Satellite sensor wave band conversion and standardization method supported by ground object spectrum database

Abstract

The invention discloses a satellite sensor wave band conversion and standardization method supported by a ground object spectrum database, and relates to the technical field of remote sensing, and the method comprises the steps: selecting a target farmland region, collecting the spectrum data corresponding to a target crop in the target farmland region, and constructing the ground object spectrum database according to the spectrum data; acquiring first remote sensing data corresponding to the target farmland area through satellite sensor equipment; based on a ground object spectrum database, constructing a satellite sensor wave band conversion model to convert the first remote sensing data into second remote sensing data; performing standardization processing on the second remote sensing data to generate target remote sensing data; calculating crop parameters corresponding to the target crop according to the target remote sensing data, monitoring the crop growth state of the target farmland area based on the crop parameters, and establishing a yield estimation model to determine the crop yield of the target farmland area, thereby eliminating the data difference between different satellite sensors, accurately monitoring the growth state of the crop, and improving the crop yield. And the accuracy and reliability of yield estimation are improved.

Description

Technical Field

[0001] The invention relates to the field of remote sensing technology, in particular to a satellite sensor band conversion and standardization method supported by a ground object spectrum database. Background Art

[0002] Satellite remote sensing technology plays an important role in crop growth monitoring and yield estimation. Using a known spectral library of representative objects, we can establish a correspondence between bands and spectral characteristics between different satellite sensors, and then convert the data obtained by different satellites to the same reference standard through mathematical algorithm models, thereby achieving interoperability between satellite sensors, unified standard processing of data, and consistency correction.

[0003] In the existing technology, in terms of crop growth monitoring, due to the differences in band settings, spectral resolution and radiation characteristics of remote sensing data obtained by different satellite sensors, traditional methods cannot efficiently realize the band conversion and standardization of satellite sensors, which directly affects the accurate extraction of crop growth information and the reliability of yield estimation, and affects the development of satellite remote sensing technology in the agricultural field.

[0004] Therefore, it is necessary to provide satellite sensor band conversion and standardization methods supported by ground object spectral database to solve the above technical problems. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a satellite sensor band conversion and standardization method supported by a ground object spectral database, which selects a target farmland area, collects spectral data corresponding to target crops in the target farmland area, and constructs a ground object spectral database according to the spectral data; obtains first remote sensing data corresponding to the target farmland area through a satellite sensor device; based on the ground object spectral database, constructs a satellite sensor band conversion model to convert the first remote sensing data into second remote sensing data; standardizes the second remote sensing data to generate target remote sensing data; calculates crop parameters corresponding to the target crop according to the target remote sensing data, monitors the crop growth status of the target farmland area based on the crop parameters, and establishes a yield estimation model to determine the crop yield of the target farmland area, so that the data differences between different satellite sensors can be eliminated through band conversion and standardization, and the accuracy and comparability of remote sensing data can be improved; the standardized reflectance data and vegetation index and other parameters can be used to more accurately monitor the growth status and health of crops; the spectral characteristics and crop growth model can be combined to establish a scientific yield estimation model to improve the accuracy and reliability of yield estimation; and it can provide strong technical support for precision agriculture, which helps farmers and agricultural managers make more scientific planting decisions and management measures.

[0006] In order to achieve the above object, the present invention is implemented by the following technical solutions:

[0007] The present invention is a satellite sensor band conversion and standardization method supported by a ground object spectral database, comprising:

[0008] Selecting a target farmland area, collecting spectral data corresponding to target crops in the target farmland area, and constructing a ground object spectral database based on the spectral data;

[0009] Acquiring first remote sensing data corresponding to the target farmland area through a satellite sensor device;

[0010] Based on the ground object spectral database, a satellite sensor band conversion model is constructed to convert the first remote sensing data into second remote sensing data;

[0011] performing standardization processing on the second remote sensing data to generate target remote sensing data;

[0012] The crop parameters corresponding to the target crop are calculated according to the target remote sensing data, the crop growth status of the target farmland area is monitored based on the crop parameters, and a yield estimation model is established to determine the crop yield of the target farmland area.

[0013] A further improvement of the present invention is that: the selecting of the target farmland area, collecting the spectral data corresponding to the target crops in the target farmland area, and constructing the ground object spectral database according to the spectral data specifically includes:

[0014] Selecting the target farmland area, determining the target crops in the target farmland area, and collecting spectral data of the target crops at different growth stages and environmental conditions;

[0015] Recording and storing the spectral data, and sorting, cleaning and verifying the spectral data to generate target spectral data;

[0016] The ground object spectrum database is constructed based on the target spectrum data.

[0017] A further improvement of the present invention is that the obtaining of the first remote sensing data corresponding to the target farmland area by means of a satellite sensor device specifically includes:

[0018] According to the preset monitoring requirements corresponding to the target crop, matching the resolution parameters corresponding to the preset monitoring requirements, and determining the satellite sensor device according to the resolution parameters;

[0019] Determining a preset collection route based on the target farmland area;

[0020] Based on the preset collection route, the first remote sensing data corresponding to the target farmland area is acquired through the satellite sensor device.

[0021] A further improvement of the present invention is that: based on the ground object spectrum database, constructing a satellite sensor band conversion model to convert the first remote sensing data into the second remote sensing data specifically includes:

[0022] Based on the ground object spectrum database, a band conversion coefficient is determined by a linear regression fitting method, and the satellite sensor band conversion model is constructed to convert the first remote sensing data into input remote sensing data;

[0023] Retrieving a preset atmospheric correction model to process the input remote sensing data to obtain output remote sensing data;

[0024] Obtain the measured data corresponding to the target farmland area, compare the difference between the measured data and the output remote sensing data to obtain the corresponding conversion accuracy. If the conversion accuracy is greater than or equal to the preset accuracy, the output remote sensing data is the second remote sensing data, otherwise repeat the above processing steps.

[0025] A further improvement of the present invention is that: based on the ground object spectrum database, the band conversion coefficient is determined by a linear regression fitting method, the satellite sensor band conversion model is constructed to convert the first remote sensing data into input remote sensing data, and the calculation is performed by the following band conversion formula:

[0026]

[0027] In the formula, R1(λ i ) represents the wavelength λ corresponding to the first remote sensing data i The reflectivity of the ground object at j ) represents the wavelength λ corresponding to the input remote sensing data j The reflectivity of the ground at c j represents the band conversion coefficient; c0 represents the offset coefficient.

[0028] A further improvement of the present invention is that the standardization processing of the second remote sensing data to generate target remote sensing data specifically includes:

[0029] Performing preliminary cleaning processing on the second remote sensing data to generate primary processed remote sensing data;

[0030] Performing spectral filtering on the primary processed remote sensing data to remove noise and redundant information in the second remote sensing data to generate secondary processed remote sensing data;

[0031] The envelope in the secondary processed remote sensing data is identified and removed to generate target remote sensing data.

[0032] A further improvement of the present invention is that the crop parameters corresponding to the target crop are calculated according to the target remote sensing data, the crop growth status of the target farmland area is monitored based on the crop parameters, and a yield estimation model is established to determine the crop yield of the target farmland area, specifically including:

[0033] Calculating crop parameters corresponding to the target crop according to the target remote sensing data;

[0034] Real-time monitoring of the crop growth status of the target farmland area according to the time series changes of the crop parameters;

[0035] An initial estimation model is constructed based on the crop parameters, and historical yield data of the target farmland area is retrieved to optimize the initial estimation model, the yield estimation model is generated, and the crop yield of the target farmland area is determined based on the yield estimation model.

[0036] A further improvement of the present invention is that the crop parameters corresponding to the target crop are calculated according to the target remote sensing data by the following monitoring formula:

[0037]

[0038] In the formula, ZWCS represents the crop parameters corresponding to the target crop; G represents the gain factor; N, R, and B represent the reflectance of the near-infrared band, red band, and blue band in the target remote sensing data; C1 and C2 represent the atmospheric correction coefficients; and R represents the soil adjustment parameter.

[0039] The crop parameters corresponding to the target crop are calculated based on the target remote sensing data, and the crop growth status of the target farmland area is monitored in real time according to the time series changes of the crop parameters.

[0040] A further improvement of the present invention is that the historical yield data of the target farmland area is retrieved to optimize the initial estimation model to generate the yield estimation model, which specifically includes:

[0041] Randomly initialize all weights W0 in the neural network corresponding to the initial estimation model;

[0042] According to the target remote sensing data and the current weight W, forward propagation is performed to calculate the predicted yield data According to the forecast production data The loss function L(W) is calculated based on the historical production data Y. The expression of the loss function L(W) is:

[0043]

[0044] Where L(W) represents the loss function; N represents the total number of data points of the target remote sensing data; Y i Represents the i-th historical production data; Represents the i-th predicted production data; represents the square of the i-th prediction error;

[0045] Calculate the gradient of the loss function L(W) with respect to the current weight W The gradient The expression is:

[0046]

[0047] In the formula, Represents the gradient of the loss function with respect to the current weight; Represents the partial derivative of the loss function with respect to the i-th component in the current weight, that is, the i-th component of the gradient;

[0048] According to the gradient Update the current weight W:

[0049]

[0050] Where W' represents the updated weight; W represents the current weight; α represents the learning rate, which controls the step size of the weight update; Represents the gradient of the loss function with respect to the current weight;

[0051] Repeat the above steps until the loss function converges to a preset threshold or reaches a preset maximum number of iterations, completing the optimization process of the initial estimation model and generating the corresponding production estimation model.

[0052] Compared with the related art, the satellite sensor band conversion and standardization method supported by the ground object spectral database provided by the present invention has the following beneficial effects:

[0053] The present invention selects a target farmland area, collects spectral data corresponding to target crops in the target farmland area, and constructs a ground object spectral database according to the spectral data; obtains first remote sensing data corresponding to the target farmland area through a satellite sensor device; constructs a satellite sensor band conversion model based on the ground object spectral database to convert the first remote sensing data into second remote sensing data; performs standardization processing on the second remote sensing data to generate target remote sensing data; calculates crop parameters corresponding to the target crop according to the target remote sensing data, monitors the crop growth status of the target farmland area based on the crop parameters, and establishes a yield estimation model to determine the crop yield of the target farmland area, so that data differences between different satellite sensors can be eliminated through band conversion and standardization processing, and the accuracy and comparability of remote sensing data can be improved; the growth status and health status of crops can be more accurately monitored by using standardized reflectance data and vegetation index and other parameters; a scientific yield estimation model can be established by combining spectral characteristics and crop growth models to improve the accuracy and reliability of yield estimation; and strong technical support can be provided for precision agriculture, which helps farmers and agricultural managers make more scientific planting decisions and management measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic flow chart of a satellite sensor band conversion and standardization method supported by a ground feature spectral database provided in an embodiment of the present invention;

[0055] Figure 2 A wavelength-reflectivity curve diagram of typical objects in a target farmland area provided by an embodiment of the present invention;

[0056] Figure 3 A schematic diagram of data collection by a satellite sensor device provided in an embodiment of the present invention;

[0057] Figure 4 A flowchart of a standardized process provided by an embodiment of the present invention;

[0058] Figure 5 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1, is a flow chart of a satellite sensor band conversion and standardization method supported by a ground object spectral database provided in an embodiment of the present invention, Figure 1 The execution subject of the method shown may be a software and / or hardware device. The execution subject of the present application may include but is not limited to at least one of the following: user equipment, network equipment, etc. Among them, user equipment may include but is not limited to computers, smart phones, personal digital assistants (Personal Digital Assistant, referred to as: PDA) and the electronic devices mentioned above. Network equipment may include but is not limited to a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers, wherein cloud computing is a type of distributed computing, a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. It includes steps S1 to S5, as follows:

[0061] S1, selecting a target farmland area, collecting spectral data corresponding to target crops in the target farmland area, and constructing a ground feature spectral database based on the spectral data;

[0062] S2, acquiring first remote sensing data corresponding to the target farmland area through a satellite sensor device;

[0063] S3, based on the ground feature spectral database, constructing a satellite sensor band conversion model to convert the first remote sensing data into second remote sensing data;

[0064] S4, performing standardization processing on the second remote sensing data to generate target remote sensing data;

[0065] S5, calculating crop parameters corresponding to the target crop according to the target remote sensing data, monitoring the crop growth status of the target farmland area based on the crop parameters, and establishing a yield estimation model to determine the crop yield of the target farmland area.

[0066] The selecting of a target farmland area, collecting spectral data corresponding to target crops in the target farmland area, and constructing a ground object spectral database according to the spectral data specifically includes:

[0067] Selecting the target farmland area, determining the target crops in the target farmland area, and collecting spectral data of the target crops at different growth stages and environmental conditions;

[0068] Recording and storing the spectral data, and sorting, cleaning and verifying the spectral data to generate target spectral data;

[0069] The ground object spectrum database is constructed based on the target spectrum data.

[0070] It is understandable that in the context of the deep integration of precision agriculture and remote sensing technology, in order to meet the crop monitoring and management needs of the target farmland area, we can first select a representative piece of farmland as the research object, that is, the target farmland area, and this process needs to comprehensively consider soil type, climate conditions, irrigation facilities, historical crop planting records and other factors to ensure the universality and accuracy of the results.

[0071] See also Figure 2 , is a wavelength-reflectivity curve diagram of a typical object in the target farmland area provided by an embodiment of the present invention, which can clearly define the target crops in the target farmland area. This process is often determined based on the local agricultural industrial structure, market demand and crop growth characteristics, such as common food crops such as wheat, corn, and rice, or cash crops such as vegetables and fruits. Using advanced equipment such as high-precision spectrometers, the target crops can be monitored throughout the process to obtain spectral reflectance characteristics under different growth stages (such as sowing period, seedling period, growth period, maturity period and harvest period) and different environmental conditions (such as sunlight intensity, temperature, humidity, soil nutrient content, etc.). These spectral data can reflect the crop physiological state, nutrient status and occurrence of pests and diseases, and thus a spectral database of objects can be constructed.

[0072] By using data processing techniques, including data organization (removing duplicates and unifying formats), data cleaning (eliminating outliers and suppressing noise) and data verification (cross-validation and consistency checking), high-quality target spectral data sets can be generated, thereby improving the accuracy and reliability of spectral data.

[0073] Based on the target spectral data, a ground feature spectral database can be established. This database not only contains the spectral characteristic information of the target crops, but also may cover diversified data such as related growth environment parameters and agricultural operation records. Therefore, it can integrate data collection, storage, management and analysis. Through the ground feature spectral database, the changing laws of the spectral characteristics of the target crops can be easily queried and analyzed, which is convenient for subsequent crop growth monitoring and yield prediction, thereby promoting the intelligent and precise development of agricultural production.

[0074] The step of obtaining the first remote sensing data corresponding to the target farmland area through a satellite sensor device specifically includes:

[0075] According to the preset monitoring requirements corresponding to the target crop, matching the resolution parameters corresponding to the preset monitoring requirements, and determining the satellite sensor device according to the resolution parameters;

[0076] Determining a preset collection route based on the target farmland area;

[0077] Based on the preset collection route, the first remote sensing data corresponding to the target farmland area is acquired through the satellite sensor device.

[0078] First, the corresponding preset monitoring requirements can be determined according to the specific growth cycle and physiological characteristics of the target crops. These requirements may cover crop growth status, pest and disease occurrence, soil moisture and nutrient content, etc. According to these preset monitoring requirements, the corresponding resolution parameters can be accurately matched. It should be noted that the resolution parameters directly determine the precision of remote sensing data and the accuracy of subsequent analysis.

[0079] Therefore, for scenarios that require precise monitoring of crop canopy structure or subtle characteristics of pests and diseases, high-resolution satellite sensors are required; while for large-scale farmland soil moisture monitoring, medium-resolution sensors can be selected to balance data processing efficiency and cost.

[0080] See also Figure 3 , is a schematic diagram of the satellite sensor device collecting data provided by an embodiment of the present invention. Based on the geographical location, topography and crop distribution characteristics of the target farmland area, the optimal preset collection route can be planned using GIS technology. This step fully considers factors such as the observation angle, lighting conditions and cloud cover of the satellite sensor, thereby ensuring that interference can be minimized during the data collection process and obtaining high-quality, continuous and comprehensive remote sensing data.

[0081] According to the preset collection route, the data collection task is performed by satellite sensor equipment. Satellite sensor equipment, such as multispectral cameras and synthetic aperture radar (SAR), can capture and record multiple information such as reflectivity, radiation temperature and scattering characteristics of the target farmland area by using their sensitivity to surface features in different bands to form the first remote sensing data set. These data not only contain direct information on crop growth, but also include the comprehensive impact of environmental factors such as soil and climate, which is convenient for subsequent data processing, model construction and agricultural decision-making.

[0082] Through the above-mentioned method, satellite sensor equipment can be efficiently used to obtain comprehensive and accurate first-line remote sensing data of the target farmland area, promoting precise and efficient agricultural management.

[0083] The step of constructing a satellite sensor band conversion model based on the ground object spectral database to convert the first remote sensing data into the second remote sensing data specifically includes:

[0084] Based on the ground object spectrum database, a band conversion coefficient is determined by a linear regression fitting method, and the satellite sensor band conversion model is constructed to convert the first remote sensing data into input remote sensing data;

[0085] Retrieving a preset atmospheric correction model to process the input remote sensing data to obtain output remote sensing data;

[0086] Obtain the measured data corresponding to the target farmland area, compare the difference between the measured data and the output remote sensing data to obtain the corresponding conversion accuracy. If the conversion accuracy is greater than or equal to the preset accuracy, the output remote sensing data is the second remote sensing data, otherwise repeat the above processing steps.

[0087] Based on the ground object spectrum database, the band conversion coefficient is determined by a linear regression fitting method, and the satellite sensor band conversion model is constructed to convert the first remote sensing data into input remote sensing data, and the calculation is performed by the following band conversion formula:

[0088]

[0089] In the formula, R1(λ i ) represents the wavelength λ corresponding to the first remote sensing data i The reflectivity of the ground object at j ) represents the wavelength λ corresponding to the input remote sensing data j The reflectivity of the ground at c j represents the band conversion coefficient; c0 represents the offset coefficient.

[0090] In practical applications, by collecting and processing more than 50,000 sets of spectral reflectance data of different objects, a band conversion tool can be built to realize band conversion processing between different satellites. Through the linear regression fitting method, the band conversion coefficient can be calculated, and the satellite sensor band conversion model can be built to convert the band data of any satellite into the multi-angle band data of the target satellite, so as to realize the standardized processing of satellite remote sensing big data, ensure the consistency and comparability of data, and provide reliable technical support for cross-satellite data fusion and analysis.

[0091] The step of performing standardization processing on the second remote sensing data to generate target remote sensing data specifically includes:

[0092] Performing preliminary cleaning processing on the second remote sensing data to generate primary processed remote sensing data;

[0093] Performing spectral filtering on the primary processed remote sensing data to remove noise and redundant information in the second remote sensing data to generate secondary processed remote sensing data;

[0094] The envelope in the secondary processed remote sensing data is identified and removed to generate target remote sensing data.

[0095] See also Figure 4, is a flowchart of the standardized processing provided by an embodiment of the present invention, which can perform preliminary cleaning processing on the second remote sensing data. Specifically, through statistical analysis and outlier detection methods, missing values, outliers and data points that obviously do not conform to physical laws in the second remote sensing data can be identified and corrected; through interpolation, smoothing or elimination and other means, errors and inconsistencies in the data can be effectively reduced, and once-processed remote sensing data can be generated, thereby improving the integrity and accuracy of the data.

[0096] Subsequently, the primary processed remote sensing data can be subjected to spectral filtering. Spectral filtering refers to a technique that finely controls the information of remote sensing data in the spectral dimension by designing appropriate filters based on the principle of signal processing. This step can remove noise components and redundant information in the data, such as atmospheric interference, sensor noise, and reflections from non-target objects on the surface, thereby enhancing the spectral characteristics of the target objects. The secondary processed remote sensing data generated after spectral filtering can be significantly improved in terms of spectral resolution and signal-to-noise ratio.

[0097] The envelope refers to the curve in remote sensing data that reflects the overall trend of the spectral characteristics of the object. It contains the basic spectral information of the object, but may also cover up some subtle spectral changes. In order to more accurately extract the spectral characteristics of the target object, these envelopes can be identified and removed. This process is usually combined with advanced algorithms such as spectral decomposition and curve fitting. By extracting the characteristic points of the spectral curve, constructing and removing the envelope, the target remote sensing data is finally generated.

[0098] Through the above method, not only the key information in the original data can be retained, but also the noise, redundancy and envelope in the original data can be removed, ensuring the accuracy of the spectral characteristics of the target object, which is convenient for subsequent classification, identification and monitoring.

[0099] The step of calculating the crop parameters corresponding to the target crop according to the target remote sensing data, monitoring the crop growth status of the target farmland area based on the crop parameters, and establishing a yield estimation model to determine the crop yield of the target farmland area specifically includes:

[0100] Calculating crop parameters corresponding to the target crop according to the target remote sensing data;

[0101] Real-time monitoring of the crop growth status of the target farmland area according to the time series changes of the crop parameters;

[0102] An initial estimation model is constructed based on the crop parameters, and historical yield data of the target farmland area is retrieved to optimize the initial estimation model, the yield estimation model is generated, and the crop yield of the target farmland area is determined based on the yield estimation model.

[0103] Among them, crop parameters related to target crops can be accurately extracted from target remote sensing data through specific remote sensing algorithms, such as vegetation index calculation, spectral inversion model, etc. These crop parameters include leaf area index (LAI), fractional vegetation cover (FVC), biomass, and chlorophyll content.

[0104] Next, based on the time series changes of crop parameters, real-time monitoring of crop growth status in the target farmland area can be achieved. Time series analysis can capture the dynamic changes of key growth stages of crops throughout the growth cycle, such as growth rate, phenological period, etc., and then evaluate key growth indicators such as crop health, water stress and nutritional level. This process can combine the advantages of remote sensing technology in spatiotemporal continuous observation to provide timely and accurate crop growth information for agricultural production.

[0105] Furthermore, a yield estimation model can be constructed to predict the crop yield of the target farmland area. The establishment of the initial estimation model depends on the statistical relationship between crop parameters and yield. In order to improve the accuracy and reliability of yield prediction, the historical yield data of the target farmland area can be retrieved and integrated. The initial estimation model can be trained and verified through optimization algorithms such as machine learning and deep learning, thereby generating a yield estimation model that is more in line with actual production.

[0106] Through the above method, based on the optimized yield estimation model, the crop yield of the target farmland area can be predicted, which will promote the progress of agricultural production planning, resource allocation and market decision-making, improve the level of intelligent agricultural production, and promote the sustainable development of agriculture.

[0107] The crop parameters corresponding to the target crop are calculated according to the target remote sensing data, and are calculated by the following monitoring formula:

[0108]

[0109] In the formula, ZWCS represents the crop parameters corresponding to the target crop; G represents the gain factor; N, R, and B represent the reflectance of the near-infrared band, red band, and blue band in the target remote sensing data; C1 and C2 represent the atmospheric correction coefficients; and R represents the soil adjustment parameter.

[0110] The crop parameters corresponding to the target crop are calculated based on the target remote sensing data, and the crop growth status of the target farmland area is monitored in real time according to the time series changes of the crop parameters.

[0111] The step of retrieving the historical yield data of the target farmland area to optimize the initial estimation model and generate the yield estimation model specifically includes:

[0112] Randomly initialize all weights W0 in the neural network corresponding to the initial estimation model;

[0113] According to the target remote sensing data and the current weight W, forward propagation is performed to calculate the predicted yield data According to the forecast production data The loss function L(W) is calculated based on the historical production data Y. The expression of the loss function L(W) is:

[0114]

[0115] Where L(W) represents the loss function; N represents the total number of data points of the target remote sensing data; Y i Represents the i-th historical production data; Represents the i-th predicted production data; represents the square of the i-th prediction error;

[0116] Calculate the gradient of the loss function L(W) with respect to the current weight W The gradient The expression is:

[0117]

[0118] In the formula, Represents the gradient of the loss function with respect to the current weight; Represents the partial derivative of the loss function with respect to the i-th component in the current weight, that is, the i-th component of the gradient;

[0119] According to the gradient Update the current weight W:

[0120]

[0121] Where W' represents the updated weight; W represents the current weight; α represents the learning rate, which controls the step size of the weight update; Represents the gradient of the loss function with respect to the current weight;

[0122] Repeat the above steps until the loss function converges to a preset threshold or reaches a preset maximum number of iterations, completing the optimization process of the initial estimation model and generating the corresponding production estimation model.

[0123] Through the introduction of the above embodiments, the present invention can select a target farmland area through a satellite sensor band conversion and standardization method supported by a ground object spectral database, collect spectral data corresponding to target crops in the target farmland area, and construct a ground object spectral database according to the spectral data; obtain first remote sensing data corresponding to the target farmland area through a satellite sensor device; based on the ground object spectral database, construct a satellite sensor band conversion model to convert the first remote sensing data into second remote sensing data; standardize the second remote sensing data to generate target remote sensing data; calculate crop parameters corresponding to the target crop according to the target remote sensing data, monitor the crop growth status of the target farmland area based on the crop parameters, and establish a yield estimation model to determine the crop yield of the target farmland area, so that the data differences between different satellite sensors can be eliminated through band conversion and standardization, and the accuracy and comparability of remote sensing data can be improved; using standardized reflectance data and vegetation index and other parameters, the growth status and health of crops can be more accurately monitored; spectral characteristics and crop growth models can be combined to establish a scientific yield estimation model to improve the accuracy and reliability of yield estimation; and strong technical support can be provided for precision agriculture, which helps farmers and agricultural managers make more scientific planting decisions and management measures.

[0124] See also Figure 5 , is a schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention, the electronic device 50 includes: a processor 51, a memory 52 and a computer program; wherein

[0125] The memory 52 is used to store the computer program, and the memory may also be a flash memory. The computer program is, for example, an application program, a functional module, etc. for implementing the above method.

[0126] The processor 51 is used to execute the computer program stored in the memory to implement each step performed by the device in the above method. For details, please refer to the relevant description in the above method embodiment.

[0127] Optionally, the memory 52 may be independent or integrated with the processor 51 .

[0128] When the memory 52 is a device independent of the processor 51, the device may further include:

[0129] The bus 53 is used to connect the memory 52 and the processor 51 .

[0130] The present invention also provides a readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it is used to implement the methods provided by the various embodiments described above.

[0131] Among them, the readable storage medium can be a computer storage medium or a communication medium. The communication medium includes any medium that facilitates the transmission of a computer program from one place to another. The computer storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, a readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application-specific integrated circuit (Application Specific Integrated Circuits, abbreviated as: ASIC). In addition, the ASIC can be located in a user device. Of course, the processor and the readable storage medium can also exist in a communication device as discrete components. The readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0132] The present invention also provides a program product, which includes an execution instruction, which is stored in a readable storage medium. At least one processor of a device can read the execution instruction from the readable storage medium, and at least one processor executes the execution instruction so that the device implements the methods provided in the above various embodiments.

[0133] In the embodiments of the above-mentioned devices, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A satellite sensor band conversion and standardization method supported by a ground feature spectral database, characterized in that: include: Selecting a target farmland area, collecting spectral data corresponding to target crops in the target farmland area, and constructing a ground object spectral database based on the spectral data; Acquiring first remote sensing data corresponding to the target farmland area through a satellite sensor device; Based on the ground object spectral database, a satellite sensor band conversion model is constructed to convert the first remote sensing data into second remote sensing data; performing standardization processing on the second remote sensing data to generate target remote sensing data; The crop parameters corresponding to the target crop are calculated according to the target remote sensing data, the crop growth status of the target farmland area is monitored based on the crop parameters, and a yield estimation model is established to determine the crop yield of the target farmland area.

2. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 1 is characterized in that: The selecting of a target farmland area, collecting spectral data corresponding to target crops in the target farmland area, and constructing a ground object spectral database according to the spectral data specifically includes: Selecting the target farmland area, determining the target crops in the target farmland area, and collecting spectral data of the target crops at different growth stages and environmental conditions; Recording and storing the spectral data, and sorting, cleaning and verifying the spectral data to generate target spectral data; The ground object spectrum database is constructed based on the target spectrum data.

3. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 1 is characterized in that: The step of obtaining the first remote sensing data corresponding to the target farmland area through a satellite sensor device specifically includes: According to the preset monitoring requirements corresponding to the target crop, matching the resolution parameters corresponding to the preset monitoring requirements, and determining the satellite sensor device according to the resolution parameters; Determining a preset collection route based on the target farmland area; Based on the preset collection route, the first remote sensing data corresponding to the target farmland area is acquired through the satellite sensor device.

4. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 1 is characterized in that: The step of constructing a satellite sensor band conversion model based on the ground object spectral database to convert the first remote sensing data into the second remote sensing data specifically includes: Based on the ground object spectrum database, a band conversion coefficient is determined by a linear regression fitting method, and the satellite sensor band conversion model is constructed to convert the first remote sensing data into input remote sensing data; Retrieving a preset atmospheric correction model to process the input remote sensing data to obtain output remote sensing data; Obtain the measured data corresponding to the target farmland area, compare the difference between the measured data and the output remote sensing data to obtain the corresponding conversion accuracy. If the conversion accuracy is greater than or equal to the preset accuracy, the output remote sensing data is the second remote sensing data, otherwise repeat the above processing steps.

5. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 4 is characterized in that: Based on the ground object spectrum database, the band conversion coefficient is determined by a linear regression fitting method, and the satellite sensor band conversion model is constructed to convert the first remote sensing data into input remote sensing data, and the calculation is performed by the following band conversion formula: In the formula, R1(λ i ) represents the wavelength λ corresponding to the first remote sensing data i The reflectivity of the ground object at j ) represents the wavelength λ corresponding to the input remote sensing data j The reflectivity of the ground at c j represents the band conversion coefficient; c0 represents the offset coefficient.

6. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 1 is characterized in that: The step of performing standardization processing on the second remote sensing data to generate target remote sensing data specifically includes: Performing preliminary cleaning processing on the second remote sensing data to generate primary processed remote sensing data; Performing spectral filtering on the primary processed remote sensing data to remove noise and redundant information in the second remote sensing data to generate secondary processed remote sensing data; The envelope in the secondary processed remote sensing data is identified and removed to generate target remote sensing data.

7. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 1 is characterized in that: The step of calculating the crop parameters corresponding to the target crop according to the target remote sensing data, monitoring the crop growth status of the target farmland area based on the crop parameters, and establishing a yield estimation model to determine the crop yield of the target farmland area specifically includes: Calculating crop parameters corresponding to the target crop according to the target remote sensing data; Real-time monitoring of the crop growth status of the target farmland area according to the time series changes of the crop parameters; An initial estimation model is constructed based on the crop parameters, and historical yield data of the target farmland area is retrieved to optimize the initial estimation model, the yield estimation model is generated, and the crop yield of the target farmland area is determined based on the yield estimation model.

8. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 7 is characterized in that: The crop parameters corresponding to the target crop are calculated according to the target remote sensing data, and are calculated by the following monitoring formula: In the formula, ZWCS represents the crop parameters corresponding to the target crop; G represents the gain factor; N, R, and B represent the reflectance of the near-infrared band, red band, and blue band in the target remote sensing data; C1 and C2 represent the atmospheric correction coefficients; and R represents the soil adjustment parameter. The crop parameters corresponding to the target crop are calculated based on the target remote sensing data, and the crop growth status of the target farmland area is monitored in real time according to the time series changes of the crop parameters.

9. The satellite sensor band conversion and standardization method supported by the ground feature spectral database according to claim 7, characterized in that: The step of retrieving the historical yield data of the target farmland area to optimize the initial estimation model and generate the yield estimation model specifically includes: Randomly initialize all weights W0 in the neural network corresponding to the initial estimation model; According to the target remote sensing data and the current weight W, forward propagation is performed to calculate the predicted yield data According to the forecast production data The loss function L(W) is calculated based on the historical production data Y. The expression of the loss function L(W) is: Where L(W) represents the loss function; N represents the total number of data points of the target remote sensing data; Y i represents the i-th historical production data; Represents the i-th predicted production data; represents the square of the i-th prediction error; Calculate the gradient of the loss function L(W) with respect to the current weight W The gradient The expression is: In the formula, Represents the gradient of the loss function with respect to the current weight; Represents the partial derivative of the loss function with respect to the i-th component in the current weight, that is, the i-th component of the gradient; According to the gradient Update the current weight W: Where W' represents the updated weight; W represents the current weight; α represents the learning rate, which controls the step size of the weight update; Represents the gradient of the loss function with respect to the current weight; Repeat the above steps until the loss function converges to a preset threshold or reaches a preset maximum number of iterations, completing the optimization process of the initial estimation model and generating the corresponding production estimation model.

Images