A Winter Rapeseed Phenology Monitoring Method Based on an Improved Shape Model

Abstract

The present invention discloses a method for monitoring the phenology of winter rapeseed based on an improved shape model, which includes: collecting the observation data of winter rapeseed and establishing a dynamic growth curve of winter rapeseed; selecting vegetation indices and establishing a shape model; using the shape model for phenology detection to obtain the phenological dynamics of winter rapeseed. The method of the present invention greatly saves the input of human and material costs, and further promotes the development of agricultural research and the transformation of achievements. Obtaining the crop phenological dynamics in a timely manner can provide a decision-making basis for the formulation of irrigation and fertilization systems, and while increasing crop yields, it can also provide important parameter support for the simulation of farmland ecological systems, which is of great significance in the context of the continuously increasing global food demand.

Description

Technical Field

[0001] This invention relates to the field of agricultural engineering technology, specifically to a method for monitoring the phenology of winter rapeseed based on an improved shape model. Background Technology

[0002] Crop models, as process-oriented dynamic models, systematically simulate the continuous growth process of crops using mechanistic and semi-mechanistic mathematical methods, incorporating the influence of external factors (meteorology, farmland management). Through crop model simulation, the current developmental status of crops can be accurately determined, and their future growth dynamics and final yield can be predicted. They have been widely applied in agricultural forecasting, climate change impact assessment, agricultural decision-making, and optimization of cultivation patterns.

[0003] There are two main categories of research methods for phenological monitoring both domestically and internationally: manual monitoring and remote sensing monitoring. Traditional phenological monitoring methods rely on manual observation to record vegetation phenological dynamics. While this method is easy to implement and yields high-precision data, it requires significant human and material resources, making it difficult to implement on a large regional scale. The main advantage of remote sensing observation lies in the use of sensors with different revisit periods and spatial resolutions, which greatly improves the efficiency and accuracy of capturing changes in vegetation phenological dynamics. Currently, vegetation phenological remote sensing monitoring has become one of the hot topics in phenological research in recent years. However, most existing research on plant phenological remote sensing monitoring is based on specific phenological stages of vegetation (such as the beginning and end of the growing season, and flowering). In practical applications, especially in agricultural production, the need for timely acquisition of phenological information for multiple stages throughout the entire growth period necessitates further improvement of existing phenological remote sensing monitoring methods. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by this invention is that the current method of establishing shape models based on remote sensing monitoring for the analysis and prediction of winter rapeseed phenology requires a large number of parameters for calibration. However, parameter calibration is often based on experimental data at a single point scale. When the model is extended to regional applications, crop parameters may change with the expansion of the scale.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for monitoring the phenology of winter rapeseed based on an improved shape model, comprising:

[0008] Collect observational data on winter rapeseed and establish a dynamic growth curve for winter rapeseed;

[0009] Select vegetation indices and establish a shape model;

[0010] Phenological dynamics of winter rapeseed were obtained by using shape models for phenological detection.

[0011] As a preferred embodiment of the winter rapeseed phenological monitoring method based on the improved shape model described in this invention, the observation data includes phenological information of winter rapeseed, state variables during the growth and development process, and remote sensing images of canopy reflectance at different growth stages of winter rapeseed.

[0012] As a preferred embodiment of the winter rapeseed phenological monitoring method based on an improved shape model described in this invention, the vegetation index is an indicator used to measure the growth and abundance of vegetation, including the red-edged chlorophyll index.

[0013]

[0014] Where CI represents the red-edged chlorophyll index; R NIR Indicates the near-infrared spectral band; R R-E Indicates the red-edge spectral band;

[0015] Normalized Difference Vegetation Index:

[0016]

[0017] Wherein, NDVI represents the Normalized Difference Vegetation Index; R RED Indicates the red spectral band;

[0018] Enhanced vegetation index:

[0019]

[0020] Where EVI represents the enhanced vegetation index; α is a constant; β is a constant; R BLUE It indicates the blue spectrum band.

[0021] As a preferred embodiment of the winter rapeseed phenological monitoring method based on the improved shape model described in this invention, the method includes: establishing the shape model by selecting an objective function based on the characteristics of the crop growth curve, matching the predefined phenological stage X0 to the model curve; and using the optimal parameter equation to perform feature matching between the input observation data and the shape model, adjusting the observation data through the optimal parameters during the matching process.

[0022] As a preferred embodiment of the winter rapeseed phenological monitoring method based on the improved shape model described in this invention, the optimal parameter equation is expressed as:

[0023] g(x) = yscale × h(xscale × x0 + tshift)

[0024] Where, g(x) represents the optimized parameter equation; yscale represents the y-axis scale; h is the established shape model; xscale represents the x-axis scale, 0.5 < xscale < 1.6; x0 represents the predefined crop phenological days; tshift represents the offset on the time axis, -70 < tshift < 70;

[0025] Adjust the observed data, which is expressed as:

[0026] g1(x) = yscale1 × h1(xscale × x0 + tshift) + yscale2 × h2(xscale × x0 + tshift)

[0027] Where, g1(x) represents the optimized parameter equation; yscale1 represents the y-axis scale 1, 0.1 < yscale1 < 1.5; h1 represents the first part of the model; yscale2 represents the y-axis scale 2, 0.1 < yscale2 < 1.5; h2 represents the second part of the model.

[0028] As a preferred scheme of the winter rapeseed phenology monitoring method based on the improved shape model described in the present invention, where: establishing the shape model further includes: selecting two functions of the asymmetric Gaussian function and the Fourier series function to establish the model, and comparing with the double logistic function;

[0029] The asymmetric Gaussian function is:

[0030]

[0031] Where, AGF VIs represents the asymmetric Gaussian function; j is a fitting parameter that determines the amplitude of the model curve; k is a fitting parameter that determines the amplitude of the model curve; e is a constant; DAS represents the number of days after sowing; f represents the center position of the first peak of the curve, h represents the center position of the second peak of the curve; g represents the width of the first peak; i represents the width of the second peak;

[0032] The Fourier series function is:

[0033]

[0034] Where, FF VIs represents the Fourier series function; l is a fitting parameter; m is a fitting parameter; n is a fitting parameter; i represents the vegetation index value of the bare soil background; p is the fundamental frequency of the signal;

[0035] The dual logic function is represented as follows:

[0036]

[0037] Among them, VI max denoted by , where represents the maximum value of the vegetation index; 'a' represents the maximum curvature when the curve is rising, and 'c' represents the maximum curvature when the curve is falling; 'b' is a fitting parameter related to the curvature of the vegetation index curve; and 'd' is a fitting parameter related to the curvature of the vegetation index curve.

[0038] As a preferred embodiment of the winter rapeseed phenological monitoring method based on the improved shape model described in this invention, the phenological detection using the shape model is expressed as follows:

[0039] X est = xscale × x0 + tshift

[0040] Among them, X est This represents the number of phenological days estimated by the model.

[0041] Secondly, the present invention also provides a device for monitoring the phenology of winter rapeseed based on an improved shape model, including a data acquisition system for acquiring phenological information of winter rapeseed, state variables during the growth and development process, and remote sensing images of canopy reflectance at different growth stages of winter rapeseed, and uploading them to a data processing module.

[0042] The data processing module establishes a dynamic growth curve for winter rapeseed, selects vegetation indices, builds a shape model, and transmits it to the phenological detection module.

[0043] The phenology detection module performs phenological detection based on the established shape model to obtain the phenological dynamics of winter rapeseed.

[0044] Thirdly, the present invention also provides a computing device, including: a memory and a processor;

[0045] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the winter rapeseed phenological monitoring method based on the improved shape model.

[0046] Fourthly, the present invention also provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the method for monitoring winter rapeseed phenology based on an improved shape model.

[0047] The beneficial effects of this invention are as follows: The method of this invention greatly saves on the input of manpower and material resources, further promoting the development and transformation of agricultural research results. Timely acquisition of crop phenological dynamics can provide a basis for decision-making in the formulation of irrigation and fertilization systems, increasing crop yields while also providing important parameter support for farmland ecosystem simulation, which is of great significance in the context of ever-increasing global food demand. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0049] Figure 1 The overall flowchart of a winter rapeseed phenological monitoring method based on an improved shape model is provided in one embodiment of the present invention;

[0050] Figure 2 A schematic diagram of the shape model before and after improvement of a winter rapeseed phenological monitoring method based on an improved shape model, provided as an embodiment of the present invention;

[0051] Figure 3 A schematic diagram of a winter rapeseed phenological monitoring method based on an improved shape model provided in the second embodiment of the present invention;

[0052] Figure 4 A schematic diagram of phenological monitoring for a winter rapeseed phenological monitoring method based on an improved shape model, provided for the second embodiment of the present invention;

[0053] Figure 5 The second embodiment of the present invention provides a winter rapeseed phenological monitoring method based on an improved shape model, with winter rapeseed growth curves established based on three VIs;

[0054] Figure 6 The second embodiment of the present invention provides a method for monitoring the phenology of winter rapeseed based on an improved shape model, which is based on a combination of three mathematical functions and the VI index.

[0055] Figure 7 The second embodiment of the present invention provides a comparison of the accuracy of target curve matching under different sowing scenarios using an improved shape model for monitoring winter rapeseed phenology. Detailed Implementation

[0056] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0057] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0059] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0060] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0061] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0062] Example 1

[0063] Reference Figure 1 and 2As an embodiment of the present invention, a method for monitoring the phenology of winter rapeseed based on an improved shape model is provided, comprising:

[0064] S1: Collect observation data of winter rapeseed and establish a dynamic growth curve of winter rapeseed.

[0065] Furthermore, accurately acquiring crop phenological information and state variables during crop growth and development is the prerequisite and foundation for accurate yield estimation by the entire assimilation system. This study established dynamic growth curves for winter rapeseed by collecting remote sensing images of canopy reflectance at different growth stages. Finally, based on an improved shape model method, phenological information at multiple stages within the winter rapeseed growth period was extracted and identified. The study constructed multiple rapeseed phenological monitoring models by combining different vegetation indices and functions, evaluated the ability of different vegetation index and function fitting methods to monitor and invert phenological stages, and determined the optimal monitoring model for crop phenological stages, providing data accuracy assurance for model assimilation.

[0066] It should be noted that a dynamic growth curve is a mathematical model that describes the growth process of an individual organism or population over a certain period of time. It typically represents time on the horizontal axis and growth indicators such as individual mass, quantity, or volume on the vertical axis. It can be used to study growth rate, growth trend, and growth patterns.

[0067] Dynamic growth curves come in various forms, among which the most common are the S-shaped curve, the exponential curve, and the linear curve. The S-shaped curve is one of the most common models for describing the growth of an organism, showing a trend of gradually accelerating growth, reaching a peak, and then gradually slowing down until stabilizing. The exponential curve shows that the growth rate of the organism remains at a relatively constant level, exhibiting a rapid growth trend. The linear curve shows that the growth rate of the organism remains at a constant rate, exhibiting a uniform growth trend.

[0068] S2: Select vegetation indices and establish shape models.

[0069] Furthermore, vegetation indices were selected. The absorption spectra of different vegetation types vary due to differences in canopy structure, pigment concentration, and leaf water content; therefore, a single spectral band cannot fully characterize the phenological features of vegetation. It should be noted that:

[0070] Studies have shown that calculating indices using combinations of multiple vegetation-sensitive spectral bands is more effective in representing vegetation phenological characteristics than single-band spectral reflectance. Examples include the NDVI, constructed based on the strong absorption of red light (650 nm) by plant leaf tissues and the strong reflectance of near-infrared light (700-100 nm) by cells on the underside of leaves; and the normalized water index, constructed using the near-infrared (860 nm) and short-wave infrared (1240 nm) bands, which are most sensitive to changes in leaf water content. In particular, the NDVI has been widely used since its inception for estimating vegetation parameters such as phenological stages and leaf area index (LAI). With the upgrading of sensors, the number of spectral bands that can be included in vegetation phenology research is constantly increasing. To address the problems of traditional vegetation indices being easily saturated in high-biomass areas and susceptible to interference from soil background, snow, and atmospheric noise, researchers have introduced more spectral bands to construct VIs to overcome these shortcomings, such as Enhanced Vegetation Index (EVI), 2-Bands Enhanced Vegetation Index (EVI2), and Red-edge Chlorophyll Index (CIred-edge). To compare the explanatory power of different VIs for the phenological characteristics of winter rapeseed and their adaptability to phenological model building when constructing curves, CIred-edge, NDVI, and EVI were selected for comparative discussion based on existing research.

[0071]

[0072] Table 1. Relevant vegetation indices for model establishment.

[0073] Furthermore, the shape model, proposed by Sakamoto et al., is a method for obtaining crop phenological information based on function transformation and feature matching. First, a target function h(x) is selected and a model is established based on the characteristics of the crop growth curve. A predefined phenological stage X0 is matched to the model curve. The optimal parameter equation (Equation 1-1) is used to perform feature matching between the input observation data and the shape model. During the matching process, the observation data is adjusted using optimal parameters (yscale, xscale, and tshift). After the optimal parameters are determined, the target crop phenological stage is estimated according to Equation 1-3. Figure 2a). This method enables the scope of phenological monitoring not to be limited to curve characteristic values (such as inflection points, extreme points, etc.), and at the same time avoids the problem of difficulty in extracting phenological information of vegetation with small seasonal variations and relatively gentle temporal VI curves. Sakamoto et al. used this method to monitor four phenological periods of corn and soybeans respectively, and the RMSE of the estimated phenological days was 2.4 - 7.4 days, proving the potential of the shape model in extracting vegetation phenological information. Different from the optimal parameter equation with three parameters proposed by Sakamoto and Zenget al., this invention adds a parameter in the y direction and also separates h(x) (Equation 1 - 2). In addition to the target curve with a single peak, the improved model can adapt to more complex curve shapes. Figure 2 b). Since the vegetation temporal curve is matched by stretching and shrinking the geometric scale of the model, the local fluctuations caused by noise values in the growth curve can be minimized. However, when the parameters of the matching model are used for different research areas and vegetation types, ground phenological observation data are required for calibration. Therefore, the accuracy and precision of ground observations and the sample size of observations (i.e., the number of years and locations) directly affect the accuracy of using the shape model.

[0074] g(x) = yscale × h(xscale × x0 + tshift) (1 - 1)

[0075] Among them, g(x) represents the optimized parameter equation; yscale represents the y - axis scale; h is the established shape model; xscale represents the x - axis scale, 0.5 < xscale < 1.6; x0 represents the predefined crop phenological days; tshift represents the offset on the time axis, - 70 < tshift < 70.

[0076] Adjust the observed data, expressed as:

[0077] g1(x) = yscale1 × h1(xscale × x0 + tshift) + yscale2 × h2(xscale × x0 + tshift) (1 - 2)

[0078] Among them, g1(x) represents the optimized parameter equation; yscale1 represents the y - axis scale 1, 0.1 < yscale1 < 1.5; h1 represents the first part of the model; yscale2 represents the y - axis scale 2, 0.1 < yscale2 < 1.5; h2 represents the second part of the model.

[0079] X est = xscale × x0 + tshift (1 - 3)

[0080] Among them, X estThis represents the number of phenological days estimated by the model.

[0081] Furthermore, unlike the single-peak VI curves of crops such as wheat, corn, and soybean, the phenological profile of winter rapeseed is complex, with multiple peaks throughout its growth period. This makes remote sensing phenological extraction methods such as thresholding difficult to apply. Meanwhile, the flowering period of rapeseed exhibits easily identifiable characteristics on the established VI curve; as rapeseed enters the flowering stage, the VI continuously decreases, reaching its minimum value at full bloom. Therefore, all reported studies on rapeseed phenological stages have used this significant characteristic to monitor the flowering period. This study focuses on the seedling-flowering stage of winter rapeseed, aiming to dynamically identify and monitor multiple phenological stages within this stage. Based on the characteristics of the target VI curve in this stage, two functions, the asymmetric Gaussian function (AGF) and the Fourier series function (FF), were selected for model establishment. These were compared with the double logistic function (DLF), which is most widely used in shape model establishment. The formulas for the three function models are as follows:

[0082] The asymmetric Gaussian function is:

[0083]

[0084] Among them, AGF VIs denoted as an asymmetric Gaussian function; j is the fitting parameter that determines the amplitude of the model curve; k is the fitting parameter that determines the amplitude of the model curve; e is a constant; DAS represents the number of days after sowing; f represents the center position of the first peak of the curve; h represents the center position of the second peak of the curve; g represents the width of the first peak; i represents the width of the second peak.

[0085] The Fourier series function is:

[0086]

[0087] Among them, FF VIs The function represents the Fourier series function; l is the fitting parameter; m is the fitting parameter; n is the fitting parameter; i represents the vegetation index value of the bare soil background; p is the fundamental frequency of the signal;

[0088] The two logic functions are represented as follows:

[0089]

[0090] Among them, VI maxdenoted by , where represents the maximum value of the vegetation index; 'a' represents the maximum curvature when the curve is rising, and 'c' represents the maximum curvature when the curve is falling; 'b' is a fitting parameter related to the curvature of the vegetation index curve; and 'd' is a fitting parameter related to the curvature of the vegetation index curve.

[0091] S3: Use shape models to detect phenology and obtain the phenological dynamics of winter rapeseed.

[0092] Furthermore, to ensure the representativeness of the shape model, the ES sowing period, with the most frequent observations, was used as the base dataset for constructing the shape model. Observational data from 9 plots during the seedling-flowering stage were used to construct the rapeseed shape model. The remaining 27 plots from the other 3 sowing periods were used for model evaluation.

[0093] It should be noted that phenological dynamics refer to the periodic physiological phenomena that occur in plants and animals with seasonal changes, mainly including flowering, leaf sprouting, fruit ripening, and migration. These biological phenomena are closely related to environmental factors such as temperature, light, and rainfall, and exhibit a certain degree of adaptability.

[0094] Phenological dynamics are crucial for the stable functioning of ecosystems and the adaptability of species. For example, the flowering and pollination periods of plants are closely related to the emergence and activity of insects, and their interactions affect plant reproduction and insect food sources; animal migration is related to seasonal changes in food supply, and animals need to adjust their activity range and dietary habits according to seasonal changes.

[0095] This embodiment also provides a computing device, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the winter rapeseed phenological monitoring method based on the improved shape model proposed in the above embodiment.

[0096] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the winter rapeseed phenological monitoring method based on an improved shape model as proposed in the above embodiments.

[0097] The storage medium proposed in this embodiment belongs to the same inventive concept as the winter rapeseed phenological monitoring method based on the improved shape model proposed in the above embodiment. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0098] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0099] Example 2

[0100] Reference Figure 3-7 As an embodiment of the present invention, a method for monitoring the phenology of winter rapeseed based on an improved shape model is provided. To verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0101] like Figure 5 As shown, the three VIs curves exhibited significant dynamic differences under four sowing scenarios. With delayed sowing (approximately 15 days between sowings), the entire growth period of rapeseed shortened from 230 days to 185 days, and the number of days to reach flowering decreased from 175 days to 140 days. Taking NDVI as an example, the growth dynamics of rapeseed throughout its entire growth period are described as follows: For ES sowing, the NDVI value increased rapidly during the winter rapeseed seedling stage, reaching a peak at DAS30 and then declining until spring temperatures rose, ending the overwintering period (around DAS120). With the end of the overwintering period, rapeseed transitioned from vegetative growth to reproductive growth, and the NDVI value began to rise until reaching a second peak, marking the beginning of the rapeseed flowering period (DAS160). As the flowering period progressed, NDVI reached its lowest point during full bloom, and then began to increase again as petals gradually withered and siliques appeared. The VIs growth curves showed similar dynamics under all sowing scenarios, but the curves exhibited significant differences in geometric shape. As the sowing date was delayed, the first peak of the VIs curve decreased significantly. Furthermore, compared to NDVI, EVI and CIred-edge effectively mitigated the problem of NDVI saturation under high biomass conditions. There were no significant differences in the phenological dynamics of rapeseed under different densities at the same sowing date.

[0102] Regarding the morphological characteristics of VIs curves, crops such as wheat, corn, soybean, and rice all exhibit typical single-peak curves. In this invention, all rapeseed time-series VIs curves show typical multi-peak characteristics. The main reason for this is the influence of temperature on crop growth and the response of the spectrum to changes in crop canopy structure. As the sowing date is delayed, the temperature gradually decreases, and the growth rate of rapeseed slows down. During the 0-60 DAS period after sowing in the first and second sowing periods (ES and NS1), the average daily temperature is approximately 20℃. This suitable temperature allows for rapid canopy development in rapeseed during this stage, and the VI curve rapidly increases to its peak value. In contrast, in the third and fourth sowing periods, the temperature drops to around 5℃, slowing the canopy development rate, and the curve's growth rate and peak value are much lower than in the ES and NS1 scenarios. With the arrival of winter, winter rapeseed enters the overwintering period (leaf-wilt period), and the leaves that have undergone early vegetative growth begin to wither, causing the VI value of the curve to gradually decrease. Until the temperature rises again, rapeseed enters the reproductive growth stage, and the stems and lateral branches begin to develop, reaching another peak value after flowering. Unlike other crops (such as wheat, corn, and soybeans) that maintain a high VI value after reaching the reproductive growth stage, the VI value of winter rapeseed decreases again after reaching the flowering stage (DAS160). This is because the green canopy is gradually covered by bright yellow petals during the flowering period. The carotenoids in the petals absorb blue light and reflect green and red light bands, thus the curve profile shows a multi-peak characteristic.

[0103] Comparing different sowing scenarios, the total growing days of winter rapeseed decreased by approximately 45 days with delayed sowing. Previous studies have reported that delayed sowing reduces crop cover and accelerates its growth process, thus shortening the crop's growing days.

[0104] The shape model creation result is as follows Figure 6 As shown, in terms of fitting accuracy, different mathematical functions combined with VIs all exhibited good fitting results (RMSE < 0.2, R² > 0.7). However, only the AGF and FF functions captured the double-peak characteristic of the VIs curve. The DLF function maintained a single-peak shape in fitting all three VIs curves. Therefore, in fitting CIred-edge, the more significant peak characteristics of CIred-edge compared to NDVI and EVI further reduced the explanatory power of DLF for the curve. AGF, on the other hand, performed excellently in fitting CIred-edge (R² > 0.7). 2 =0.92, RMSE=0.10), FF ranks second, with its fitted curve accuracy slightly lower than AGF. The DLF model gives the lowest R-value among the three models. 2 And RMSE values. After determining the parameters of each shape model, the phenological dynamics of winter rapeseed under the other three sowing scenarios were monitored and extracted using the improved optimal parameter adjustment equation, and the dynamic response mechanism of winter rapeseed growth when the sowing date changes was further discussed.

[0105] The improved shape model's target curve matching results for winter rapeseed under different sowing scenarios are as follows: Figure 7 As shown. Among the models built based on the three functions, only the AGF model retained the shape features of the target curve in all scenarios. In most cases, the AGF model has a good interpretability for the dynamic changes of the rapeseed VI curve, but the accuracy of the AGF model decreases when the difference in the amplitude of the VI curve peaks and troughs is small, such as NDVI (NS1 and NS2) and EVI (NS1 and NS2). Compared with the AGF model, the FF model only retained the shape features of the curve in the NS1 validation. This is because the dynamic changes of NS1 and ES curves are basically consistent, while ES provides the dataset for model building. When the sowing date is delayed and the curve shape changes significantly, the FF model begins to fail to capture the dynamic changes of the VI curve. Through multi-scenario validation, the DLF model is not suitable for model application under multi-feature conditions. At the same time, the research results show that the curve matching accuracy is not completely equivalent to the phenological period estimation accuracy, because the curve matching accuracy is calculated by the VI value output after feature matching and the VI value of the target curve, while the phenological estimation value is obtained by stretching the x0 change using Equation 1-3 through the optimal scaling parameter output by the model, which is independent of the curve amplitude.

[0106] After the target curve matching was completed under different sowing scenarios, the optimal scaling parameters output by the model are shown in Table 2.

[0107]

[0108] Table 2 Optimal scaling parameters for model-matched growth curves

[0109] As mentioned earlier, the parameters xscale and tshift primarily adjust the shape model in the time domain, while yscale1 and 2 mainly control the amplitude of the VI curve. Taking the CIred-edge-AGF results as an example, the xscale value is slightly greater than 1, indicating that the growth cycle of the latter three sowing scenarios is shorter than that of the ES sowing scenario. To match the time span of the shape model, a stretching transformation of the time domain is required. A positive tshift value indicates that all shape models must be shifted to the right to approach the target curve. As the sowing date is delayed, the tshift values ​​of the three late-sowing treatments gradually increase, indicating that the rapeseed growth process is further accelerated. The difference in yscale1 and 2 values ​​indicates the difference in the amplitude of the rapeseed growth curve, that is, the stage difference between the vegetative growth and reproductive growth of winter rapeseed before and after overwintering.

[0110] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for monitoring the phenology of winter rapeseed based on an improved shape model, characterized in that, Comprising: Collecting the observation data of winter rape, and establishing a dynamic growth curve of winter rape; Selecting vegetation indices and establishing a shape model; Using the shape model for phenological detection to obtain the phenological dynamics of winter rape; The observation data includes the phenological information of winter rape, the state variables during the growth and development process, and the remote sensing images of the canopy reflectance at different growth nodes of winter rape; The vegetation index is an index used to measure the growth and lushness of vegetation, including Red-edge chlorophyll index: wherein CI represents a red edge chlorophyll index; R NIR represents a near infrared spectral band; R R-E represents a red edge spectral band; Normalized difference vegetation index: wherein NDVI represents the normalized difference vegetation index; R RED represents the red spectral band; Enhanced vegetation index: wherein EVI represents an enhanced vegetation index; a is a constant; β is a constant; R BLUE represents a blue spectral band; Establishing the shape model includes: selecting an objective function according to the characteristics of the crop growth curve to establish the shape model, and matching the predefined phenological stage X0 to the model curve; using the optimization parameter equation to perform feature matching on the input observation data and the shape model, and adjusting the observation data through the optimization parameters during the matching process; The optimization parameter equation is expressed as: g(x) = yscale × h(xscale × x0 + tshift) Where, g(x) represents the optimization parameter equation; yscale represents the y-axis scale; h is the established shape model; xscale represents the x-axis scale, 0.5 < xscale < 1.6; x0 represents the predefined crop phenological days; tshift represents the offset on the time axis, -70 < tshift < 70; Adjusting the observation data is expressed as: g1(x) = yscale1 × h1(xscale × x0 + tshift) + yscale2 × h2(xscale × x0 + tshift) Where, g1(x) represents the optimization parameter equation; yscale1 represents the y-axis scale 1, 0.1 < yscale1 < 1.5; h1 represents the first part of the model; yscale2 represents the y-axis scale 2, 0.1 < yscale2 < 1.5; h2 represents the second part of the model; Establishing the shape model further includes: selecting two functions of the asymmetric Gaussian function and the Fourier series function to establish the model, and comparing with the double logistic function; The asymmetric Gaussian function is: where AGF VIs represents an asymmetric Gaussian function; j is a fitting parameter that determines the amplitude of the model curve; k is a fitting parameter that determines the amplitude of the model curve; e is a constant; DAS represents days after sowing; f represents the center position of the first peak of the curve, h represents the center position of the second peak of the curve; g represents the width of the first peak; i represents the width of the second peak; The Fourier series function is: Among them, FF VIs The function represents the Fourier series function; l is the fitting parameter; m is the fitting parameter; n is the fitting parameter; i represents the vegetation index value of the bare soil background; p is the fundamental frequency of the signal; The double logistic function is expressed as: Among them, VI max denoted by , where represents the maximum value of the vegetation index; 'a' represents the maximum curvature when the curve is rising, and 'c' represents the maximum curvature when the curve is falling; 'b' is a fitting parameter related to the curvature of the vegetation index curve; and 'd' is a fitting parameter related to the curvature of the vegetation index curve.

2. The method for monitoring winter rapeseed phenology based on an improved shape model as described in claim 1, characterized in that: Using the shape model for phenological detection is expressed as: X est =xscale×x0+tshift Among them, X est This represents the number of phenological days estimated by the model.

3. A monitoring device employing the winter rapeseed phenological monitoring method based on an improved shape model as described in any one of claims 1 to 2, characterized in that, Including A collection system that collects the phenological information of winter rape, the state variables during the growth and development process, and the remote sensing images of the canopy reflectance at different growth nodes of winter rape, and uploads them to the data processing module; A data processing module that establishes a dynamic growth curve of winter rape, selects vegetation indices, establishes a shape model, and transmits it to the phenological detection module; A phenological detection module that performs phenological detection according to the established shape model to obtain the phenological dynamics of winter rape.

4. A computing device, comprising: A memory and a processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the winter rape phenological monitoring method based on the improved shape model described in any one of claims 1 to 2 are implemented.

5. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the winter rapeseed phenological monitoring method based on an improved shape model as described in any one of claims 1 to 2.

Images