Grain-bean hybrid crop growth estimation method and device based on multispectral technology

Abstract

The present application relates to the technical field of remote sensing monitoring of crops, and provides a method and device for estimating the growth of a mixed crop of grass and beans based on multispectral technology, wherein the method includes collecting multispectral data of a mixed area, calculating a vegetation index and a texture index based on the multispectral data, determining the biomass and nitrogen content of the crops in the mixed area, and constructing multiple estimation models for estimating biomass based on a general linear regression algorithm and a partial least squares regression algorithm, and screening out a target estimation model from the multiple estimation models, and estimating the growth of crops in the mixed area through the target estimation model. It is possible to combine the texture index that expresses the difference characteristics of spatial details with the vegetation index to establish an estimation model with high estimation accuracy, and to establish an estimation model for mixed areas with complex phenotypic characteristics of crop objects and high estimation difficulty, so as to achieve non-destructive, efficient and accurate estimation of the biomass and nitrogen content of mixed crops of grass and beans.

Description

Technical Field

[0001] The present application relates to the field of crop remote sensing monitoring technology, and in particular to a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology. The present application also relates to a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology, a computing device, and a computer-readable storage medium. Background Art

[0002] The study found that, for example, oats, continuous cropping of gramineous crops, such as oats, has problems that limit its high-yield cultivation, such as decreased soil fertility and increased weeds. Mixed sowing with legumes can avoid the above planting problems and effectively improve the production performance and nutritional quality of forage. Mixed sowing of gramineous and legumes can make full use of land resources and climate resources, realize intensive planting, improve land utilization rate, and improve soil structure. It is an effective planting method to promote high yield and increase production of crops. The stems and branches of the arrow pea are tender, palatable, cold-resistant, and drought-resistant. It can fertilize and fix nitrogen. It is a widely used high-quality and high-yield legume forage. Mixed sowing of oats and arrow peas is a typical mixed sowing model for the current construction of high-quality artificial forage bases. Reasonable mixed sowing combinations and mixed sowing ratios are the prerequisites for the mixed sowing of gramineous and legumes to realize their production potential. They can promote interspecific synergy and improve resource utilization efficiency, so that the established artificial grassland can maintain ideal productivity.

[0003] Biomass and nitrogen are important indicators for measuring forage growth and nutritional abundance, and they play an important role in the process of forage field management. The ability to monitor the growth and nutritional status of crops in the field in real time with low cost and high efficiency is the focus of many scholars. For mixed artificial grasslands, it is extremely important to monitor the changes in biomass and nutritional content in real time. According to the actual situation of field planting, field management measures such as water and fertilizer should be adjusted in time; the optimal harvest period can also be determined based on the dynamic changes of biomass and quality, so as to maximize the benefits of forage yield and quality. The traditional method of measuring biomass and quality is destructive sampling. Although it has a high degree of accuracy, it is non-dynamic, lagging, destructive, and has high time and economic costs.

[0004] The integration of advanced information and communication technology with agricultural production technology plays an important role in achieving precision agricultural management. As precision agriculture continues to attract the attention of many scholars at home and abroad, its concept continues to deepen, technology continues to develop, and its application in actual agricultural production is gradually becoming a reality. With the help of information technology, people can monitor forage production and manage livestock conditions more quickly and dynamically. As a non-destructive method, drones and remote sensing technology can obtain standardized and repeated information at different spatial and temporal scales.

[0005] Satellite or airborne imagery has been tested in many studies to assess the plant composition, structure, quality or quantity of grasslands. Typically, point-level spectral reflectance data are collected in the visible to short-wave infrared region of the electromagnetic spectrum. Such platforms can be equipped with different types of sensors, such as RGB sensors, including red, green and blue, and multispectral sensors that include near-infrared spectral bands (wavelength range between 400 and 1000 nm). These are often used to phenotypic analysis of various agronomic traits of forages, including biomass, yield, disease resistance, and vegetation cover, as well as analysis of forage nutritional qualities such as nitrogen content, crude protein, and crude fiber. Despite the high estimation potential of spectral techniques, their relatively low spatial and temporal resolution still limits their application on crops.

[0006] In order to compensate for the interference of soil or atmosphere in spectral reflectance and increase the sensitivity of vegetation characteristics, vegetation index can be used as a way to reduce these interferences. The more common vegetation index uses red and near-infrared bands and is called the Normalized Difference Vegetation Index (NDVI). However, NDVI also has its own limitations because it becomes saturated when the biomass is too high and changes with soil color and climate conditions. Therefore, a non-destructive, efficient and accurate estimation method is urgently needed to estimate the biomass and nitrogen content of mixed crops of grass and beans. Summary of the invention

[0007] In view of this, the embodiment of the present application provides a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology to solve the technical defects existing in the prior art. The embodiment of the present application also provides a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology, a computing device, and a computer-readable storage medium.

[0008] According to a first aspect of an embodiment of the present application, a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology is provided, comprising:

[0009] Collecting multispectral data of a mixed-crop area, wherein at least two crops are mixed in the mixed-crop area;

[0010] Calculating a vegetation index based on the multispectral data, and calculating a texture index associated with the vegetation index;

[0011] determining the biomass and nitrogen content of crops in the intercropping area;

[0012] Based on a general linear regression algorithm, a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index are constructed;

[0013] Based on a partial least squares regression algorithm, constructing a third estimation model for estimating the biomass through the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index;

[0014] According to a preset screening principle, a target estimation model is selected from the first estimation model, the second estimation model, the third estimation model and the fourth estimation model, and the growth conditions of the crops in the mixed-cropping area are estimated according to the target estimation model.

[0015] Optionally, collecting multispectral data of the interbreeding area includes:

[0016] Based on an image acquisition device including a multispectral camera, multispectral data of the mixed-seed area is collected, wherein the collection bands of the multispectral data include blue light, green light, red light, red edge and near infrared.

[0017] Optionally, calculating a vegetation index according to the multispectral data includes:

[0018] The vegetation index including the normalized difference vegetation index, the normalized difference red edge vegetation index, the green normalized difference vegetation index, the leaf chlorophyll index and the optimized soil adjusted vegetation index is calculated according to the multispectral data.

[0019] Optionally, the calculating a texture index associated with the vegetation index includes:

[0020] Based on the symbiotic grayscale matrix, the initial texture index associated with the vegetation index in different bands is calculated, and a hybrid area image is generated according to the calculation results, wherein the texture index corresponding to any block in the hybrid area image is determined by the average reflectivity of the block, and the texture index includes mean, contrast, difference, entropy, energy, homogeneity, and second-order moment.

[0021] Optionally, the constructing of a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a general linear regression algorithm comprises:

[0022] Based on a general linear regression algorithm, a leave-one-out cross validation method is used to construct the first estimation model for estimating the biomass through the vegetation index and the texture index;

[0023] Based on a general linear regression algorithm and using a leave-one-out cross validation method, the second estimation model for estimating the nitrogen content through the vegetation index and the texture index is constructed.

[0024] Optionally, the constructing of a third estimation model for estimating the biomass by using the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content by using the vegetation index and the texture index based on a partial least squares regression algorithm comprises:

[0025] Based on the partial least squares regression algorithm, a leave-one-out cross validation method is adopted to construct the third estimation model for estimating the biomass through the vegetation index and the texture index;

[0026] Based on the partial least squares regression algorithm and using the leave-one-out cross validation method, the fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index is constructed.

[0027] Optionally, selecting a target estimation model from the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model according to a preset screening principle includes:

[0028] determining a Pearson correlation coefficient associated with the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model;

[0029] determining a root mean square error associated with the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model;

[0030] A target estimation model is determined according to the Pearson correlation coefficient and the root mean square error.

[0031] According to a second aspect of an embodiment of the present application, a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology is provided, comprising:

[0032] A collection module is configured to collect multispectral data of a mixed-crop area, wherein at least two crops are mixed in the mixed-crop area;

[0033] A calculation module is configured to calculate a vegetation index according to the multispectral data, and calculate a texture index associated with the vegetation index;

[0034] A determination module configured to determine the biomass and nitrogen content of the crops in the mixed-cropping area;

[0035] A first construction module is configured to construct a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a general linear regression algorithm;

[0036] A second construction module is configured to construct a third estimation model for estimating the biomass through the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a partial least squares regression algorithm;

[0037] The screening module is configured to select a target estimation model from among the first estimation model, the second estimation model, the third estimation model and the fourth estimation model according to a preset screening principle, and estimate the growth of crops in the mixed-crop area according to the target estimation model.

[0038] According to a third aspect of an embodiment of the present application, a computing device is provided, including:

[0039] Memory and processor;

[0040] The memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, the steps of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology are implemented.

[0041] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology are implemented.

[0042] According to a fifth aspect of an embodiment of the present application, a chip is provided, which stores a computer program, and when the computer program is executed by the chip, the steps of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology are implemented.

[0043] The method for estimating the growth of a mixed crop of grass and beans based on multispectral technology provided in the present application is to collect multispectral data of a mixed area, wherein at least two crops are mixed in the mixed area; calculate the vegetation index according to the multispectral data, and calculate the texture index associated with the vegetation index; determine the biomass and nitrogen content of the crops in the mixed area; construct a first estimation model for estimating the biomass through the vegetation index and the texture index based on a general linear regression algorithm, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index; construct a third estimation model for estimating the biomass through the vegetation index and the texture index based on a partial least squares regression algorithm, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index; select a target estimation model from the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model through a preset screening principle, and estimate the growth of the crops in the mixed area according to the target estimation model. A non-destructive, efficient, and accurate estimation of the biomass and nitrogen content of mixed crops of grass and beans is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a flow chart of a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology provided in one embodiment of the present application;

[0046] Figure 2 This is a schematic diagram of the structure of a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology provided in one embodiment of the present application;

[0047] Figure 3 It is a structural block diagram of a computing device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0048] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0049] The terms used in one or more embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of the present application. The singular forms of "a", "said" and "the" used in one or more embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of the present application refers to and includes any or all possible combinations of one or more associated listed items.

[0050] It should be understood that, although the terms first, second, etc. may be used to describe various information in one or more embodiments of the present application, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0051] In the present application, a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology is provided. The present application also relates to a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology, a computing device, and a computer-readable storage medium, which are described in detail one by one in the following embodiments.

[0052] Remote sensing technology has great potential for development in the agricultural field. The continuous maturity of remote sensing technology has improved the resolution of data space, so the detailed information of the geometric features of the objects can be expressed more fully. The texture index, as a feature for expressing the difference in spatial details, has also begun to be added to practical applications. Among them, the estimation model based on the combination of vegetation index and texture index based on linear regression and machine learning algorithms may have a strong accuracy, but it will be affected to a certain extent by factors such as crop type, cultivation method, and cultivation environment. In previous studies, crops such as wheat, corn, rapeseed, and rice were the main research objects, and there was less research on grass crops, and the cultivation method was relatively simple.

[0053] Among them, the texture features of the gray-level co-occurrence matrix based on statistical methods are more common and widely used in research methods such as plant classification, forage biomass estimation, and nitrogen abundance and deficiency diagnosis. In addition to the linear regression analysis method, machine learning has also been used in remote sensing estimation technology.

[0054] Crop yield estimation and nutrient deficiency diagnosis have important practical production guidance value. First, the yield and quality of crops can be estimated based on their growth, leaf color and other apparent characteristics. Although this method is fast, intuitive and easy to operate, it is highly subjective and has poor accuracy. Second, sampling and chemical methods can be used to obtain crop plant samples to determine biological yield and nitrogen content, but sample preparation and nitrogen determination processes are time-consuming, labor-intensive, and cumbersome, and are destructive sampling, and growth information cannot be fed back in time during the critical period of crop growth. Third, with the help of handheld devices, drones and other equipment, crop phenotypic indicators can be acquired non-destructively, key vegetation indices can be screened, and prediction models between vegetation indices and biomass and nitrogen concentrations can be established. However, the problem with using spectral bands and vegetation indices alone is that the spatial variation of grayscale values ​​(pixels) in each pixel neighborhood is ignored, and the estimation accuracy is low. Fourth, previous studies using remote sensing technology to estimate artificial grassland biomass mainly focused on the planting of a single forage species, and the established estimation model cannot be extended to the estimation of biomass and quality of mixed artificial grassland. In view of this, the present application proposes a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology, which realizes the establishment of an estimation model by combining the texture index, which is a feature for expressing the difference in spatial details, with the vegetation index. This can improve the estimation accuracy and, for the mixed areas where grass and beans are planted, enable the establishment of an estimation model with more complex phenotypic characteristics of the crop objects and higher difficulty.

[0055] Figure 1A flowchart of a method for estimating the growth of a mixed crop of grass and beans based on multispectral technology according to an embodiment of the present application is shown, which specifically includes the following steps:

[0056] Step S102: collecting multispectral data of a mixed-crop area, wherein at least two crops are mixed in the mixed-crop area;

[0057] Step S104: calculating a vegetation index according to the multispectral data, and calculating a texture index associated with the vegetation index;

[0058] Step S106: determining the biomass and nitrogen content of the crops in the mixed-crop area;

[0059] Step S108: constructing a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a general linear regression algorithm;

[0060] Step S110: constructing a third estimation model for estimating the biomass through the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a partial least squares regression algorithm;

[0061] Step S112: selecting a target estimation model from among the first estimation model, the second estimation model, the third estimation model and the fourth estimation model according to a preset screening principle, and estimating the growth of crops in the mixed-cropping area according to the target estimation model.

[0062] The crops mixed in the mixed-cropping area include gramineous crops and legume crops. After collecting multispectral data of the mixed-cropping area of ​​gramineous and legume crops, the vegetation index and the texture index are determined according to the multispectral data.

[0063] Subsequently, the biomass and nitrogen content were measured by mowing and sampling the crops in the mixed-cropping area, weighing and drying them, and then modeling was carried out through the general linear regression algorithm and the partial least squares regression algorithm. Multiple models for evaluating biomass and nitrogen content through vegetation index and texture index were obtained. Finally, through screening, the accuracy and error of the models obtained by partial least squares regression modeling and general linear regression modeling were compared. According to the comparison results, the model with better accuracy was selected, that is, the target estimation model for estimating dry matter yield and nitrogen fixation.

[0064] Furthermore, in step S102, the process of collecting multispectral data of the mixed area is specifically implemented as follows in this embodiment:

[0065] Based on an image acquisition device including a multispectral camera, multispectral data of the mixed-seed area is collected, wherein the collection bands of the multispectral data include blue light, green light, red light, red edge and near infrared.

[0066] Among them, image acquisition equipment includes handheld devices, drones, etc. In actual usage scenarios, the multispectral camera is installed on the drone to capture multispectral images. The drone is equipped with a universal joint to stabilize the platform, which effectively improves the image acquisition efficiency.

[0067] Based on this, the multispectral camera used includes 5 channels, namely 5 bands, namely 450 nm blue light, 560 nm green light, 650 nm red light, 730 nm red edge and 840 nm near infrared, to capture multispectral images of plant canopies.

[0068] Furthermore, in step S104, the process of calculating the vegetation index according to the multispectral data is specifically implemented as follows in this embodiment:

[0069] The vegetation index including the normalized difference vegetation index, the normalized difference red edge vegetation index, the green normalized difference vegetation index, the leaf chlorophyll index and the optimized soil adjusted vegetation index is calculated according to the multispectral data.

[0070] Among them, according to the obtained multispectral data, the normalized vegetation index NDVI, the normalized difference red edge vegetation index NDRE, the green normalized difference vegetation index GNDVI, the leaf chlorophyll index LCI and the optimized soil conditioned vegetation index OSAVI are calculated. Specifically, as shown in Table 1 below:

[0071] Table 1 Vegetation index comparison table

[0072] Vegetation Index name formula NDVI Normalized Difference Vegetation Index (NIR-R) / (NIR+R) NDRE Normalized difference red edge vegetation index (NIR−RE) / (NIR+RE) GNDVI Green Normalized Difference Vegetation Index (NIR−G) / (NIR+G) LCI Leaf chlorophyll index (NIR−RE) / (NIR+R) OSAVI Optimizing Soil Adjustment Vegetation Index (NIR−R) / (NIR+R+0.16)

[0073] Among them, NIR represents the spectral characteristics of the near-infrared band, R represents the red light band, G represents the green light band, and RE represents the red edge band.

[0074] Furthermore, in step S102, the process of calculating the texture index associated with the vegetation index is specifically implemented as follows in this embodiment:

[0075] Based on the symbiotic grayscale matrix, the initial texture index associated with the vegetation index in different bands is calculated, and a hybrid area image is generated according to the calculation results, wherein the texture index corresponding to any block in the hybrid area image is determined by the average reflectivity of the block, and the texture index includes mean, contrast, difference, entropy, energy, homogeneity, and second-order moment.

[0076] Specifically, the texture index is shown in Table 2 below:

[0077] Table 2 Selected texture index

[0078]

[0079] Where i is the grayscale of any point (x, y), and j is the grayscale of another point deviating from this point; P i,j is the frequency of occurrence of pixels with a grayscale of j and a distance of d from (x, y).

[0080] Based on this, the calculation of vegetation index and the proportion of different spectral bands are different, which are used for specific applications and use the spectral and texture information carried by drones to estimate the agronomic traits and nutritional quality of forage. However, the problem with using spectral bands and vegetation indices alone is that the spatial variation of grayscale values ​​(pixels) within each pixel neighborhood is ignored. In this embodiment, the above gap can be compensated by analyzing the texture index. Since texture is a spatial variation within an image and is related to the structure and heterogeneity of vegetation, the combination of vegetation index and texture index can achieve a more accurate estimation of forage biomass and nutritional quality.

[0081] For example, in actual application scenarios, for the model for estimating the biomass and nitrogen content of grassland mixed with oats and samphire, a mixed sowing ratio experiment of 5 kinds of oats and samphire was set up in a plain area in 2021-2022 and 2022-2023, and the grass biomass and nitrogen content of each mixed sowing ratio in 5-6 growth stages were quantitatively determined. At the same time, a multi-spectral camera on an unmanned aerial vehicle was used to obtain 5 spectral reflectances of different growth stages and different mixed sowing ratios, 5 vegetation indices were calculated, and 7 texture indices were extracted. A general linear model and a partial least squares regression model were constructed to estimate the biomass and nitrogen content of the mixed grassland.

[0082] Specifically, forage oats and peas, the varieties of which are "Fuweng" and "Chuanbei", were sown on October 21, 2021 and October 21, 2022, respectively. A total of five sowing methods were set, including T1 (grain: bean = 1:0), T2 (grain: bean = 3:1), T3 (grain: bean = 1:1), T4 (grain: bean = 1:3), and T5 (grain: bean = 0:1). The mixed sowing ratio is calculated according to the number of seeds, and the total planting density is 300 seeds / m 2 . Row sowing, oats and peas mixed in the same row, row spacing 30 cm. Plot area 15 m 2 (3 m × 5 m), randomized block design, repeated 3 times, a total of 15 plots ( Figure 3 750 g of phosphate fertilizer was applied as basal fertilizer in each plot, and 295 g of urea was applied as topdressing in each plot during the tillering stage.

[0083] The process of biomass and nitrogen content determination was to take samples at 9 times, including 2021.12.29 (S1-1), 2022.1.16 (S1-2), 2022.2.9 (S1-3), 2022.3.17 (S1-4), 2022.4.6 (S1-5), 2022.12.30 (S2-1), 2023.1.10 (S2-2), 2023.2.9 (S2-3), 2023.2.28 (S2-4), 2023.3.24 (S2-5), and 2023.4.10 (S2-6), and use drone aerial images to randomly cut 100 m3 of each plot. cm sample segment, put the sample into a blast drying oven at 105℃ for 30 minutes, dry it at 65℃ for 20 hours, and weigh it separately to calculate the yield per unit area, i.e. biomass. The sample dried to constant weight was crushed, passed through a 1 mm sieve, and wrapped in tin foil. The fixed nitrogen content, i.e. nitrogen content, was determined using a Dumas combustion nitrogen analyzer (NDA701, VELP, Italy).

[0084] Multispectral data were collected by drones equipped with multispectral cameras. The drones flew at an altitude of 50 m, with a ground spatial resolution of 2.2 cm and 85% longitudinal overlap and 85% lateral overlap between images. Clear and cloudless weather was selected to shoot drone images. The shooting time was generally from 10 to 12 o'clock on the same day, and automatic exposure was used during shooting. The bands used in this study were blue light (450 nm), green light (560 nm), red light (650 nm), red edge (730 nm) and near infrared (840 nm). The calibrated calibration blanket was placed within the flight range before shooting. ArcGIS 10.8.1 was used for image conversion format and reflectance extraction software, and image reconstruction, geometric correction and image stitching were performed in DJI Terra 3.6.1.

[0085] The vegetation index of different bands was calculated by DJI Terra 3.6.1 to generate the test plot image. The test plot image was imported into ArcGIS 10.8.1 and the area of ​​interest of each experimental plot was circled. The vegetation index of each band in the plot was calculated according to the average reflectivity in each area of ​​interest.

[0086] PyCharm was used to calculate the texture index of vegetation index in different bands using the symbiotic grayscale matrix and generate the experimental field image. The generated experimental field image with texture index was imported into ArcGIS 10.8.1 and the area of ​​interest of each experimental field plot was circled. The texture index under each vegetation index in the field was calculated based on the average reflectance in each area of ​​interest.

[0087] Further, in step S108, based on the general linear regression algorithm, a first estimation model for estimating biomass by vegetation index and texture index, and a second estimation model for estimating nitrogen content by vegetation index and texture index are constructed. In this embodiment, the specific implementation method is as follows:

[0088] Based on the general linear regression algorithm, the leave-one-out cross-validation method is used to construct the first estimation model for estimating the biomass through the vegetation index and the texture index; based on the general linear regression algorithm, the leave-one-out cross-validation method is used to construct the second estimation model for estimating the nitrogen content through the vegetation index and the texture index.

[0089] Further, in step S110, based on the partial least squares regression algorithm, a third estimation model for estimating biomass by vegetation index and texture index, and a fourth estimation model for estimating nitrogen content by vegetation index and texture index are constructed. In this embodiment, the specific implementation method is as follows:

[0090] Based on the partial least squares regression algorithm, the leave-one-out cross-validation method is used to construct the third estimation model for estimating the biomass through the vegetation index and the texture index; based on the partial least squares regression algorithm, the leave-one-out cross-validation method is used to construct the fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index.

[0091] Furthermore, in step S112, the process of selecting the target estimation model from the first estimation model, the second estimation model, the third estimation model and the fourth estimation model according to the preset screening principle is specifically implemented as follows in this embodiment:

[0092] Determine the Pearson correlation coefficient associated with the first estimation model, the second estimation model, the third estimation model and the fourth estimation model; determine the root mean square error associated with the first estimation model, the second estimation model, the third estimation model and the fourth estimation model; determine the target estimation model based on the Pearson correlation coefficient and the root mean square error.

[0093] Among them, two modeling methods were used to compare the optimal estimation model of dry matter yield and nitrogen fixation in mixed grassland. First, the vegetation index and texture index with good correlation with field data were screened by general linear regression, and then the variable projection importance index of partial least squares regression modeling was used to screen factors for secondary modeling. Both models were modeled using the leave-one-out cross-validation method to determine the optimal number of factors. The Pearson correlation coefficient (R 2) was used to evaluate the correlation between biomass, nitrogen content and vegetation index and texture index. The root mean square error (RMSE) was used to evaluate the estimation accuracy of different models. The accuracy and error of the two models, partial least squares regression modeling and general linear regression modeling, were compared. Based on the results, the model with better accuracy was selected to estimate dry matter yield and nitrogen fixation.

[0094] Using the above example, there were 11 observations from 2021 to 2023, of which 8 observations (S1-1, S1-3, S1-4, S1-5, S2-1, S2-2, S2-3, S2-5) were used for general linear regression and partial least squares regression training models, and the remaining 3 observations (S1-2, S2-4, S2-6) were used as independent data sets for partial least squares regression model validation. The variable projection importance index (VIP) was calculated using the statistical analysis software SIMCA-P 11.5, and VIP>1 was used as the threshold for screening key vegetation indices. General linear regression and partial least squares regression analysis were performed using R 4.3.0.

[0095] Specifically, for the first estimation model, according to actual measurements, through R 2 The RMSE and RMS represent each simple vegetation index, and the biomass is estimated under general linear regression. The accuracy of the estimated value is also expressed in R 2 The combined model of vegetation index and texture index was characterized by RMSE, and the biomass was estimated under general linear regression. The accuracy of the estimated values ​​was obtained. The results showed that the optimal estimation result of the combined model of vegetation index and texture index was more accurate than the optimal estimation result related to the single vegetation index.

[0096] For the third estimation model, 18 factors were selected to model the biomass yield of mixed grasses under PLSR based on the modeling results of general linear regression. First, according to the results of general linear regression and combined with the vip threshold (vip>1), the PLSR model based on the simple vegetation index was selected, among which the vip values ​​of NDRE, LCI, and NDVI were greater than 1. After adding the texture index to NDRE, LCI, and NDVI, the vip values ​​of 5 independent variables were greater than 1, namely GNDVI, LCI, NDRE, NDRE_cont, and LCI_cont.

[0097] The PLSR model based on NDRE, LCI and NDVI provides an estimate of mixed grass biomass yield. 2As shown in the RMSE, among all vegetation indices, NDRE contributes the most to the prediction of mixed grass biomass using the PLSR model, and the PLSR model based on vegetation index and texture characteristics provides a more accurate estimation of mixed grass biomass than the general linear model based on vegetation index and texture characteristics.

[0098] In order to evaluate the contribution of texture index to the estimation of mixed grass biomass, the R values ​​of the training and validation datasets were calculated. 2 Compared with RMSE, the vegetation index and texture index were used as inputs of the PLSR model. Compared with the PLSR model based only on the vegetation index, the PLSR model with added texture features did not significantly improve the estimation accuracy of mixed grass biomass. From the vip values ​​of mixed grass, it can be seen that GNDVI, NDRE, and LCI are ranked in the top three in the ranking based solely on the vegetation index and in the comprehensive ranking of the vegetation index plus the texture index. This also shows that the vegetation index alone has a relatively ideal effect on predicting the biomass yield of mixed grasses. Further, compared with the PLSR model with the texture index added, the R of the validation data set was 2 The improvement was 216.7% and the RMSE value was reduced by 23.1%.

[0099] For the second estimation model, according to the actual measurement, the R 2 The RMSE and RMS characterize each simple vegetation index, and the nitrogen content is estimated under the general linear regression. The accuracy of the estimated value is also expressed in R 2 The combined model of vegetation index and texture index is characterized by RMSE, and the nitrogen content is estimated under general linear regression. The accuracy of the estimated value is obtained. The results show that the optimal estimation result related to the simple vegetation index is more accurate than the combined model of vegetation index and texture index.

[0100] For the fourth estimation model, 17 factors were selected to model the N concentration of mixed grass under PLSR based on the modeling results of general linear regression. First, according to the results of general linear regression and combined with the vip threshold (vip>1), the PLSR model based on the simple vegetation index was selected, in which the vip values ​​of NDRE and LCI were greater than 1. After adding texture features to NDRE and LCI, the vip values ​​of 6 independent variables were greater than 1, namely NDRE, LCI, LCI_mean, GNDVI, NDRE_mean, and NDRE_cont.

[0101] The NDRE and LCI models provide estimates of the nitrogen concentration of mixed grasses. 2From the RMSE, we can see that among all vegetation indices, NDRE contributes the most to the prediction of N concentration of mixed grass using PLSR model. 2 Compared with the PLSR model based on the simple vegetation index, the PLSR model with texture features did not significantly improve the estimation accuracy of the N concentration of mixed grass. From the vip value of mixed grass, it can be seen that NDRE and LCI are ranked in the top two in the ranking based on the vegetation index alone and in the comprehensive ranking of the vegetation index plus the texture index. This also shows that the simple vegetation index has a relatively ideal effect on predicting the N concentration of mixed grass. Further, compared with the PLSR model based on the simple vegetation index and the PLSR model with the texture index, the R of the validation data set is 2 The improvement was 12.3% and the RMSE value was reduced by 6%.

[0102] Corresponding to the above method embodiment, the present application also provides an embodiment of a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology. Figure 2 FIG. 1 is a schematic diagram showing a structure of a device for estimating the growth of a mixed crop of grass and beans based on multispectral technology provided by an embodiment of the present application. Figure 2 As shown, the device comprises:

[0103] The acquisition module 202 is configured to acquire multispectral data of a mixed-crop area, wherein at least two crops are mixed in the mixed-crop area;

[0104] A calculation module 204 is configured to calculate a vegetation index according to the multispectral data, and calculate a texture index associated with the vegetation index;

[0105] A determination module 206, configured to determine the biomass and nitrogen content of the crops in the mixed-crop area;

[0106] The first construction module 208 is configured to construct a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a general linear regression algorithm;

[0107] The second construction module 210 is configured to construct a third estimation model for estimating the biomass through the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a partial least squares regression algorithm;

[0108] The screening module 212 is configured to select a target estimation model from among the first estimation model, the second estimation model, the third estimation model and the fourth estimation model according to a preset screening principle, and estimate the growth of crops in the mixed-crop area according to the target estimation model.

[0109] In an optional embodiment, the acquisition module 202 is further configured to:

[0110] Based on an image acquisition device including a multispectral camera, multispectral data of the mixed-seed area is collected, wherein the collection bands of the multispectral data include blue light, green light, red light, red edge and near infrared.

[0111] In an optional embodiment, the calculation module 204 is further configured to:

[0112] The vegetation index including the normalized difference vegetation index, the normalized difference red edge vegetation index, the green normalized difference vegetation index, the leaf chlorophyll index and the optimized soil adjusted vegetation index is calculated according to the multispectral data.

[0113] In an optional embodiment, the calculation module 204 is further configured to:

[0114] Based on the symbiotic grayscale matrix, the initial texture index associated with the vegetation index in different bands is calculated, and a hybrid area image is generated according to the calculation results, wherein the texture index corresponding to any block in the hybrid area image is determined by the average reflectivity of the block, and the texture index includes mean, contrast, difference, entropy, energy, homogeneity, and second-order moment.

[0115] In an optional embodiment, the first building module 208 is further configured to:

[0116] Based on the general linear regression algorithm, the leave-one-out cross-validation method is used to construct the first estimation model for estimating the biomass through the vegetation index and the texture index; based on the general linear regression algorithm, the leave-one-out cross-validation method is used to construct the second estimation model for estimating the nitrogen content through the vegetation index and the texture index.

[0117] In an optional embodiment, the second building module 210 is further configured to:

[0118] Based on the partial least squares regression algorithm, the leave-one-out cross-validation method is used to construct the third estimation model for estimating the biomass through the vegetation index and the texture index; based on the partial least squares regression algorithm, the leave-one-out cross-validation method is used to construct the fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index.

[0119] In an optional embodiment, the screening module 212 is further configured to:

[0120] Determine the Pearson correlation coefficient associated with the first estimation model, the second estimation model, the third estimation model and the fourth estimation model; determine the root mean square error associated with the first estimation model, the second estimation model, the third estimation model and the fourth estimation model; determine the target estimation model based on the Pearson correlation coefficient and the root mean square error.

[0121] The multispectral technology-based grain-bean mixed crop growth estimation device provided by the present application collects multispectral data of a mixed area, wherein at least two crops are mixed in the mixed area; calculates a vegetation index according to the multispectral data, and calculates a texture index associated with the vegetation index; determines the biomass and nitrogen content of crops in the mixed area; constructs a first estimation model for estimating the biomass through the vegetation index and the texture index based on a general linear regression algorithm, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index; constructs a third estimation model for estimating the biomass through the vegetation index and the texture index based on a partial least squares regression algorithm, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index; selects a target estimation model from the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model through a preset screening principle, and estimates the growth of crops in the mixed area according to the target estimation model. A non-destructive, efficient, and accurate estimation of the biomass and nitrogen content of the grain-bean mixed crop is achieved.

[0122] The above is a schematic scheme of a kind of grain-bean hybrid crop growth estimation device based on multispectral technology of this embodiment. It should be noted that the technical scheme of the grain-bean hybrid crop growth estimation device based on multispectral technology and the technical scheme of the grain-bean hybrid crop growth estimation method based on multispectral technology belong to the same concept, and the details of the technical scheme of the grain-bean hybrid crop growth estimation device based on multispectral technology that are not described in detail can all refer to the description of the technical scheme of the grain-bean hybrid crop growth estimation method based on multispectral technology. In addition, each component in the device embodiment should be understood as a functional module that must be established to implement each step of the program flow or each step of the method, and each functional module is not an actual functional segmentation or separation limitation. The device claim defined by such a group of functional modules should be understood as a functional module architecture that mainly implements the solution through a computer program recorded in the specification, and should not be understood as a physical device that mainly implements the solution through hardware.

[0123] Figure 3The block diagram of a computing device 300 provided according to an embodiment of the present application is shown. The components of the computing device 300 include but are not limited to a memory 310 and a processor 320. The processor 320 is connected to the memory 310 via a bus 330, and the database 350 is used to store data.

[0124] The computing device 300 also includes an access device 340 that enables the computing device 300 to communicate via one or more networks 360. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 340 may include one or more of any type of network interface (e.g., a network interface card (NIC)) whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a World Wide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.

[0125] In one embodiment of the present application, the above components of the computing device 300 and Figure 3 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 3 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of the present application. Those skilled in the art may add or replace other components as needed.

[0126] The computing device 300 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. The computing device 300 may also be a mobile or stationary server.

[0127] The processor 320 is used to execute computer executable instructions for each step of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology.

[0128] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above-mentioned method for estimating the growth of a mixed crop of grass and beans based on multispectral technology belong to the same concept, and the details not described in detail in the technical scheme of the computing device can be found in the description of the technical scheme of the above-mentioned method for estimating the growth of a mixed crop of grass and beans based on multispectral technology.

[0129] An embodiment of the present application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, are used to execute each step of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology.

[0130] The above is a schematic scheme of a computer-readable storage medium of the present embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the above-mentioned method for estimating the growth of a mixed crop of grass and beans based on multispectral technology belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be found in the description of the technical scheme of the above-mentioned method for estimating the growth of a mixed crop of grass and beans based on multispectral technology.

[0131] An embodiment of the present application further provides a chip storing a computer program, which, when executed by the chip, implements the steps of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology.

[0132] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0134] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0135] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0136] The preferred embodiments of the present application disclosed above are only used to help explain the present application. The optional embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the present application. The present application selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is only limited by the claims and their full scope and equivalents.

Claims

1. A method for estimating the growth of a mixed crop of grass and beans based on multispectral technology, characterized in that: include: Collecting multispectral data of a mixed-crop area, wherein at least two crops are mixed in the mixed-crop area; According to the multispectral data, a vegetation index is calculated, and a texture index associated with the vegetation index is calculated. Specifically, according to the multispectral data, the vegetation index including the normalized vegetation index, the normalized difference red edge vegetation index, the green normalized difference vegetation index, the leaf chlorophyll index and the optimized soil adjustment vegetation index is calculated. Based on the symbiotic grayscale matrix, the initial texture index associated with the vegetation index in different bands is calculated, and a mixed-seed area image is generated according to the calculation result, wherein the texture index corresponding to any block in the mixed-seed area image is determined by the average reflectivity of the block, and the texture index includes mean, contrast, difference, entropy, energy, homogeneity and second-order moment; Determining the biomass and nitrogen content of crops in the intercropping area; Based on a general linear regression algorithm, a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index are constructed; Based on a partial least squares regression algorithm, constructing a third estimation model for estimating the biomass through the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index; According to a preset screening principle, a target estimation model is selected from the first estimation model, the second estimation model, the third estimation model and the fourth estimation model, and the growth conditions of the crops in the mixed-cropping area are estimated according to the target estimation model.

2. The method for estimating the growth of a mixed crop of grass and beans based on multispectral technology according to claim 1, characterized in that: The method of collecting multispectral data of the mixed-breeding area includes: Based on an image acquisition device including a multispectral camera, multispectral data of the mixed-seed area is collected, wherein the collection bands of the multispectral data include blue light, green light, red light, red edge and near infrared.

3. The method for estimating the growth of a mixed crop of grass and beans based on multispectral technology according to claim 1, characterized in that: The method of constructing a first estimation model for estimating the biomass by using the vegetation index and the texture index and a second estimation model for estimating the nitrogen content by using the vegetation index and the texture index based on a general linear regression algorithm comprises: Based on a general linear regression algorithm, a leave-one-out cross validation method is used to construct the first estimation model for estimating the biomass through the vegetation index and the texture index; Based on a general linear regression algorithm and using a leave-one-out cross validation method, the second estimation model for estimating the nitrogen content through the vegetation index and the texture index is constructed.

4. The method for estimating the growth of a mixed crop of grass and beans based on multispectral technology according to claim 1, characterized in that: The method of constructing a third estimation model for estimating the biomass by using the vegetation index and the texture index and a fourth estimation model for estimating the nitrogen content by using the vegetation index and the texture index based on the partial least squares regression algorithm comprises: Based on the partial least squares regression algorithm, a leave-one-out cross validation method is adopted to construct the third estimation model for estimating the biomass through the vegetation index and the texture index; Based on the partial least squares regression algorithm and using the leave-one-out cross validation method, the fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index is constructed.

5. The method for estimating the growth of a mixed crop of grass and beans based on multispectral technology according to claim 1, characterized in that: The selecting a target estimation model from the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model according to a preset screening principle includes: determining a Pearson correlation coefficient associated with the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model; determining a root mean square error associated with the first estimation model, the second estimation model, the third estimation model, and the fourth estimation model; A target estimation model is determined according to the Pearson correlation coefficient and the root mean square error.

6. A device for estimating the growth of a mixed crop of grass and beans based on multispectral technology, characterized in that: include: A collection module is configured to collect multispectral data of a mixed-crop area, wherein at least two crops are mixed in the mixed-crop area; A calculation module is configured to calculate a vegetation index according to the multispectral data, and calculate a texture index associated with the vegetation index. Specifically, according to the multispectral data, the vegetation index including a normalized vegetation index, a normalized difference red edge vegetation index, a green normalized difference vegetation index, a leaf chlorophyll index and an optimized soil conditioned vegetation index is calculated. Based on a symbiotic grayscale matrix, an initial texture index associated with the vegetation index in different bands is calculated, and a hybrid area image is generated according to the calculation result, wherein the texture index corresponding to any block in the hybrid area image is determined by the average reflectivity of the block, and the texture index includes mean, contrast, difference, entropy, energy, homogeneity, and second-order moment; A determination module configured to determine the biomass and nitrogen content of the crops in the mixed-cropping area; A first construction module is configured to construct a first estimation model for estimating the biomass through the vegetation index and the texture index, and a second estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a general linear regression algorithm; A second construction module is configured to construct a third estimation model for estimating the biomass through the vegetation index and the texture index, and a fourth estimation model for estimating the nitrogen content through the vegetation index and the texture index based on a partial least squares regression algorithm; The screening module is configured to select a target estimation model from among the first estimation model, the second estimation model, the third estimation model and the fourth estimation model according to a preset screening principle, and estimate the growth of crops in the mixed-crop area according to the target estimation model.

7. A computing device, characterized in that include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the steps of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing computer instructions, characterized in that: When the instruction is executed by the processor, the steps of the method for estimating the growth of a mixed crop of grass and beans based on multi-spectral technology as described in any one of claims 1 to 5 are implemented.

Images