Mulching film reuse farmland corn yield prediction method, system, equipment and medium

Abstract

The invention discloses a corn yield prediction method, system, equipment and medium for a mulching film reuse farmland, and relates to the technical field of agriculture, and the method comprises the steps: collecting the key growth period images of the current mulching film reuse farmland corn, including a mulching film mulching RGB image, a thermal imaging image of soil temperature, and a multispectral image of soil moisture and vegetation characteristics; extracting a mulching film coverage area and a mulching film damage area according to the mulching film coverage RGB image and the gray-level co-occurrence matrix to obtain a mulching film damage rate; establishing a linear regression model of the soil moisture content and the NDVI by using the multispectral image to estimate the soil moisture content; obtaining soil temperature data according to the soil temperature thermal imaging image; a yield collaborative change prediction model of the mulching film reuse farmland in the key growth period of the corn is constructed, and the estimated yield of the corn in a certain year in the future is obtained; according to the method, accurate prediction of the corn yield of the multi-year mulching film reuse farmland is realized.

Description

Technical Field

[0001] The present invention relates to the field of agricultural technology, and in particular to a method, system, equipment and medium for predicting corn yield in a film-reused farmland. Background Art

[0002] Corn is the main food crop in the dryland areas of the Yellow River Basin in my country, and occupies an important position in regional food production. As the region is located inland, the climate is dry, the rainfall is small, and the evaporation is large, mulching technology is one of the important agronomic measures to solve the above problems. With the increase in the area of ​​mulch-covered farmland, the demand for mulch in my country remains above one million tons per year, especially in the arid areas of the Yellow River Basin, where the mulch coverage rate is higher. However, since the main components of agricultural mulch are stable and difficult to degrade in the soil, and because inappropriate recycling methods will increase the degree of mulch fragmentation and gradually expand the area of ​​residual film pollution in the soil, the problem of "white pollution" in farmland is becoming more and more serious. Under the urgent need for sustainable agricultural development, mulch reuse has become an important means to improve agricultural production efficiency, reduce costs, and reduce non-point source pollution. However, during the repeated use of mulch, it is easy to be damaged, which has a complex impact on the soil moisture and soil temperature of farmland.

[0003] Traditional corn yield prediction methods are mostly based on ground surveys and statistical data, which are not only time-consuming, labor-intensive and costly, but also difficult to fully and timely reflect the actual situation of mulch-film reused farmland. UAV remote sensing technology is increasingly widely used in the agricultural field due to its high efficiency, flexibility and high resolution. By carrying various sensors on drones, multi-source data of large areas of farmland can be quickly obtained, including image data, spectral data, etc. These data provide new possibilities for accurately predicting corn yields in mulch-film reused farmland. Therefore, under the background of the country's vigorous promotion of green and high-quality agricultural development and the booming precision agriculture, there is an urgent need for a technology that can give full play to the advantages of drone remote sensing technology, closely combine the characteristics of mulch-film reused farmland, and realize accurate prediction of corn yields.

[0004] Patent 202311299358.4 discloses an evaluation method for the effects of irrigation amount and mulch type on cotton growth, physiology and yield, including steps such as mulching, moisture change monitoring, and yield determination. This method has certain defects because it does not consider the effects of biodegradable mulch damage on soil moisture and cotton yield. Patent 202210428234.0 discloses a crop yield prediction method and system based on drone multispectral image fusion, including steps such as collecting multispectral images, determining texture features, and determining yield index. This method has certain defects. On the one hand, this method only collects remote sensing effects before maturity and cannot fully reflect the growth status of crops; on the other hand, this method does not consider the effects of soil moisture and temperature changes on yield. Patent 202210913918.X discloses a wheat yield prediction method based on drone multi-source remote sensing data, including steps such as carrying a drone remote sensing platform, obtaining remote sensing images, and establishing a yield prediction model. This method has certain defects. The yield prediction model constructed by this method does not consider soil influencing factors and cannot provide support for agricultural production decisions.

[0005] In short, the existing technology does not consider the impact of soil environmental changes after the film is damaged on crop growth and yield. The constructed yield prediction model is single, and the accuracy of estimating corn yield in farmland with multi-year film reuse is low, and it cannot provide effective water and fertilizer management decisions for subsequent agricultural production. Summary of the invention

[0006] In view of the shortcomings of the prior art that the impact of soil environmental changes after film mulch damage on crop growth and yield is not considered, the constructed yield prediction model is single, and the accuracy of estimating corn yield in farmland with multi-year film mulch reuse is low, the present invention proposes a method, system, equipment and medium for predicting corn yield in farmland with film mulch reuse. By utilizing unmanned aerial vehicle remote sensing images, high-resolution data of large areas of farmland can be quickly acquired, and film mulch damage areas can be accurately identified and quantified. The impact of multiple factors such as film mulch damage, soil moisture and soil temperature on corn growth and yield is comprehensively considered, and a prediction model for the coordinated changes of multiple factors and yield is established, thereby solving the problems existing in the prior art.

[0007] A method for predicting corn yield in a mulch film reused farmland comprises the following steps:

[0008] Collect images of corn in the current mulch-film-reused farmland throughout its growth period, including RGB images of mulch film coverage, multispectral images of soil moisture, thermal images of soil temperature, and multispectral images of vegetation characteristics;

[0009] Based on the RGB image of mulch film coverage, the threshold segmentation method was used to extract the mulch film coverage area, and the mulch film damage area was extracted based on the gray level co-occurrence matrix GLCM. The mulch film damage rate was obtained according to the number of pixels in the mulch film coverage area and the mulch film damage area. The normalized vegetation index NDVI was calculated using multispectral images of soil moisture and vegetation characteristics, and a linear regression model between the measured data of soil moisture content and NDVI was established to estimate the soil moisture content. The temperature value corresponding to each pixel in the soil temperature thermal imaging image was extracted to obtain the soil temperature data. Based on the linear relationship between the mulch film damage rate, soil moisture content and soil temperature data and the current corn yield, the growth period weight coefficient was introduced to construct a yield synergistic change prediction model for corn in the whole growth period of mulch film reuse farmland.

[0010] The data of mulch damage rate, soil moisture content and soil temperature during the key growth period of corn in mulch-reused farmland in the next year are input into the yield synergistic change prediction model to obtain the estimated corn yield for that year.

[0011] Furthermore, the method of extracting the film-covered area by using a threshold segmentation method based on the film-covered RGB image specifically includes the following steps:

[0012] The RGB image of mulch film is converted into a grayscale image through the grayscale formula. The grayscale formula is expressed as:

[0013] Gray=0.299R+0.587G+0.114B

[0014] The film-covered area is extracted by using a threshold segmentation method, and the thresholds of the film-covered area and the corn plant background are set. The image is binarized to extract the complete film-covered area; the film-covered area includes a damaged area and a completely covered area.

[0015] Furthermore, the extraction of ground film damaged area based on gray level co-occurrence matrix GLCM includes the following steps:

[0016] Calculate the entropy values ​​of the area fully covered by the mulch film and the area damaged by the mulch film (E1 and E2, E2>E1), and set the threshold E t (E1<E t <E2), if E>E t , it can be determined as the damaged area of ​​ground film; its calculation formula is:

[0017]

[0018] Wherein, E is the entropy value; P(i,j) is the gray level co-occurrence matrix; i and j represent different gray level combinations, and i = 0, 1, 2..., L-1; j = 0, 1, 2..., L-1.

[0019] Furthermore, the ground film damage rate is obtained according to the number of pixels in the ground film covered area and the ground film damaged area, which is expressed as:

[0020]

[0021] Among them, DR represents the film damage rate, n represents the number of pixels in the film damage area, and N represents the number of pixels in the film covered area.

[0022] Furthermore, before constructing the prediction model for the yield synergistic change of corn in the whole growth period of the film-reused farmland, the film damage rate, soil moisture content and soil temperature data are encoded and processed, which specifically includes the following steps:

[0023] The Z-Score standardization method was used to calculate the damage rate of ground film x 1j , soil moisture content x 2j and soil temperature x 3j After data processing, the standardized film damage rate z is obtained. 1j , soil moisture content 2j and soil temperature z 3j ; where j = 1, 2, 3, 4, 5 represent the seedling stage, jointing stage, tasseling stage, grain filling stage and maturity stage respectively;

[0024] The z of each growth period 1j 、z 2j and z 3j Form a feature vector V j , respectively:

[0025] V1=[z 11 ,z 21 ,z 31 ],V2=[z 12 ,z 22 ,z 32 ],V3=[z 13 ,z 23 ,z 33 ],V4=[z 14 ,z 24 ,z 34 ],V5=[z 15 ,z 25 ,z 35 ]

[0026] Sequential coding is used to assign coding values ​​to each growth stage; the seedling stage is coded as 1, the jointing stage is coded as 2, the tasseling stage is coded as 3, the filling stage is coded as 4, and the maturity stage is coded as 5. The coding feature vector is expressed as:

[0027] V c1 =[1,z 11 ,z 21,z 31 ],V c2 =[2,z 12 ,z 22 ,z 32 ],V c3 =[3,z 13 ,z 23 ,z 33 ],V c4 =[4,z 14 ,z 24 ,z 34 ],V c5 =[5,z 15 ,z 25 ,z 35 ].

[0028] Furthermore, the yield synergistic change prediction model of corn in the whole growth period of the film-mulching farmland is expressed as:

[0029]

[0030] Among them, β0 is the intercept term, which represents the basic yield under ideal environmental conditions; β ij is the regression coefficient, X 1j is the coded ground film damage rate, X 2j is the coded soil moisture content, X 3j is the coded soil temperature, Y is the corn yield, i=1, 2, 3 correspond to the film damage rate, soil moisture content, and soil temperature respectively, j=1, 2, 3, 4, 5 correspond to the seedling stage, jointing stage, tasseling stage, filling stage, and maturity stage respectively; ε is the error term; ω j is the weight coefficient of the fertility period, and the weight coefficient of the fertility period satisfies And 0≤ω j ≤1.

[0031] The present invention also includes a corn yield prediction system for film-mulching farmland, comprising:

[0032] The acquisition module is used to collect images of corn in the current mulch-film reused farmland throughout its growth period, including RGB images of mulch film coverage, multispectral images of soil moisture, thermal images of soil temperature, and multispectral images of vegetation characteristics;

[0033] The model building module is used to extract the mulch coverage area based on the mulch coverage RGB image using the threshold segmentation method, and extract the mulch damage area based on the gray level co-occurrence matrix GLCM, and obtain the mulch damage rate based on the number of pixels in the mulch coverage area and the mulch damage area; calculate the normalized vegetation index NDVI using multispectral images of soil moisture and vegetation characteristics, establish a linear regression model between the measured data of soil moisture content and NDVI, and estimate the soil moisture content; extract the temperature value corresponding to each pixel in the soil temperature thermal imaging image to obtain the soil temperature data; based on the linear relationship between the mulch damage rate, soil moisture content and soil temperature data and the current corn yield, introduce the growth period weight coefficient, and construct a yield synergistic change prediction model for corn in the key growth period of mulch reuse farmland;

[0034] The prediction module is used to input the film damage rate, soil moisture content and soil temperature data of the whole growth period of corn in the film-reused farmland in the next year into the yield coordinated change prediction model to obtain the estimated corn yield of that year.

[0035] The present invention also includes a computer device for predicting corn yield in film-reused farmland, including: a memory, a processor, and a computer program stored in the memory. When the processor executes the computer program, the steps of the method for predicting corn yield in film-reused farmland are implemented.

[0036] The present invention also includes a readable storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, they are used to execute the steps of the method for predicting corn yield in film-reused farmland.

[0037] The present invention provides a method for predicting corn yield in a film-mulching farmland, which has the following beneficial effects:

[0038] The present invention collects remote sensing images throughout the entire growth period to comprehensively reflect the growth status of corn, and constructs a prediction model for the coordinated changes in film breakage rate, soil moisture content and soil temperature and yield in the key growth period of corn in film-reused farmland, establishes the direct and interactive relationships between the three factors and the yield, and adds a weight distribution item for the growth period by considering the different sensitivities of corn in different growth periods to film breakage rate, soil moisture content and soil temperature, thereby achieving accurate prediction of corn yield in film-reused farmland for many years and providing decision-making support for the film-reused corn planting model. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a flowchart of a method for predicting corn yield in a film-reused farmland according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0041] The present invention proposes a method for predicting corn yield in mulch-film reused farmland based on UAV remote sensing. It comprehensively considers the dynamic changes of soil moisture and soil temperature after mulch film damage and the impact on corn yield, collects remote sensing images throughout the growth period to comprehensively reflect the growth status of corn, and constructs a mulch film damage-soil moisture-soil temperature-yield coordinated change prediction model to achieve accurate yield prediction and provide strong support for agricultural production decision-making. Figure 1 As shown, the method specifically comprises the following steps:

[0042] S1. Collect relevant indicators and drone remote sensing images of key growth periods of corn in mulch-film reused farmland: collect measured data on mulch film damage rate, corn growth, soil moisture content and soil temperature at the seedling, jointing, tasseling, grain filling and maturity stages, and measure corn yield after corn harvest. Use drones equipped with visible light cameras, multispectral cameras and thermal infrared cameras to obtain RGB images of mulch film coverage, multispectral images of soil moisture, thermal images of soil temperature and multispectral images of vegetation characteristics.

[0043] S2. Identification of damaged areas of ground film:

[0044] 1) The RGB color image is converted into a grayscale image using the grayscale formula (Gray = 0.299R + 0.587G + 0.114B), and the threshold segmentation method is used to extract the mulch film covered area. The thresholds of the mulch film covered area (including damaged and fully covered) and the background such as corn plants are set, and the image is binarized to extract the complete mulch film covered area.

[0045] 2) Extract the damaged area of ​​ground film based on gray level co-occurrence matrix (GLCM) and use the formula Calculate the entropy values ​​of the area fully covered by the mulch film and the area damaged by the mulch film (E1 and E2, E2>E1), and set the threshold E t (E1<E t <E2), if E>E t , which can be determined as the damaged area of ​​ground film.

[0046] 3) The number of pixels in the damaged film area is recorded as n, and the number of pixels in the film covered area is recorded as N. The film damage rate is calculated as follows:

[0047] S3. The normalized difference vegetation index (NDVI) was calculated using the acquired multispectral image data, and a linear regression model between the measured soil moisture data and NDVI was established. The correlation coefficient (R2), root mean square error (RMSE) and other indicators were used to evaluate the accuracy of the model, so as to accurately estimate the soil moisture content.

[0048] S4. After performing radiation correction, geometric correction and registration on the acquired thermal infrared image, the temperature value corresponding to each pixel is extracted to obtain the soil temperature data.

[0049] S5. Z-Score standardization method was used to calculate the damage rate of ground film (x 1j ), soil moisture content (x 2j ) and soil temperature (x 3j ) data were processed to obtain the standardized film damage rate (z 1j ), soil moisture content (z 2j ) and soil temperature (z 3j ), where j = 1, 2, 3, 4, 5 represent the seedling stage, jointing stage, tasseling stage, grain filling stage and maturity stage respectively. 1j 、z 2j and z 3j Form a comprehensive feature vector V j , respectively:

[0050] V1=[z 11 ,z 21 ,z 31 ],V2=[z 12 ,z 22 ,z 32 ],V3=[z 13 ,z 23 ,z 33 ],V4=[z 14 ,z 24 ,z 34 ],V5=[z 15 ,z 25 ,z 35 ]

[0051] Sequential coding is used to assign coding values ​​to each growth stage. The seedling stage is coded as 1, the jointing stage is coded as 2, the tasseling stage is coded as 3, the filling stage is coded as 4, and the maturity stage is coded as 5. The coding feature vector is expressed as follows:

[0052] V c1 =[1,z 11 ,z 21 ,z 31 ],V c2 =[2,z 12 ,z 22 ,z32 ],V c3 =[3,z 13 ,z 23 ,z 33 ],V c4 =[4,z 14 ,z 24 ,z 34 ],V c5 =[5,z 15 ,z 25 ,z 35 ].

[0053] S6, considering the coded ground film damage rate (X 1j ), soil moisture content (X 2j ) and soil temperature (X 3j ) and corn yield (Y). At the same time, the growth period weight coefficient was introduced to construct a prediction model of mulch damage rate-soil moisture content-soil temperature-yield synergistic changes in the key growth period of corn in mulch reuse farmland:

[0054]

[0055] Assume that the weight of seedling period is ω1, the weight of jointing period is ω2, the weight of tasseling period is ω3, the weight of grain filling period is ω4, and the weight of maturity period is ω5. These weight coefficients satisfy And 0≤ω j ≤1.

[0056] Incorporating the weight coefficient into the model, the model is modified as follows:

[0057]

[0058] Among them, β0 is the intercept term, which represents the basic yield under ideal environmental conditions, and β ij is the regression coefficient (i=1, 2, 3 correspond to the damage rate of plastic film, soil moisture content, and soil temperature respectively; j=1, 2, 3, 4, 5 correspond to the seedling stage, jointing stage, tasseling stage, filling stage, and maturity stage respectively), ε is the error term, which is usually assumed to have a mean of 0 and a variance of σ 2 The normal distribution of .

[0059] The collected data on film damage rate, soil moisture content, soil temperature and yield were randomly divided into training set (70%), validation set (15%) and test set (15%). The training set data was used to determine the regression coefficient using the least squares method, and the validation set and test set were used to evaluate the model to obtain the optimal yield prediction model.

[0060] S5. Substitute the relevant characteristic parameters of the second year's corn field with mulch film reuse into the above yield prediction model to obtain the estimated corn yield By collecting the corn yield data of the same agricultural production area, determine the general level of local corn yield The difference between the two is If It is possible to consider continuing to carry out plastic film reuse for corn planting. Otherwise, it is necessary to consider changing the planting method.

[0061] The present invention reduces the plastic film breakage rate during the corn growth stage: By using this method, it is possible to obtain data on plastic film breakage rate, soil moisture content, and soil temperature through an unmanned aerial vehicle, greatly reducing the impact of human activities on plastic film breakage during the corn growth stage, effectively extending the service life of the covered plastic film, and reducing farmland residual film pollution. Improve the accuracy of corn yield measurement: By using the co-variation prediction model of plastic film breakage rate - soil moisture content - soil temperature - yield, not only can the impact effect of soil water and heat changes caused by continuous plastic film breakage on yield be considered, but also the sensitivity of corn growth at different stages to the above three parameters is embedded. Compared with the traditional corn yield measurement method, the application of this model reduces labor costs and improves the yield measurement accuracy. Provide the appropriate service life for plastic film reuse: By predicting the corn yield of the plastic film reuse farmland in the next year, the applicability of the plastic film reuse planting mode can be evaluated, the effective plastic film reuse years can be determined, and the economic and ecological benefits can be maximized.

[0062] Example:

[0063] S1. Select the farmland where plastic film reuse for corn planting is carried out in the first year. Conduct field sampling of plastic film breakage rate, soil moisture content, and soil temperature, and collect remote sensing images during the seedling stage, jointing stage, tasseling stage, filling stage, and maturity stage of corn. Under clear weather conditions at each growth stage, use an unmanned aerial vehicle equipped with a visible light camera, a multispectral camera, and a thermal infrared camera to photograph the entire farmland according to a predetermined flight route, respectively obtain the RGB image of plastic film coverage, the multispectral image of soil moisture and vegetation characteristics, and the thermal imaging image of soil temperature, and select 10 corn plants for yield measurement after the corn is harvested.

[0064] S2. Perform gray-scale processing on the RGB image of plastic film coverage, and use the threshold segmentation method to extract the plastic film coverage area. Through multiple experiments and observations, set the threshold between the plastic film coverage area (including breakage and complete coverage) and the background such as corn plants, perform binary processing on the image, and extract the complete plastic film coverage area. Extract the plastic film breakage area based on GLCM. Calculate the entropy values (E1 and E2) of the completely covered plastic film area and the plastic film breakage area through the corresponding formula, and set the threshold E t (E1 < E t < E2). According to the judgment rule, if the E of a certain area > E t , then determine that this area is the plastic film breakage area. Calculate the plastic film breakage rate at each growth stage according to the plastic film breakage rate calculation formula, and verify and supplement it through the measured plastic film breakage rate.

[0065] S3. NDVI is calculated by extracting near infrared band (NIR) and red light band (R) from multispectral images, and a linear model between the measured value of soil moisture content and NDVI is established. After the model is calibrated and verified, the soil moisture content is estimated.

[0066] S4. Perform preprocessing operations such as radiation correction, geometric correction and registration on the acquired thermal infrared images to ensure the accuracy and consistency of the image data. After processing, extract the temperature value corresponding to each pixel to obtain the soil temperature data of each growth period, and conduct comparative analysis through the measured soil temperature to further improve the accuracy of soil temperature data obtained by the image.

[0067] S5. Z-Score standardization method was used to calculate the damage rate of ground film (x 1j ), soil moisture content (x 2j ) and soil temperature (x 3j ) data. Taking the seedling stage as an example, assuming that the standardized film damage rate (z 1j ), soil moisture content (z 2j ), soil temperature (z 3j ). Similarly, the data of jointing stage, tasseling stage, grain filling stage and maturity stage were also standardized. 1j 、z 2j and z 3j Form a comprehensive feature vector V j . A coding value is assigned to each growth stage according to the sequential coding, and the seedling stage, jointing stage, tasseling stage, grain filling stage and maturity stage are 1, 2, 3, 4 and 5 respectively.

[0068] S6, considering the coded ground film damage rate (X 1j ), soil moisture content (X 2j ), soil temperature (X 3j ) and corn yield (Y), and the weight coefficient of the reproductive period (ω j ), the weight coefficients of each growth period can be determined based on past experience and the characteristics of agricultural production in the region, and a prediction model of mulch damage rate-soil moisture content-soil temperature-yield synergistic changes in the key growth period of corn in mulch-reused farmland was constructed:

[0069]

[0070] The regression coefficients were determined by the least squares method using the training set data. After a series of calculations and iterations, the specific values ​​of each regression coefficient were obtained. The model was evaluated using the validation set and the test set. By comparing the differences between the prediction results under different model parameters and the actual yield data, the model parameters were continuously adjusted to finally obtain the optimal yield prediction model.

[0071] S7, using the above model to predict the corn yield in the second year, and obtain the estimated corn yield The general level of local corn production The difference between the two is pass Determine whether to continue planting corn with mulch reuse.

[0072] The key points of the present invention are as follows: (1) Dynamic monitoring of film damage in film-reused corn farmland, obtaining remote sensing images through unmanned aerial vehicles, clarifying the film damage area based on the grayscale co-occurrence matrix, and calculating the film damage rate; (2) Constructing a comprehensive feature vector of film damage rate, soil moisture content, and soil temperature, and encoding the growth period, which mainly considers the intrinsic correlation and interaction between the three parameters in different growth periods, and effectively improves the accuracy level of the least squares method based on these feature vectors; (3) Constructing a film damage rate-soil moisture content-soil temperature-yield coordinated change prediction model considering the key growth period of corn in film-reused farmland, firstly establish the direct and interactive relationship between the three factors and the yield, and by considering the different sensitivities of corn in different growth periods to film damage rate, soil moisture content and soil temperature, increase the weight distribution item of the growth period, realize the accurate prediction of corn yield in film-reused farmland for many years, and provide decision-making support for the film-reused corn planting mode.

[0073] Based on the same inventive concept, the present invention also proposes a corn yield prediction system for film-mulching farmland, comprising:

[0074] The acquisition module is used to collect images of the key growth period of corn in the current mulch-reused farmland, including RGB images of mulch coverage, multispectral images of soil moisture, thermal images of soil temperature, and multispectral images of vegetation characteristics.

[0075] The model building module is used to extract the mulch coverage area based on the mulch coverage RGB image using the threshold segmentation method, and extract the mulch damage area based on the gray level co-occurrence matrix GLCM, and obtain the mulch damage rate according to the number of pixels in the mulch coverage area and the mulch damage area; calculate the normalized vegetation index NDVI using the multispectral image of soil moisture and vegetation characteristics, establish a linear regression model between the measured soil moisture data and NDVI, and estimate the soil moisture content; extract the temperature value corresponding to each pixel in the soil temperature thermal imaging image to obtain the soil temperature data; based on the linear relationship between the mulch damage rate, soil moisture content and soil temperature data and the current corn yield, introduce the growth period weight coefficient, and construct a yield synergistic change prediction model for corn in the key growth period of mulch reuse farmland.

[0076] The prediction module is used to input the mulch damage rate, soil moisture content and soil temperature data of the key growth period of corn in the mulch reuse farmland in the next year into the yield coordinated change prediction model to obtain the estimated corn yield for that year.

[0077] The present invention also proposes a computer device for predicting corn yield in film-reused farmland, comprising: a memory, a processor and a computer program stored in the memory, and the processor implements the steps of a method for predicting corn yield in film-reused farmland when executing the computer program.

[0078] The present invention also proposes a readable storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, they are used to execute the steps of a method for predicting corn yield in a mulch-film reused farmland.

[0079] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for predicting corn yield in film-mulching farmland, characterized in that: The following steps are involved: Collect images of corn in the current mulch-film-reused farmland throughout its growth period, including RGB images of mulch film coverage, multispectral images of soil moisture, thermal images of soil temperature, and multispectral images of vegetation characteristics; Based on the RGB image of mulch film coverage, the threshold segmentation method was used to extract the mulch film coverage area, and the mulch film damage area was extracted based on the gray level co-occurrence matrix GLCM. The mulch film damage rate was obtained according to the number of pixels in the mulch film coverage area and the mulch film damage area. The normalized vegetation index NDVI was calculated using multispectral images of soil moisture and vegetation characteristics, and a linear regression model between the measured data of soil moisture content and NDVI was established to estimate the soil moisture content. The temperature value corresponding to each pixel in the soil temperature thermal imaging image was extracted to obtain the soil temperature data. Based on the linear relationship between the mulch film damage rate, soil moisture content and soil temperature data and the current corn yield, the growth period weight coefficient was introduced to construct a yield synergistic change prediction model for corn in the whole growth period of mulch film reuse farmland. The data of mulch damage rate, soil moisture content and soil temperature during the key growth period of corn in mulch-reused farmland in the next year are input into the yield synergistic change prediction model to obtain the estimated corn yield for that year.

2. The method for predicting corn yield in film-mulching farmland according to claim 1, characterized in that: The method of extracting the film-covered area by using a threshold segmentation method based on the film-covered RGB image specifically includes the following steps: The grayscale formula Gray is used to convert the RGB image of ground film coverage into a grayscale image. The grayscale formula is expressed as: Gray=0.299R+0.587G+0.114B, Among them, R, G, B represent red, green, and blue colors; The film-covered area is extracted by using a threshold segmentation method, and the thresholds of the film-covered area and the corn plant background are set. The image is binarized to extract the complete film-covered area; the film-covered area includes a damaged area and a completely covered area.

3. The method for predicting corn yield in film-mulching farmland according to claim 2, characterized in that: The method of extracting the damaged area of ​​ground film based on the gray level co-occurrence matrix GLCM comprises the following steps: Calculate the entropy values ​​of the area fully covered by the mulch film and the area damaged by the mulch film (E1 and E2, E2>E1), and set the threshold E t (E1<E t <E2), if the selected area’s E>E t , which can be determined as the damaged area of ​​ground film; its calculation formula is: Wherein, E is the entropy value; P(i,j) is the gray level co-occurrence matrix; i and j represent different gray level combinations, and i = 0, 1, 2..., L-1; j = 0, 1, 2..., L-1.

4. The method for predicting corn yield in film-mulching farmland according to claim 1, characterized in that: According to the number of pixels in the mulch-covered area and the mulch-damaged area, the mulch-damaged rate is obtained, which is expressed as: Among them, DR represents the film damage rate, n represents the number of pixels in the film damage area, and N represents the number of pixels in the film covered area.

5. The method for predicting corn yield in film-mulching farmland according to claim 1, characterized in that: Before constructing the prediction model for the yield synergistic change of corn in the whole growth period of the mulch film reuse farmland, the data of mulch film damage rate, soil moisture content and soil temperature are encoded and processed, which specifically includes the following steps: The Z-Score standardization method was used to calculate the damage rate of ground film x 1j , soil moisture content x 2j and soil temperature x 3j After data processing, the standardized film damage rate z is obtained. 1j , soil moisture content 2j and soil temperature z 3j ; where j = 1, 2, 3, 4, 5 represent the seedling stage, jointing stage, tasseling stage, grain filling stage and maturity stage respectively; The z of each growth period 1j 、z 2j and z 3j Form a feature vector V j , respectively: V1=[with 11 ,with 21 ,with 31 ],V2=[z 12 ,with 22 ,with 32 ],V3=[z 13 ,with 23 ,with 33 ],V4=[z 14 ,with 24 ,with 34 ],V5=[z 15 ,with 25 ,with 35 ] Sequential coding is used to assign coding values ​​to each growth stage; the seedling stage is coded as 1, the jointing stage is coded as 2, the tasseling stage is coded as 3, the filling stage is coded as 4, and the maturity stage is coded as 5. The coding feature vector is expressed as: In c1 =[1,z 11 ,with 21 ,with 31 ],In c2 =[2,z 12 ,with 22 ,with 32 ],In c3 =[3,z 13 ,with 23 ,with 33 ],In c4 =[4,z 14 ,with 24 ,with 34 ], In c5 =[5,z 15 ,with 25 ,with 35 ]。 6. The method for predicting corn yield in film-mulching farmland according to claim 5, characterized in that: The yield synergistic change prediction model of corn in the whole growth period of the film-mulching farmland is expressed as: Among them, β0 is the intercept term, which represents the basic yield under ideal environmental conditions; β ij is the regression coefficient, X 1j is the coded ground film damage rate, X 2j is the coded soil moisture content, X 3j is the coded soil temperature, Y is the corn yield, i=1, 2, 3 correspond to the film damage rate, soil moisture content, and soil temperature respectively, j=1, 2, 3, 4, 5 correspond to the seedling stage, jointing stage, tasseling stage, filling stage, and maturity stage respectively; ε is the error term; ω j is the weight coefficient of the fertility period, and the weight coefficient of the fertility period satisfies And 0≤ω j ≤1.

7. A system for predicting corn yield in film-mulching farmland, characterized in that: include: The acquisition module is used to collect images of corn in the current mulch-film reused farmland throughout its growth period, including RGB images of mulch film coverage, multispectral images of soil moisture, thermal images of soil temperature, and multispectral images of vegetation characteristics; The model building module is used to extract the mulch coverage area based on the mulch coverage RGB image using the threshold segmentation method, and extract the mulch damage area based on the gray-level co-occurrence matrix GLCM, and obtain the mulch damage rate based on the number of pixels in the mulch coverage area and the mulch damage area; the normalized vegetation index NDVI is calculated using the multispectral image of soil moisture and vegetation characteristics, and a linear regression model between the measured soil moisture data and NDVI is established to estimate the soil moisture content; the temperature value corresponding to each pixel in the soil temperature thermal imaging image is extracted to obtain the soil temperature data; based on the linear relationship between the mulch damage rate, soil moisture content and soil temperature data and the current corn yield, the growth period weight coefficient is introduced to construct a yield synergistic change prediction model for corn in the whole growth period of mulch reuse farmland; The prediction module is used to input the film damage rate, soil moisture content and soil temperature data of the key growth period of corn in the film-reused farmland in the next year into the yield coordinated change prediction model to obtain the estimated corn yield for that year.

8. A computer device for predicting corn yield in film-reused farmland, characterized in that: include: A memory, a processor, and a computer program stored in the memory, wherein when the processor executes the computer program, the steps of the method for predicting corn yield in a mulch-film reused farmland according to any one of claims 1 to 6 are implemented.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor, they are used to execute the steps of the method for predicting corn yield in film-reused farmland according to any one of claims 1 to 6.

Images