constructed to identify the image of the farmland area boundary points and construct a pixel matrix of the farmland area boundary to obtain pixel matrix data information of the farmland area boundary; U4. Based on the pixel matrix data information of the farmland area boundary, the pixel points of the farmland area boundary are optimized using the whale optimization algorithm based on the golden sine factor to obtain the optimized pixel matrix data information of the farmland area boundary; U5. Based on the pixel matrix data information of the optimized farmland area boundary, the area of ​​the farmland area is estimated using a multivariate integral algorithm for fitting the farmland area boundary to obtain the area data information of the farmland area; In step U2, the use of the improved multi-focus image fusion algorithm to fuse images of the farmland area includes: U21. Based on the pre-processed image data information of the plurality of groups of farmland areas, constructing a Laplace transform function F of each group of farmland area images, ， Among them, y i is the image data information of the i-th group of farmland area after preprocessing, β i is the gradient factor of the pre-processed image of the i-th group of farmland areas, and the clarity of the image of each group of farmland areas is characterized to obtain the clarity data information of the image of each group of farmland areas; U22. Based on the image clarity data information of each group of farmland areas, a feature extraction function H of the farmland area is established. ， Among them, r i is the image clarity data information of the i-th group of farmland areas, λ1, λ2 and λ3 are the feature extraction factors of the farmland area images, and feature extraction is performed on the farmland area images to obtain the feature matrix data information of each group of farmland areas; U23. Based on the characteristic matrix data information of each group of farmland areas, construct a fusion function R of the farmland area image, ， Among them, z i is the characteristic matrix data information of the i-th group of farmland areas, ω i is the weight coefficient of the feature matrix of the i-th group of farmland areas, n is the sample capacity, and the images of the farmland areas are fused to obtain the image data information of the fused farmland areas.

2. The method for constructing and measuring farmland plots based on remote sensing images according to claim 1, characterized in that: In step U1 , the image preprocessing includes image denoising and artifact removal, image correction, and image enhancement.

3. The method for constructing and measuring farmland plots based on remote sensing images according to claim 2, characterized in that: The image enhancement is to enhance the image of the farmland area using an improved Retinex algorithm, including: U11. Based on the image data information of the farmland area, construct a pixel matrix of the image of the farmland area to obtain the image pixel matrix data information of the farmland area; U12. Based on the image pixel matrix data information of the farmland area, construct an image pixel enhancement function G, ， Where x is the image pixel matrix data information of the farmland area, α is the Gaussian surrounding space constant, * is the convolution operation, and e is a natural constant; U13. Based on the image pixel enhancement function G, the image of the farmland area is enhanced to obtain enhanced image data information of the farmland area.

4. The method for constructing and measuring farmland plots based on remote sensing images according to claim 1, characterized in that: The constraints of the feature extraction factors λ1, λ2 and λ3 of the farmland area image are: ， The weight coefficient ω of the characteristic matrix of the i-th group of farmland areas i for, ， Among them, z i is the characteristic matrix data information of the i-th group of farmland areas.

5. The method for constructing and measuring farmland plots based on remote sensing images according to claim 1, characterized in that: In step U3, the construction of a deep learning convolutional neural network model integrating the OSTU algorithm to identify images of farmland area boundary points includes: U31. Based on the image data information of the fused farmland area, a grayscale histogram of the farmland area is constructed, and a cumulative distribution function Q of the grayscale histogram of the farmland area is established. ， Wherein, a is the image data information of the fused farmland area, η1, η2 and η3 are the gain constant parameters of the farmland area image, and the boundary threshold of the farmland area image is calculated to obtain the data information of the boundary threshold of the farmland area image; U32. Input the data information of the boundary threshold of the farmland area image into the deep learning convolutional neural network model for training and learning, and determine the kernel function P of the convolutional neural network model. ， Among them, b is the data information of the boundary threshold of the farmland area image, ρ1, ρ2 and ρ3 are the boundary recognition factors of the farmland area, and the trained deep learning convolutional neural network model is obtained; U33. Based on the trained deep learning convolutional neural network model, the fused image data information of the farmland area is input, the image of the boundary points of the farmland area is identified, and the pixel point matrix of the boundary of the farmland area is constructed to obtain the pixel point matrix data information of the boundary of the farmland area.

6. The method for constructing and measuring farmland plots based on remote sensing images according to claim 1, characterized in that: In step U4, the optimization of pixel points at the boundary of the farmland area using the whale optimization algorithm based on the golden sine factor includes: U41. Based on the pixel matrix data information of the farmland area boundary, the whale population is initialized, the maximum number of iterations L is determined, and the data information of the initialized whale population is obtained; U42. Based on the data information of the initialized whale population, establish the fitness function S of the individual whale population, ， Among them, c is the data information of the initialized whale population, σ1, σ2 and σ3 are the fitness determining factors of the individual whale population, and the fitness values ​​of the individual whale population are calculated to obtain the fitness value data information of the individual whale population; U43. Based on the fitness value data information of the individual whale population, establish the objective function V, ， ， ， Among them, h is the fitness value data information of the whale population individuals, δ1 and δ2 are golden sine factors, and the pixel points at the boundary of the farmland area are optimized to obtain the optimized pixel point matrix data information of the boundary of the farmland area.

7. The method for constructing and measuring farmland plots based on remote sensing images according to claim 1, characterized in that: In step U5, the multivariate integration algorithm for fitting the boundaries of the farmland region is used to estimate the area of ​​the farmland region, including: U51. Based on the pixel matrix data information of the optimized farmland area boundary, a multi-point fitting function O of the farmland area is established. ， Among them, g is the pixel matrix data information of the optimized farmland area boundary, θ1, θ2, θ3, θ4 and θ5 are fitting constant parameters, and the pixel points of the farmland area boundary are fitted to obtain the curve data information of the upper and lower parts of the fitted farmland area boundary; U52. Based on the curve data information of the upper and lower parts of the farmland area boundary after fitting, establish a multivariate integral function M of the farmland area, ， Among them, f1 is the curve of the upper half of the farmland area boundary after fitting, f2 is the curve of the lower half of the farmland area boundary after fitting, γ1, γ2 and γ3 are the integral constant factors of the farmland area; U53. Based on the multivariate integral function M of the farmland area, the area of ​​the farmland area is estimated to obtain area data information of the farmland area.

8. A farmland plot construction and acreage measurement system based on remote sensing images, including computer equipment, characterized in that: The computer device is programmed or configured to execute the steps of the method for constructing and measuring farmland plots based on remote sensing images as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program programmed or configured to execute the method for constructing and measuring farmland plots based on remote sensing images as described in any one of claims 1 to 7.

Images