A comprehensive management system for forestry resource data

Abstract

The present invention relates to the technical field of forestry data management, and specifically discloses a comprehensive management system for forestry resource data, which is used to solve the problem that the existing traditional defoliation monitoring method relies on spectral differences, but due to the leaves not falling off, it is difficult to effectively identify through single spectral inversion. The present invention obtains the spectral data and image data of the tree crown, extracts the spectral data and image data of the tree crown in the pest state and the healthy state, preliminarily analyzes whether there are diseases in the leaves according to the pixel values extracted from the image data, and finally analyzes whether there are diseases in the tree crown according to the spectral data, which helps to improve the comprehensiveness and accuracy of disease detection, reduce misjudgment and missed judgment, and the dual analysis mechanism ensures the sensitive identification of early diseases, and can provide data support for timely intervention.

Description

Technical Field

[0001] The present invention relates to the technical field of forestry data management. More specifically, the present invention relates to a comprehensive management system for forestry resource data. Background Art

[0002] Remote sensing monitoring of forest pests and diseases is an important part of forest resource protection and ecological management. However, for remote sensing-based pest and disease monitoring, especially for the monitoring of crown defoliation, due to the sudden, hidden, and rapid spread characteristics of pests and diseases, traditional remote sensing monitoring methods have certain limitations in terms of time and space, resulting in the inability to respond promptly to the monitoring and control of pests and diseases, increasing the risk of damage to the forest ecosystem. In the existing literature "Qi Xinglan, Cao Zuning, Liu Jian, et al. Research Progress on Forest Pest and Disease Monitoring Based on Satellite Remote Sensing Images [J]. Forest Resources Management, 2020, (02): 181-186. DOI: 10.13466 / j.cnki.lyzygl.2020.02.027.", the research status of using satellite remote sensing images to monitor forest pests and diseases at home and abroad is comprehensively analyzed, the main technical methods for applying satellite remote sensing images to monitor forest pests and diseases are systematically summarized, the main existing problems are deeply analyzed and research prospects are carried out. Some pests and diseases (such as longhorn beetles and spruce leaf rust) can cause the leaves of trees to turn yellow, wither, or lose green, but the leaves still remain on the crown for a certain period of time and do not form significant defoliation characteristics. Traditional defoliation monitoring methods rely on spectral differences, but due to the leaves not falling off, it is difficult to effectively identify through single-spectral inversion. To solve the above problems, a technical solution is provided herein. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, the present invention provides a comprehensive management system for forestry resource data, which is used to solve the problem that traditional defoliation monitoring methods rely on spectral differences, but due to the leaves not falling off, it is difficult to effectively identify through single-spectral inversion, so as to solve the problems raised in the above background art.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A comprehensive forestry resource data management system includes a multi-dimensional data acquisition module, a diseased leaf distribution recognition module, and a dynamic disease management module. The multi-dimensional data acquisition module is used to obtain the spectral data and image data of the tree crown. The diseased leaf distribution recognition module includes a first data acquisition unit, a data processing unit, a diseased leaf distribution analysis unit, a non-diseased leaf distribution analysis unit, and a dynamic leaf distribution analysis unit. The first data acquisition unit is used to extract the first image data, the second image data, and the third image data. The data processing unit is used to perform segmentation processing on the first image data, the second image data, and the third image data respectively and extract their pixel values to obtain the first pixel data, the second pixel data, and the third pixel data. The diseased leaf distribution analysis unit is used to analyze the diseased leaf distribution coefficient according to the second pixel data. The non-diseased leaf distribution analysis unit is used to analyze the non-diseased leaf distribution coefficient according to the third pixel data. The dynamic leaf distribution analysis unit is used to preliminarily analyze whether there is a disease in the leaves based on the first pixel data, the diseased leaf distribution coefficient, and the non-diseased leaf distribution coefficient. The formula of the primary leaf disease discrimination model is:

[0006] ;

[0007] In the formula: is the primary leaf disease discrimination value, is the non-diseased leaf distribution coefficient, is the diseased leaf distribution coefficient, is the number of the first pixel data in the third pixel sequence, is the q-th first pixel data in the third pixel sequence, is the mean value of the first pixel data in the third pixel sequence.

[0008] As a further solution of the present invention, the multi-dimensional data acquisition module includes a multi-spectral camera, a first-level data extraction unit, and a second-level data extraction unit. The multi-spectral camera is used to obtain the first spectral data and the first image data of the tree crown within the target range in real time. The first-level data extraction unit is used to obtain the second spectral data of the historical pest-infected tree crown and the third spectral data of the historical pest-free tree crown. The second-level data extraction unit is used to obtain the second image data of the historical pest-infected tree crown and the third image data of the historical pest-free tree crown.

[0009] As a further solution of the present invention, the diseased leaf distribution analysis unit is used to analyze the diseased leaf distribution coefficient according to the second pixel data. Specifically: the second pixel data is arranged in ascending order to obtain the first pixel sequence, the maximum value in the first pixel sequence is extracted as the first pixel value, the minimum value in the first pixel sequence is extracted as the second pixel value, the mean value of the second pixel data is calculated based on the first pixel sequence to obtain the first pixel distribution coefficient, and a diseased leaf distribution coefficient calculation model is constructed according to the second pixel data, the first pixel value, the second pixel value, and the first pixel distribution coefficient to calculate the diseased leaf distribution coefficient. The calculation formula of the diseased leaf distribution coefficient calculation model is:

[0010] ;

[0011] In the formula: is the diseased leaf distribution coefficient, is the number of the second pixel data in the first pixel sequence, is the i-th second pixel data in the first pixel sequence, is the first pixel distribution coefficient, is the first pixel value, is the second pixel value.

[0012] As a further solution of the present invention, the non-diseased leaf distribution analysis unit is used to analyze the non-diseased leaf distribution coefficient according to the third pixel data. Specifically: the third pixel data is arranged in ascending order to obtain the second pixel sequence, the maximum value in the second pixel sequence is extracted as the third pixel value, the minimum value in the second pixel sequence is extracted as the fourth pixel value, the mean value of the third pixel data is calculated based on the second pixel sequence to obtain the second pixel distribution coefficient, and a non-diseased leaf distribution coefficient calculation model is constructed according to the third pixel data, the fourth pixel value, the third pixel value, and the second pixel distribution coefficient to calculate the non-diseased leaf distribution coefficient. The calculation formula of the non-diseased leaf distribution coefficient calculation model is:

[0013] ;

[0014] In the formula: is the non-diseased leaf distribution coefficient, is the number of the third pixel data in the second pixel sequence, is the j-th third pixel data in the second pixel sequence, is the second pixel distribution coefficient, is the third pixel value, is the fourth pixel value.

[0015] As a further solution of the present invention, the dynamic blade distribution analysis unit is used to analyze whether there is a disease in the blade based on the first pixel data, the disease blade distribution coefficient, and the non-disease blade distribution coefficient. Specifically: extract the first pixel data, obtain the third pixel sequence by arranging the first pixel data in ascending order, construct a primary blade disease discrimination model based on the third pixel sequence, the disease blade distribution coefficient, and the non-disease blade distribution coefficient, and preliminarily determine whether there is a disease in the blade.

[0016] As a further solution of the present invention, the dynamic disease management module includes a second data acquisition unit, a spectral curve drawing unit, a disease spectral coefficient calculation unit, and a secondary blade spectral analysis unit;

[0017] The second data acquisition unit is used to extract the first spectral data, the second spectral data, and the third spectral data;

[0018] The spectral curve drawing unit is used to draw spectral curves according to the first spectral data, the second spectral data, and the third spectral data respectively;

[0019] The disease spectral coefficient calculation unit is used to calculate the canopy disease spectral coefficient according to the second spectral data and the third spectral data;

[0020] The secondary blade spectral analysis unit is used to secondarily analyze whether there is a disease in the canopy according to the first spectral data and the canopy disease spectral coefficient.

[0021] As a further solution of the present invention, the disease spectral coefficient calculation unit is used to calculate the canopy disease spectral coefficient according to the second spectral data and the third spectral data. Specifically: obtain the second reflection coefficient based on the second spectral data, obtain the third reflection coefficient based on the third spectral data, arrange the second reflection coefficient in ascending order of wavelength to obtain the second reflection sequence, arrange the third reflection coefficient in ascending order of wavelength to obtain the third reflection sequence, import the second reflection sequence and the third reflection sequence into the canopy disease spectral coefficient calculation model to calculate the canopy disease spectral coefficient. The formula of the canopy disease spectral coefficient calculation model is:

[0022] ;

[0023] In the formula: is the canopy disease spectral coefficient, is the total number of wavelength points, is the weight factor of the r-th wavelength point, is the reflection coefficient corresponding to the r-th wavelength point in the second reflection sequence, is the reflection coefficient corresponding to the r-th wavelength point in the third reflection sequence.

[0024] As a further solution of the present invention, the secondary leaf spectral analysis unit is used to secondarily analyze whether there is a disease in the tree crown according to the first spectral data and the spectral coefficient of the tree crown disease. Specifically, the first reflection coefficients are arranged in ascending order of wavelength to obtain the first reflection sequence, and a secondary tree crown disease analysis model is constructed according to the first reflection sequence and the spectral coefficient of the tree crown disease to secondarily analyze whether there is a disease in the tree crown. The formula of the secondary tree crown disease analysis model is:

[0025] ;

[0026] In the formula: is the discriminant value of the secondary leaf disease, is the spectral coefficient of the tree crown disease, is the reflection coefficient corresponding to the r-th wavelength point in the first reflection sequence, is the average value of the reflection coefficients corresponding to each wavelength point in the first reflection sequence, is the total number of wavelength points.

[0027] The technical effects and advantages of a forestry resource data comprehensive management system of the present invention: By obtaining the spectral data and image data of the tree crown, extracting the spectral data and image data of the tree crown in the pest state and the healthy state, initially analyzing whether there is a disease in the leaves according to the pixel values extracted from the image data, and finally analyzing whether there is a disease in the tree crown according to the spectral data, it helps to improve the comprehensiveness and accuracy of disease detection, reduce misjudgment and missed judgment, and the dual analysis mechanism ensures sensitive identification of early diseases, and can provide data support for timely intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a schematic structural diagram of a forestry resource data comprehensive management system provided by the present invention;

[0029] Figure 2 is a schematic diagram showing the changes and characteristics of the tree crown from the healthy state to different disease stages provided by the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0030] Next, the technical solutions in the present invention will be clearly and completely described in conjunction with the drawings in the present invention. Obviously, the described technical solutions are only a part of the present invention, rather than all of them. All other technical solutions obtained by those of ordinary skill in the art based on the technical solutions in the present invention without making creative efforts fall within the scope of protection of the present invention.

[0031] Figure 1 is a schematic structural diagram of a forestry resource data comprehensive management system provided by the present invention. As Figure 1As shown in the figure, a comprehensive forestry resource data management system includes a multi-dimensional data acquisition module, a diseased leaf distribution recognition module, and a dynamic disease management module; the multi-dimensional data acquisition module is respectively connected to the diseased leaf distribution recognition module and the dynamic disease management module, and the diseased leaf distribution recognition module is connected to the dynamic disease management module;

[0032] The multi-dimensional data acquisition module is used to obtain the spectral data and image data of the tree crown;

[0033] The diseased leaf distribution recognition module is used to preliminarily analyze whether there are diseases on the leaves according to the image data;

[0034] The dynamic disease management module is used to finally analyze whether there are diseases on the tree crown according to the spectral data.

[0035] Specifically, the multi-dimensional data acquisition module includes a multispectral camera, a primary data extraction unit, and a secondary data extraction unit;

[0036] The multispectral camera is used to obtain the first spectral data and the first image data of the tree crown within the target range in real time;

[0037] The primary data extraction unit is used to obtain the second spectral data of the historically pest-infected tree crown and the third spectral data of the historically pest-free tree crown;

[0038] The secondary data extraction unit is used to obtain the second image data of the historically pest-infected tree crown and the third image data of the historically pest-free tree crown.

[0039] It should be noted that the spectral data includes the leaf reflectance coefficients corresponding to different wavelengths.

[0040] The multispectral camera obtains the first spectral data and the first image data of the tree canopy within the target range in real time, reflecting the health status of the current tree canopy, which helps to obtain the latest tree canopy information in a timely manner, supports dynamic monitoring and early disease warning, provides high-precision real-time reflectance coefficients, and captures the spectral characteristics of the tree canopy; the primary and secondary data extraction units respectively provide the spectral data and image data of the pest-infected and pest-free tree canopies in history, constituting a comparison benchmark. A benchmark model is established through historical data to facilitate the analysis of the deviation of real-time data, support machine learning or statistical modeling based on historical data, and improve the accuracy of disease identification; the spectral data includes the leaf reflectance coefficients corresponding to different wavelengths, which can quantify the reflection characteristics of the tree canopy in specific spectral bands. The spectral data is sensitive to the chlorophyll content, water content, and health status of the leaves, and can identify the early changes caused by diseases; the image data provides the spatial distribution information of the tree canopy, compensating for the lack of local features of the spectral data. The disease area is located through image analysis to clarify the disease range and distribution. Combining spectral characteristics with spatial image features supports more accurate disease diagnosis; the primary data extraction unit obtains the spectral data of the pest-infected and healthy tree canopies in history, forms a comparison sequence based on different wavelengths, and provides a reference curve for the spectral characteristics of pests, facilitating the analysis of the spectral shift of real-time data and supporting the training of machine learning models based on spectral reflectance differences to improve the disease identification efficiency; the secondary data extraction unit provides the image data of the pest-infected and healthy tree canopies, forms a historical image library, supports the deep learning model to extract spatial features, and identifies the disease patterns of the tree canopy. Integrating spectral and image features improves the accuracy of disease classification and location; by fusing real-time spectral and image data, and historical pest-infected and healthy data, a complete data link is constructed, providing strong technical support for the accurate identification and dynamic monitoring of tree canopy diseases. The spectral data, as the core, quantifies the changes in reflectance coefficients at different wavelengths, and together with the image data and historical comparison, realizes early, intelligent, and efficient disease diagnosis.

[0041] Specifically, the disease leaf distribution recognition module includes a first data acquisition unit, a data processing unit, a disease leaf distribution analysis unit, a non-disease leaf distribution analysis unit, and a dynamic leaf distribution analysis unit; the first data acquisition unit is respectively connected to the data processing unit, the disease leaf distribution analysis unit, the non-disease leaf distribution analysis unit, and the dynamic leaf distribution analysis unit, and the disease leaf distribution analysis unit and the non-disease leaf distribution analysis unit are respectively connected to the dynamic leaf distribution analysis unit.

[0042] The first data acquisition unit is used to extract the first image data, the second image data, and the third image data;

[0043] The data processing unit is used to respectively perform segmentation processing on the first image data, the second image data, and the third image data and extract their pixel values to obtain the first pixel data, the second pixel data, and the third pixel data;

[0044] The diseased leaf distribution analysis unit is used to analyze the diseased leaf distribution coefficient according to the second pixel data;

[0045] The non-diseased leaf distribution analysis unit is used to analyze the non-diseased leaf distribution coefficient according to the third pixel data;

[0046] The dynamic leaf distribution analysis unit is used to preliminarily analyze whether there is a disease in the leaf based on the first pixel data, the diseased leaf distribution coefficient, and the non-diseased leaf distribution coefficient.

[0047] Specifically, the diseased leaf distribution analysis unit is used to analyze the diseased leaf distribution coefficient according to the second pixel data. Specifically, the first pixel sequence is obtained by arranging the second pixel data in ascending order, the maximum value in the first pixel sequence is extracted as the first pixel value, the minimum value in the first pixel sequence is extracted as the second pixel value, the mean value of the second pixel data is calculated based on the first pixel sequence to obtain the first pixel distribution coefficient, and the diseased leaf distribution coefficient calculation model is constructed according to the second pixel data, the first pixel value, the second pixel value, and the first pixel distribution coefficient to calculate the diseased leaf distribution coefficient. The calculation formula of the diseased leaf distribution coefficient calculation model is:

[0048] ;

[0049] In the formula: is the diseased leaf distribution coefficient, is the number of the second pixel data in the first pixel sequence, is the i-th second pixel data in the first pixel sequence, is the first pixel distribution coefficient, is the first pixel value, is the second pixel value.

[0050] The diseased leaf distribution recognition module decomposes the task into multiple functional units, including data acquisition, processing, distribution analysis, and dynamic analysis, which improves the efficiency of task allocation. Each unit focuses on a specific function, reducing error accumulation. It acquires the first image data (real-time monitoring), the second image data (disease history), and the third image data (healthy history), dynamically combines real-time data with historical data, supports more flexible disease analysis, provides diverse input data sources, and enhances the applicability of the system. By integrating the first pixel data, the diseased leaf distribution coefficient, and the non-diseased leaf distribution coefficient, it dynamically determines whether there is a disease in the leaf, achieving real-time dynamic analysis of the diseased leaf distribution. The diseased leaf distribution analysis unit extracts the optical or pixel characteristics of the diseased leaves, calculates the diseased leaf distribution coefficient, and quantifies the disease degree through pixel-level distribution characteristics, avoiding subjective judgment or local errors. The non-diseased leaf distribution analysis unit extracts the characteristics of healthy leaves, compares them with the distribution coefficient of diseased leaves, provides a healthy benchmark, improves the accuracy of disease recognition, and supports fine-grained disease classification and distribution feature analysis. By fusing real-time data with historical data, quantifying the distribution characteristics of diseased leaves, and dynamically judging the disease status, its modular architecture and scientific formula for the diseased leaf distribution coefficient provide efficient and accurate disease diagnosis capabilities, enabling precise positioning, dynamic monitoring, and intelligent management of diseases, which is of great significance in disease prevention and control and ecological protection.

[0051] Specifically, the non-diseased leaf distribution analysis unit is used to analyze the non-diseased leaf distribution coefficient according to the third pixel data. Specifically, it arranges the third pixel data in ascending order to obtain the second pixel sequence, extracts the maximum value in the second pixel sequence as the third pixel value, extracts the minimum value in the second pixel sequence as the fourth pixel value, calculates the mean value of the third pixel data based on the second pixel sequence to obtain the second pixel distribution coefficient, and constructs a diseased leaf distribution coefficient calculation model according to the third pixel data, the fourth pixel value, the third pixel value, and the second pixel distribution coefficient to calculate the non-diseased leaf distribution coefficient. The calculation formula of the non-diseased leaf distribution coefficient calculation model is:

[0052] ;

[0053] In the formula: is the non-diseased leaf distribution coefficient, is the number of third pixel data in the second pixel sequence, is the j-th third pixel data in the second pixel sequence, is the second pixel distribution coefficient, is the third pixel value, is the fourth pixel value.

[0054] Analyze the change trend of healthy leaves through the statistical characteristics of the distribution of non-diseased leaves, assist in early disease detection, provide a control for the disease leaf distribution coefficient, and enhance the reliability of disease diagnosis; dynamically compare the non-diseased leaf distribution coefficient with the disease leaf distribution coefficient, which helps to identify the distribution differences between diseased areas and healthy areas, and strengthen the ability to identify abnormal distributions of diseased leaves. Especially in the early disease or mild disease stage, through the statistical benchmark of non-diseased distribution, further refine the disease diagnosis results, such as distinguishing the subtle differences between mild diseases and healthy leaves; the spatial distribution of the second pixel sequence is sorted through pixel data to analyze the overall spatial distribution characteristics of non-diseased leaves, provide the spatial distribution characteristics of non-diseased areas, provide a reference for the precise positioning of subsequent diseased areas, support the construction of a distribution comparison map of healthy and diseased leaves, and intuitively display the health status of the tree crown; the non-diseased leaf distribution system captures subtle changes in leaf status through the normalization and deviation analysis of pixel data. Especially when pixel values fluctuate due to environmental changes (such as light and humidity), the normalization process of the formula can reduce the risk of misjudgment and enhance the sensitivity of disease diagnosis. Even if there are no obvious disease symptoms on the leaves, potential abnormalities can be identified, improving the system's adaptability to complex environments (such as uneven light and background interference); the real-time calculation of the non-diseased leaf distribution coefficient supports the dynamic monitoring of the change trend of healthy leaves. When the non-diseased leaf distribution coefficient changes significantly, an alarm can be triggered in combination with the disease leaf distribution coefficient to dynamically track the distribution changes of healthy leaves, providing data support for long-term monitoring. By normalizing the pixel data of healthy leaves and analyzing the distribution characteristics, the non-diseased leaf distribution coefficient is scientifically calculated. Its benefits include providing a distribution benchmark for healthy leaves, supporting the precise positioning of diseased areas, dynamically feedbacking the health status, improving the diagnosis accuracy and sensitivity, and optimizing the disease prediction model, which is of great significance in disease monitoring, early warning, and system adaptability.

[0055] Specifically, the dynamic leaf distribution analysis unit is used to analyze whether there is a disease in the leaf according to the first pixel data, the disease leaf distribution coefficient, and the non-diseased leaf distribution coefficient. Specifically: extract the first pixel data, obtain the third pixel sequence by arranging the first pixel data in ascending order, construct a primary leaf disease discrimination model based on the third pixel sequence, the disease leaf distribution coefficient, and the non-diseased leaf distribution coefficient, and preliminarily judge whether there is a disease in the leaf. The formula of the primary leaf disease discrimination model is:

[0056] ;

[0057] In the formula: is the primary leaf disease discrimination value, is the non-diseased leaf distribution coefficient, is the disease leaf distribution coefficient, is the number of the first pixel data in the third pixel sequence, is the q-th first pixel data in the third pixel sequence, is the mean value of the first pixel data in the third pixel sequence;

[0058] Obtain the primary leaf disease discrimination value and the preset primary leaf disease discrimination threshold. If the primary leaf disease discrimination value is greater than or equal to the preset primary leaf disease discrimination threshold, it is preliminarily determined that the leaf has a disease; if the primary leaf disease discrimination value is less than the preset primary leaf disease discrimination threshold, it is preliminarily determined that the leaf does not have a disease.

[0059] By integrating the first pixel data, the disease leaf distribution coefficient, and the non-disease leaf distribution coefficient, the primary discrimination model is used to achieve a preliminary determination of the leaf disease state. By comprehensively considering the global distribution characteristics (disease and non-disease coefficients) and local distribution characteristics (the first pixel data), the fusion of multi-dimensional features is realized, enhancing the adaptability of the disease discrimination model to complex distribution situations; comparing the disease distribution characteristics with the healthy benchmark, improving the diagnostic accuracy, strengthening the sensitivity to abnormal disease distribution characteristics, especially in the early stage of disease or mild disease stage, improving the discrimination ability for complex disease distributions (such as the mixture of disease and healthy areas); based on the dynamic calculation of the first pixel data, the third pixel sequence is arranged in ascending order, analyzing the local distribution characteristics of the first pixel data, using the mean deviation to measure the dynamic change of the pixel data, which can quickly capture the abnormal fluctuations in the real-time data, adapt to the dynamic monitoring requirements, and update the disease discrimination value in real time to provide support for the dynamic tracking of the crown health status; through the relative change (rather than the absolute value) of the pixel value, the influence of light change or noise on the monitoring result is reduced, enhancing the anti-interference ability of the system, adapting to complex environmental conditions, improving the sensitivity to mild diseases or potential diseases, and avoiding missed judgments.

[0060] Specifically, the dynamic disease management module includes a second data acquisition unit, a spectral curve drawing unit, a disease spectral coefficient calculation unit, and a secondary leaf spectral analysis unit; the second data acquisition unit is respectively connected to the spectral curve drawing unit, the disease spectral coefficient calculation unit, and the secondary leaf spectral analysis unit, and the disease spectral coefficient calculation unit is connected to the secondary leaf spectral analysis unit;

[0061] The second data acquisition unit is used to extract the first spectral data, the second spectral data, and the third spectral data;

[0062] The spectral curve drawing unit is used to draw spectral curves according to the first spectral data, the second spectral data, and the third spectral data respectively;

[0063] The disease spectral coefficient calculation unit is used to calculate the crown disease spectral coefficient according to the second spectral data and the third spectral data;

[0064] The secondary leaf spectral analysis unit is used to secondarily analyze whether there is a disease in the tree crown according to the first spectral data and the tree crown disease spectral coefficient.

[0065] Specifically, the disease spectral coefficient calculation unit is used to calculate the tree crown disease spectral coefficient according to the second spectral data and the third spectral data. Specifically: obtain the second reflection coefficient based on the second spectral data, obtain the third reflection coefficient based on the third spectral data, arrange the second reflection coefficient in ascending order of wavelength to obtain the second reflection sequence, arrange the third reflection coefficient in ascending order of wavelength to obtain the third reflection sequence, import the second reflection sequence and the third reflection sequence into the tree crown disease spectral coefficient calculation model to calculate the tree crown disease spectral coefficient. The formula of the tree crown disease spectral coefficient calculation model is:

[0066] ;

[0067] In the formula: is the tree crown disease spectral coefficient, is the total number of wavelength points, is the weight factor of the r-th wavelength point, is the reflection coefficient corresponding to the r-th wavelength point in the second reflection sequence, is the reflection coefficient corresponding to the r-th wavelength point in the third reflection sequence.

[0068] Through the collaborative work of multiple units, comprehensively analyze the spectral characteristics and dynamic changes of canopy diseases, and obtain the first spectral data (monitoring target), the second spectral data (historical diseased canopies), and the third spectral data (historical healthy canopies) in real time. Integrate the real-time data and historical data to provide a basis for dynamic analysis, and support the timeliness and accuracy of disease monitoring; the spectral curve plotting unit plots the healthy, diseased, and real-time monitoring spectral curves respectively, intuitively showing the differences in spectral characteristics. By comparing the curves, quickly identify the changing trends of the spectral characteristics of diseases, which is convenient for analysts to visually judge the disease distribution and severity; the disease spectral coefficient calculation formula synthesizes the differences at different wavelength points, quantifies the overall changes in the spectral characteristics of diseases, and assigns higher weights to sensitive wavelengths (such as the 680–750 nm red edge band and the 750–900 nm near-infrared band) to improve the sensitivity of disease detection; the secondary leaf spectral analysis unit integrates the first spectral data monitored in real time and the canopy disease spectral coefficient to further analyze the diseases. The secondary analysis can refine the disease detection results and reduce misjudgments (such as spectral anomalies caused by changes in light conditions or background interference). Combine the real-time spectral data with the historical disease benchmark data to improve the reliability and stability of diagnosis; visually compare the real-time spectral data with the healthy and diseased spectral data. The spectral characteristic changes caused by diseases (such as red edge shift and near-infrared reflection decrease) can be directly presented through curve differences, supporting the early detection of diseases, especially when the leaves have not significantly fallen off in the initial stage of the disease (such as the slight chlorosis stage); according to the reflection differences at different wavelengths, provide global quantitative indicators of the spectral characteristics of diseases, which can refine the classification of disease types and severity, support the analysis of disease distribution characteristics by wavelength, and optimize the prevention and control strategies.

[0069] Specifically, the secondary leaf spectral analysis unit is used to secondarily analyze whether there are diseases in the canopy according to the first spectral data and the canopy disease spectral coefficient. Specifically: arrange the first reflection coefficients in ascending order of wavelength to obtain the first reflection sequence, and construct a secondary canopy disease analysis model according to the first reflection sequence and the canopy disease spectral coefficient to secondarily analyze whether there are diseases in the canopy. The formula of the secondary canopy disease analysis model is:

[0070] ;

[0071] In the formula: is the secondary leaf disease discrimination value, is the canopy disease spectral coefficient, is the reflection coefficient corresponding to the r-th wavelength point in the first reflection sequence, is the mean value of the reflection coefficients corresponding to each wavelength point in the first reflection sequence, is the total number of wavelength points;

[0072] Extract the discriminant value of the secondary leaf disease, and compare the discriminant value of the secondary leaf disease with the preset discriminant threshold of the secondary leaf disease. If the discriminant value of the secondary leaf disease is greater than or equal to the preset discriminant threshold of the secondary leaf disease, it is determined that the leaf has a disease; if the discriminant value of the secondary leaf disease is less than the preset discriminant threshold of the secondary leaf disease, the leaf has no disease.

[0073] Such as Figure 2 The change diagram of the tree crown from a healthy state to different disease stages shown, such as Figure 2 In (a), the change of the tree crown from a healthy state to different disease stages is shown. Among them, the leaves in A1 are in a healthy state, the leaves show bright green, representing the best state of the tree being healthy and disease-free. The leaves in A2 are in the green attack stage, starting to lose green but not significantly discolored. The leaves in A3 are in the red attack stage, gradually turning red and the disease is obvious. The leaves in A4 are in the gray attack stage. At this time, the leaves wither further and the spectral reflectance decreases significantly; such as Figure 2 In (b), the spectral reflectance characteristics of the tree crown in a healthy state and different disease stages are shown. The spectral range covers 400–900 nm, including visible light (400–700 nm) and near-infrared light (700–900 nm). Different color curves correspond to different tree crown states. Among them, the green curve is the spectral reflectance curve in the A1 state, the yellow curve is the spectral reflectance curve in the A2 state, the red curve is the spectral reflectance curve in the A3 state, and the gray curve is the spectral reflectance curve in the A4 state. The reflectance of the tree crown in A1 is the strongest at green light (near 500 nm) because healthy leaves contain a large amount of chlorophyll and green light is partially reflected. The reflectance of other states gradually decreases with the aggravation of the disease. The red edge (680–750 nm) drifts at the initial stage of the disease (A2), and the drift is more significant in A3 and A4. The disease in the near-infrared (750–900 nm) causes the reflectance to decrease, especially in A4.

[0074] By utilizing the first spectral data and the spectral coefficients of canopy diseases, further disease analysis is carried out based on the distribution characteristics of the reflection coefficients. By integrating the disease spectral coefficients and the real-time spectral data, misjudgments or missed judgments caused by single analysis are reduced. Considering the global distribution characteristics and local abnormal characteristics of wavelength points, the accuracy of diagnosis is enhanced. The degree of deviation between the first reflection sequence and the mean captures the abnormal changes in the canopy spectrum. The disease spectral coefficients weigh the contributions of each wavelength point to the disease characteristics. Even if the spectral changes caused by early diseases (such as slight chlorosis of leaves) are small, the secondary analysis model can still sensitively identify the disease characteristics and improve the adaptability to complex spectral distributions (such as differences in lighting conditions or background interference). The overall degree of deviation of each wavelength point in the first reflection sequence from the mean reflects the global consistency of the canopy spectrum. By combining the global and local characteristics, the detection ability for different types of diseases (such as local spot diseases or overall discoloration diseases) is improved, and the robustness of the system to data noise (such as changes in acquisition conditions) is enhanced. Through the secondary analysis of the first spectral data and the disease spectral coefficients, it is possible to quickly determine whether there are diseases in the canopy, facilitate a quick response to the risk of disease spread, dynamically track the progress trend of canopy diseases, and support long-term monitoring and management.

[0075] In the embodiments of the present invention, by obtaining the spectral data and image data of the canopy, and extracting the spectral data and image data of the canopy in the pest state and the healthy state, initially analyzing whether there are diseases in the leaves according to the pixel values extracted from the image data, and finally analyzing whether there are diseases in the canopy according to the spectral data, it helps to improve the comprehensiveness and accuracy of disease detection, reduce misjudgments and missed judgments, and the dual analysis mechanism ensures the sensitive identification of early diseases, and can provide data support for timely intervention.

[0076] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0077] Finally: The above is only the preferred solution of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An integrated forestry resource data management system, including a multi-dimensional data collection module, a diseased leaf distribution recognition module, and a dynamic disease management module, characterized in that, The multi-dimensional data acquisition module is used to obtain the spectral data and image data of the tree crown; the multi-dimensional data acquisition module includes a multispectral camera; the multispectral camera is used to obtain the first spectral data and the first image data of the tree crown within the target range in real time; The multi-dimensional data acquisition module further includes a primary data extraction unit and a secondary data extraction unit; the primary data extraction unit is used to obtain the second spectral data of the historically pest-infected tree crown and the third spectral data of the historically pest-free tree crown; the secondary data extraction unit is used to obtain the second image data of the historically pest-infected tree crown and the third image data of the historically pest-free tree crown; The diseased leaf distribution recognition module includes a first data acquisition unit, a data processing unit, a diseased leaf distribution analysis unit, a non-diseased leaf distribution analysis unit, and a dynamic leaf distribution analysis unit; the first data acquisition unit is used to extract the first image data, the second image data, and the third image data; the data processing unit is used to perform segmentation processing on the first image data, the second image data, and the third image data respectively and extract their pixel values to obtain the first pixel data, the second pixel data, and the third pixel data; The diseased leaf distribution analysis unit is used to analyze the diseased leaf distribution coefficient according to the second pixel data; the non-diseased leaf distribution analysis unit is used to analyze the non-diseased leaf distribution coefficient according to the third pixel data; The dynamic leaf distribution analysis unit is used to preliminarily analyze whether there is a disease in the leaf based on the first pixel data, the diseased leaf distribution coefficient, and the non-diseased leaf distribution coefficient, extract the first pixel data, obtain the third pixel sequence by arranging the first pixel data in ascending order, construct a primary leaf disease discrimination model based on the third pixel sequence, the diseased leaf distribution coefficient, and the non-diseased leaf distribution coefficient, and preliminarily determine whether there is a disease in the leaf. The formula of the primary leaf disease discrimination model is: ； Where: is the discrimination value of primary leaf diseases, is the distribution coefficient of non-diseased leaves, is the distribution coefficient of diseased leaves, is the number of the first pixel data in the third pixel sequence, is the q-th first pixel data in the third pixel sequence, is the mean value of the first pixel data in the third pixel sequence; The dynamic disease management module includes a second data acquisition unit, a spectral curve drawing unit, a disease spectral coefficient calculation unit, and a secondary leaf spectral analysis unit; The second data acquisition unit is used to extract the first spectral data, the second spectral data, and the third spectral data; The spectral curve drawing unit is used to draw spectral curves according to the first spectral data, the second spectral data, and the third spectral data respectively; The disease spectral coefficient calculation unit is used to calculate the tree crown disease spectral coefficient according to the second spectral data and the third spectral data; The secondary leaf spectral analysis unit is used to perform secondary analysis on whether there is a disease in the tree crown according to the first spectral data and the tree crown disease spectral coefficient.

2. The integrated forestry resource data management system according to claim 1, wherein The diseased leaf distribution analysis unit is used to analyze the diseased leaf distribution coefficient according to the second pixel data. Specifically, the second pixel data is sorted in ascending order to obtain the first pixel sequence. The maximum value in the first pixel sequence is extracted as the first pixel value, and the minimum value in the first pixel sequence is extracted as the second pixel value. The mean value of the second pixel data is calculated based on the first pixel sequence to obtain the first pixel distribution coefficient. A diseased leaf distribution coefficient calculation model is constructed according to the second pixel data, the first pixel value, the second pixel value, and the first pixel distribution coefficient to calculate the diseased leaf distribution coefficient. The calculation formula of the diseased leaf distribution coefficient calculation model is: ； Wherein: is the distribution coefficient of diseased leaves, is the number of second pixel data in the first pixel sequence, is the i-th second pixel data in the first pixel sequence, is the first pixel distribution coefficient, is the first pixel value, is the second pixel value.

3. The integrated forestry resource data management system according to claim 2, characterized in that, The non-diseased leaf distribution analysis unit is used to analyze the non-diseased leaf distribution coefficient according to the third pixel data. Specifically, the third pixel data is sorted in ascending order to obtain the second pixel sequence. The maximum value in the second pixel sequence is extracted as the third pixel value, and the minimum value in the second pixel sequence is extracted as the fourth pixel value. The mean value of the third pixel data is calculated based on the second pixel sequence to obtain the second pixel distribution coefficient. A diseased leaf distribution coefficient calculation model is constructed according to the third pixel data, the fourth pixel value, the third pixel value, and the second pixel distribution coefficient to calculate the non-diseased leaf distribution coefficient. The calculation formula of the non-diseased leaf distribution coefficient calculation model is: ； Wherein: is the non-disease leaf distribution coefficient, is the number of the third pixel data in the second pixel sequence, is the j-th third pixel data in the second pixel sequence, is the second pixel distribution coefficient, is the third pixel value, is the fourth pixel value.

4. The integrated forestry resource data management system according to claim 1, characterized in that The diseased crown spectral coefficient calculation unit is used to calculate the diseased crown spectral coefficient according to the second spectral data and the third spectral data. Specifically, the second reflection coefficient is obtained based on the second spectral data, and the third reflection coefficient is obtained based on the third spectral data. The second reflection coefficient is sorted in ascending order of wavelength to obtain the second reflection sequence, and the third reflection coefficient is sorted in ascending order of wavelength to obtain the third reflection sequence. The second reflection sequence and the third reflection sequence are imported into the diseased crown spectral coefficient calculation model to calculate the diseased crown spectral coefficient. The formula of the diseased crown spectral coefficient calculation model is: ； In the formula: is the spectral coefficient of canopy diseases, is the total number of wavelength points, is the weighting factor of the r-th wavelength point, is the reflection coefficient corresponding to the r-th wavelength point in the second reflection sequence, is the reflection coefficient corresponding to the r-th wavelength point in the third reflection sequence.

5. The comprehensive forestry resource data management system according to claim 1, characterized in that The secondary leaf spectral analysis unit is used to secondarily analyze whether there is a disease in the crown according to the first spectral data and the diseased crown spectral coefficient. Specifically, the first reflection coefficient is sorted in ascending order of wavelength to obtain the first reflection sequence. A secondary diseased crown analysis model is constructed according to the first reflection sequence and the diseased crown spectral coefficient to secondarily analyze whether there is a disease in the crown. The formula of the secondary diseased crown analysis model is: ； Wherein: is the discriminant value of the secondary leaf disease, is the spectral coefficient of the crown disease, is the reflection coefficient corresponding to the r-th wavelength point in the first reflection sequence, is the mean value of the reflection coefficients corresponding to each wavelength point in the first reflection sequence, is the total number of wavelength points.

Images