Intelligent soilless culture environment monitoring system and method based on big data

Abstract

The application discloses an intelligent soilless cultivation environment monitoring system and method based on big data, and relates to the technical field of soilless cultivation environment monitoring.The application comprises the following steps: dividing a soilless cultivation environment area into a plurality of sub-areas according to equal area size, and setting the area of each sub-area according to specific cultivation environment and monitoring requirements to ensure that the plants in each sub-area are representative.The application uses a convolutional neural network to extract features from collected image data and predict plant diseases and insect pests, which can not only accurately determine the health status of each sub-area, but also automatically adjust the focus of the camera when a potential plant disease and insect pest area is found, focus on details and shorten the focal length, so that early signs of plant diseases and insect pests can be clearly captured, greatly improving the accuracy of plant disease and insect pest identification, and timely discovering plant disease and insect pest hazards to ensure that plant diseases and insect pests are intervened in the early stage, effectively preventing the spread of plant diseases and insect pests, and protecting the health of crops in the entire cultivation environment.

Description

Technical Field

[0001] This invention relates to the field of soilless cultivation environment monitoring technology, specifically to an intelligent soilless cultivation environment monitoring system and method based on big data. Background Technology

[0002] Big data-driven intelligent hydroponics environment monitoring utilizes advanced sensor technology, the Internet of Things (IoT), and big data analytics to monitor and adjust various environmental parameters in the hydroponics system in real time, such as temperature, humidity, light intensity, air circulation, and nutrient solution concentration. By collecting this data and combining it with machine learning algorithms and intelligent predictive models, the cultivation environment can be dynamically adjusted to achieve precise control and optimization, ensuring plants grow under optimal conditions. This intelligent monitoring system not only improves crop yield and quality but also identifies potential problems through data analysis, providing early warnings, reducing resource waste, and enhancing the sustainability and efficiency of agricultural production.

[0003] The existing technology has the following shortcomings:

[0004] In intelligent hydroponic cultivation environment monitoring, existing technologies typically rely on a single camera to monitor plant leaves over a large area. However, due to the limitations of the camera's field of view, its focal length is usually large, resulting in only a wide-area image being captured. This can potentially mask or overlook early signs of pests and diseases. Especially at different growth stages of plants, leaf appearance varies significantly, and images acquired from a single perspective cannot accurately distinguish the specific diseases affecting each plant. Therefore, early symptoms of pests and diseases are often not identified in time, delaying intervention and potentially leading to the spread of pests and diseases and impacting the health of crops throughout the entire cultivation area.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent hydroponic cultivation environment monitoring system and method based on big data. By dividing the cultivation area into multiple sub-regions and collecting image data, the representativeness of plant characteristics within each sub-region is ensured, thereby improving the accuracy of pest and disease monitoring. Combining convolutional neural networks for feature extraction and pest and disease prediction enables precise assessment of the health status of each sub-region. When potential pest and disease areas are identified, the camera focus is automatically adjusted to focus on details and shorten the focal length, improving the accuracy of capturing early signs of pests and diseases. This intelligent and dynamically adjusted monitoring method ensures timely identification and intervention of pests and diseases, preventing their spread and protecting crop health, thus solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring the intelligent soilless cultivation environment based on big data, comprising the following steps:

[0008] The hydroponic environment area is divided into several sub-areas of equal size. The area of ​​each sub-area is set according to the specific cultivation environment and monitoring requirements to ensure that the plants in each sub-area are representative.

[0009] Within each sub-region, images of plant leaves within that region are automatically collected by cameras and transmitted in real time to the data processing system for subsequent pest and disease detection and analysis.

[0010] The image data collected from each sub-region is summarized, and features reflecting potential pests and diseases are extracted from the summarized data. After analyzing the extracted features, the analyzed features are input into a pre-trained convolutional neural network, which then predicts the pests and diseases in the sub-region.

[0011] Based on the prediction results of the convolutional neural network, the health status of each sub-region is classified into "potential pest and disease areas" and "normal growth areas".

[0012] For areas with normal growth, continue to monitor for pests and diseases while maintaining the initial state.

[0013] For sub-areas identified as potential pest and disease areas, the camera's control module automatically adjusts the camera's focus to concentrate on that sub-area. Once focused, the camera can more accurately capture the details of the plants within that sub-area.

[0014] After the camera focuses on the potential pest and disease area, the current focal length of the camera is shortened based on the prediction results of the convolutional neural network, so as to capture the details of the target area more clearly and identify the pests and diseases.

[0015] Preferably, the hydroponics environment area is divided into several sub-regions of equal size. The specific steps are as follows:

[0016] It is necessary to clarify the total area of ​​the monitoring area of ​​the entire hydroponics environment and the area ratio required for dividing it into sub-areas. Based on the shape of the environment, the layout of the cultivation system, the types of plants, and the utilization needs of the cultivation area, the size of each sub-area should be determined.

[0017] After determining the area of ​​the sub-region, select a region division method to divide the entire cultivation environment into several small squares according to an equal grid, with each small square being a sub-region.

[0018] Preferably, features reflecting potential pests and diseases are extracted from the aggregated data. These features include the degree of contrast between leaf veins and surrounding leaf tissues and the gradual color shift of the leaf color. After analyzing the degree of contrast between leaf veins and surrounding leaf tissues and the gradual color shift of the leaf color in each sub-region, leaf vein prominence reference values ​​and color shift gradual reference values ​​are generated. The leaf vein prominence reference values ​​quantify the intensity of the contrast between leaf veins and surrounding leaf tissues, and the color shift gradual reference values ​​quantify the degree of gradual shift of leaf color in time or space.

[0019] Preferably, the vein prominence reference value and color shift reference value generated after analyzing the contrast between the leaf vein and the surrounding leaf tissue and the progressive color shift change of the leaf color are input into a pre-trained convolutional neural network. Based on the convolutional neural network, a pest and disease index is generated, and the pest and disease index is used to predict the pest and disease situation in the sub-region.

[0020] Preferably, the specific steps for generating a vein prominence reference value after analyzing the contrast between leaf veins and surrounding leaf tissue in each sub-region are as follows:

[0021] The plant images of each sub-region are processed, and the spectral reflectance features of the leaves are extracted using multispectral imaging technology. A spectral contrast matrix between the vein region and the surrounding leaf tissue is constructed using spectral contrast values. The expression for constructing the spectral contrast matrix is ​​as follows: ,in: The leaf veins and surrounding leaf tissue at the pixel level Spectral contrast value, For pixels The reflectance value in the near-infrared band of the leaf vein region, pixel The reflectance of the green light band in the tissue surrounding the leaf veins. It is in the near-infrared band. It is in the green light band;

[0022] Based on the spectral contrast matrix, a positional weighting function and a dynamic threshold function are introduced to normalize the matrix, and the reference value for vein prominence is calculated. The expression for the calculation is as follows: ,in: This serves as a reference value for highlighting leaf veins, used to quantify the intensity of vein contrast across the entire sub-region. H , W These are the height and width of the image, respectively. This is a position weighting function that assigns weights to pixels based on the importance of the leaf region. Assign different weights, The dynamic threshold function, used to remove background noise from normal leaf vein contrast, is defined as follows: ,in For pixels Distance to the center of the blade and It is a parameter used to control threshold decay.

[0023] Preferably, the specific steps for generating a progressive color shift reference value after analyzing the progressive color shift changes of the leaf color in each sub-region are as follows:

[0024] By performing multi-dimensional color analysis on the leaf images within each sub-region, a color shift feature matrix is ​​extracted. This color shift feature matrix is ​​based on the leaf's position at time t and spatial location. The color change is calculated using the following expression: ,in, The color shift feature matrix represents the spatial location at time t. The degree of color shift, Indicates at wavelength Lower spatial position Color reflectance value, Indicates the previous time point Color reflectance value, This refers to the band weighting coefficient;

[0025] Using the color shift feature matrix and spatial shift weights, the color shift progressive reference value of the sub-region is calculated to quantify the degree of progressive shift of the leaf color in time or space. The calculation expression is as follows: ,in, This serves as a color shift reference value, quantifying the overall gradual shift of leaf color within a sub-region. Spatial offset weights are set based on the spatial concentration of leaf density and color offset within a sub-region. As a smoothing factor, to prevent when The denominator may be zero at times.

[0026] Preferably, the pest and disease index generated when predicting pests and diseases in a sub-region using a pre-trained convolutional neural network is compared and analyzed with a pre-set reference threshold for the pest and disease index, and the health status of each sub-region is classified. The specific classification is as follows:

[0027] If the pest and disease index generated in a sub-region is greater than the pest and disease index reference threshold, then the health status of the sub-region is classified as a potential pest and disease area.

[0028] If the pest and disease index generated in a sub-region is less than or equal to the pest and disease index reference threshold, then the health status of that sub-region is classified as a normal growth region.

[0029] Preferably, for sub-areas identified as potential pest and disease areas, automatic focusing is achieved through the camera's control module, specifically:

[0030] First, the pest and disease index generated by the convolutional neural network determines the location coordinates of the target sub-region, and these coordinates are used as the target point of the camera.

[0031] Next, by utilizing the camera's autofocus function and combining it with depth calculation algorithms, the camera's focal plane is dynamically adjusted to ensure that the camera accurately focuses on the target sub-region.

[0032] Preferably, after the camera focuses on the potential pest and disease area, the current focal length of the camera is shortened based on the prediction results of the convolutional neural network. The specific steps are as follows:

[0033] The pest and disease index generated by a convolutional neural network in this sub-region is collected. This index is then compared with a preset reference threshold. Based on the relationship between the two, a focal length difference factor is calculated. The expression for this calculation is as follows: ,in, This is the focal length difference factor, used in subsequent calculations of the adjusted focal length. This indicates that the pest and disease index for this sub-region was generated by a convolutional neural network. This serves as a reference threshold for the pest and disease index. An adjustment factor used to quantify the impact of the degree to which the pest and disease index exceeds the reference threshold on focal length shortening;

[0034] Based on the current focal length and the calculated focal length difference factor, calculate the adjusted target focal length, ensuring it remains within the camera's focal length adjustment range. The expression for the target focal length adjustment is: ,in, The adjusted target focal length is used to achieve sharper image capture. The current focal length of the camera. and This refers to the physical adjustment range of the camera's focal length, ensuring that the calculated target focal length does not exceed the camera's hardware capabilities. and These are the maximum and minimum focal lengths of the camera, respectively.

[0035] The camera's focal length is automatically adjusted based on the calculated target focal length. Simultaneously, the focus result is verified in real-time using an image sharpness index. If the focus effect does not meet the expected set value, it is fine-tuned according to the adjustment range. The sharpness index formula is as follows: ,in, Image sharpness index, indicating the clarity of the target area in the image. Current focal length F Next pixel Image intensity value at that location, Image gradient, used to quantize changes in pixel intensity. N This represents the total number of pixels within the target area.

[0036] The intelligent soilless cultivation environment monitoring system based on big data includes a sub-region division module, an image data acquisition module, a feature extraction and pest prediction module, a health status classification module, a normal area monitoring module, a potential pest focusing module, and a focal length optimization and fine recognition module.

[0037] The sub-region division module divides the hydroponic environment area into several sub-regions of equal size. The area of ​​each sub-region is set according to the specific cultivation environment and monitoring requirements to ensure that the plants in each sub-region are representative.

[0038] The image data acquisition module automatically acquires image data of plant leaves in each sub-region using a camera, and transmits it to the data processing system in real time for subsequent pest and disease detection and analysis.

[0039] The feature extraction and pest prediction module summarizes the image data collected from each sub-region, extracts features reflecting potential pests and diseases from the summary, analyzes the extracted features, and inputs the analyzed features into a pre-trained convolutional neural network to predict the pests and diseases in the sub-region.

[0040] The health status classification module, based on the prediction results of the convolutional neural network, classifies the health status of each sub-region into "potential pest and disease areas" and "normal growth areas".

[0041] The normal area monitoring module continues to monitor pests and diseases in normal growth areas, maintaining the initial state.

[0042] The potential pest and disease focusing module automatically adjusts the camera's focus through the camera's control module for sub-areas identified as potential pest and disease areas, so that the camera focuses on the sub-area. After the camera focuses, it will capture the details of the plants in the sub-area more accurately.

[0043] The focal length optimization and fine recognition module shortens the current focal length of the camera after the camera focuses on the potential pest and disease area, based on the prediction results of the convolutional neural network, so as to capture the details of the target area more clearly and identify pests and diseases.

[0044] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0045] This invention improves the accuracy of pest and disease monitoring by dividing the cultivation area into multiple sub-regions and collecting image data from each sub-region, ensuring that the plant characteristics within each sub-region are representative. Secondly, it utilizes convolutional neural networks to extract features and predict pests and diseases from the collected image data. This not only accurately assesses the health status of each sub-region but also automatically adjusts the camera's focus when potential pest and disease areas are detected, concentrating on details and shortening the focal length. This allows for the clear capture of early signs of pests and diseases. This intelligent and dynamically adjusted monitoring method significantly improves the accuracy of pest and disease identification, enabling timely detection of potential pest and disease problems and ensuring early intervention. This effectively prevents the spread of pests and diseases and safeguards the health of crops throughout the entire cultivation environment. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0047] Figure 1 This is a flowchart of the intelligent soilless cultivation environment monitoring method based on big data according to the present invention.

[0048] Figure 2 This is a schematic diagram of the modules of the intelligent soilless cultivation environment monitoring system based on big data of the present invention. Detailed Implementation

[0049] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0050] This invention provides, for example Figure 1 The intelligent hydroponics environment monitoring method based on big data, as shown, includes the following steps:

[0051] The hydroponic environment area is divided into several sub-areas of equal size. The area of ​​each sub-area is set according to the specific cultivation environment and monitoring requirements to ensure that the plants in each sub-area are representative.

[0052] The hydroponics environment is divided into several sub-regions of equal size. The specific steps are as follows:

[0053] It is necessary to determine the total area of ​​the monitoring zone for the entire hydroponics environment and the area ratio required for dividing it into sub-zones. The size of each sub-zone should be determined based on the shape of the environment, the layout of the cultivation system, the types of plants, and the utilization requirements of the cultivation area. The total area can be obtained by measuring or acquiring a layout map of the environment, and then the standard area of ​​each sub-zone can be calculated according to the requirements.

[0054] The purpose of this step is to determine the area of ​​each sub-region to ensure that the divided sub-regions can cover every part of the cultivation area. Appropriate sub-region size not only contributes to the efficiency of pest and disease monitoring but also ensures the rational allocation and management of resources during the monitoring process. This avoids situations where excessively large monitoring areas make pest and disease signs difficult to detect, or excessively small monitoring areas result in too much redundant data.

[0055] After determining the area of ​​the sub-region, a reasonable method for dividing the region should be selected. Common methods include regular grid division, uniform region division, or division based on crop type / growth stage. If a regular grid division is used, the entire cultivation environment is divided into several small squares using equal grids; each small square is a sub-region. If division is based on crop growth stage, adjustments can be made according to plant growth density and condition to identify areas requiring focused attention.

[0056] The purpose of this step is to ensure that the cultivation area is divided into multiple sub-regions reasonably and evenly so that each sub-region can be monitored independently. Different division methods can be used to customize the regional division based on factors such as different cultivation environments, crop types, and cultivation stages. This process ensures that the monitoring content of each sub-region is representative, avoiding over-division or uneven division, thereby guaranteeing the comprehensiveness and accuracy of subsequent pest and disease identification.

[0057] When monitoring intelligent hydroponics environments, the area of ​​each sub-region is not fixed but should be flexibly set according to the characteristics of the actual cultivation environment and monitoring needs. Different cultivation environments may have different sizes, layouts, and crop types, so the area of ​​each sub-region needs to be adjusted based on these factors. For example, if the cultivation environment is large, it may be necessary to divide it into multiple smaller sub-regions to ensure that each sub-region can be effectively monitored; while in a smaller cultivation environment, fewer sub-regions may be needed. At the same time, the required monitoring accuracy also affects the area of ​​the sub-regions. If higher monitoring accuracy is required, each sub-region may need to be divided into smaller areas to allow for more detailed observation and analysis of the plant leaves in each area. Conversely, if the monitoring requirements are lower, the sub-regions can be appropriately expanded, thereby reducing the number of monitoring devices and resource consumption.

[0058] By subdividing a large monitoring area into multiple smaller monitoring units, the limitations of a single perspective in clearly identifying pests and diseases are avoided. This subdivision allows for independent monitoring and analysis of plants in different areas, improving the accuracy of pest and disease monitoring. This subdivision effectively solves the monitoring ambiguity caused by a single perspective, ensuring that signs of pests and diseases in each area can be detected promptly.

[0059] Within each sub-region, images of plant leaves within that region are automatically collected by cameras and transmitted in real time to the data processing system for subsequent pest and disease detection and analysis.

[0060] This step is fundamental to pest and disease detection. The camera captures images in real time, providing raw data for subsequent feature extraction and analysis. By acquiring images at high frequency, subtle changes in plant leaves can be captured more accurately, especially in the early stages of pests and diseases.

[0061] The image data collected from each sub-region is summarized, and features reflecting potential pests and diseases are extracted from the summarized data. After analyzing the extracted features, the analyzed features are input into a pre-trained convolutional neural network, which then predicts the pests and diseases in the sub-region.

[0062] Features reflecting potential pests and diseases were extracted from the aggregated data. These features included the contrast between leaf veins and surrounding leaf tissues, and the gradual color shift of the leaf color. After analyzing the contrast between leaf veins and surrounding leaf tissues and the gradual color shift of the leaf color in each sub-region, leaf vein prominence reference values ​​and color shift progressive reference values ​​were generated. The leaf vein prominence reference values ​​quantified the intensity of the contrast between leaf veins and surrounding leaf tissues, measuring the clarity, color contrast, and texture difference of leaf veins relative to surrounding leaf tissues in the image. This reflects whether the plant has abnormally prominent vascular bundles or degenerate surrounding leaf tissues due to disease or stress. The color shift progressive reference values ​​quantified the degree of gradual shift of leaf color over time or space, analyzing the subtle changes in leaf color (such as a gradual transition from normal green to yellowish-green, brown, red, etc.), reflecting whether the plant has been affected by pests, diseases, or environmental stresses leading to chlorophyll degradation or abnormal pigment accumulation.

[0063] A significant contrast between the veins of a leaf within a specific sub-region and the surrounding leaf tissue usually indicates a potential threat of pests or diseases in that sub-region. This high contrast may be due to pathogens or pests damaging the normal structure of the leaf tissue, leading to degeneration, chlorosis, or necrosis of the surrounding mesophyll tissue, making the veins appear more prominent. This phenomenon may reflect plant damage from fungi, viruses, or pests, such as fungal infections causing mesophyll cell collapse, viral infections leading to chlorophyll degradation, or insect damage causing localized tissue damage. This characteristic is crucial for early pest and disease identification. Quantifying the vein contrast can provide preliminary risk warnings to monitoring systems, facilitating timely intervention and preventing the spread of pests and diseases.

[0064] The specific steps for generating vein prominence reference values ​​after analyzing the contrast between leaf veins and surrounding leaf tissue within each sub-region are as follows:

[0065] The plant images of each sub-region are processed, and the spectral reflectance features of the leaves are extracted using multispectral imaging technology. A spectral contrast matrix between the vein region and the surrounding leaf tissue is constructed using spectral contrast values. The spectral contrast matrix is ​​determined by each pixel in a specific wavelength band (such as near-infrared). and green light The reflectance ratio is calculated from the spectral contrast ratio, and the expression for constructing the spectral contrast matrix is: ,in: The leaf veins and surrounding leaf tissue at the pixel level Spectral contrast value, For pixels The reflectance value in the near-infrared band of the leaf vein region, pixel The reflectance of the green light band in the tissue surrounding the leaf veins. It is a near-infrared band used to detect the water content or structural integrity of plant internal tissues. This is the green light band, used to measure the distribution and health status of chlorophyll;

[0066] The main purpose of this step is to construct a fundamental matrix that can quantify the contrast intensity between leaf veins and surrounding leaf tissue by analyzing the differences in spectral characteristics. By selecting the appropriate bands, abnormal changes in leaf veins and surrounding tissues caused by pests, diseases, or stress can be captured more sensitively.

[0067] Based on the spectral contrast matrix, a positional weighting function and a dynamic threshold function are introduced to normalize the matrix, and the reference value for vein prominence is calculated. The expression for the calculation is as follows: ,in: This serves as a reference value for highlighting leaf veins, used to quantify the intensity of vein contrast across the entire sub-region. H , W These are the height and width of the image, respectively. This is a position weighting function that assigns weights to pixels based on the importance of the leaf region. Different weights are assigned (e.g., the central region of the blade has a greater weight than the edge region). The dynamic threshold function, used to remove background noise from normal leaf vein contrast, is defined as follows: ,in For pixels Distance to the center of the blade and It is a parameter used to control threshold decay;

[0068] This step generates a global reference value for vein prominence by comprehensively considering the spectral contrast between the veins and surrounding tissues and the importance of the leaf region. The introduction of position weighting function and dynamic threshold function helps to eliminate interference from normal contrast areas and focus on abnormally prominent vein areas, further improving the accuracy and reliability of pest and disease detection.

[0069] As shown by the vein prominence reference value, a higher value indicates a stronger contrast between the veins and surrounding leaf tissue within each sub-region. This typically suggests a more severe impact from pests, diseases, or other stresses on the leaf, thus reflecting a higher likelihood of pest or disease infestation in that sub-region. Conversely, a lower value indicates a lower contrast, with more normal physiological structure and color changes in the leaf, suggesting a lower likelihood of pest or disease infestation in that sub-region. The vein prominence reference value, by quantifying the degree of contrast, provides crucial information for early warning of pests and diseases in sub-regions, helping the system to promptly locate problem areas and implement intervention measures.

[0070] When the leaf color in a sub-region exhibits a gradual change, it usually indicates a potential threat of pests or diseases in that sub-region. A gradual color change refers to the process by which the leaf color transitions from normal green to abnormal hues such as yellowish-green, brown, gray, or red. This phenomenon is often an early sign of external stress in plants (such as pathogen infection, pest infestation, or adverse environmental conditions), and may reflect chlorophyll degradation, abnormal cell metabolism, or abnormal accumulation of secondary pigments (such as carotenoids and anthocyanins). By analyzing the gradual change trajectory of leaf color, early abnormal reactions in plants can be detected before pests or diseases have spread significantly, providing a basis for timely diagnosis and intervention, and preventing the potential for further damage.

[0071] The specific steps for generating a progressive color shift reference value after analyzing the progressive color shift changes of the leaves in each sub-region are as follows:

[0072] By performing multi-dimensional color analysis on the leaf images within each sub-region, a color shift feature matrix is ​​extracted. This color shift feature matrix is ​​based on the leaf's position at time t and spatial location. The color changes are calculated, focusing on reflecting the gradual trend of color change across different wavelengths. The calculation expression is as follows: ,in, The color shift feature matrix represents the spatial location at time t. The degree of color shift, Indicates at wavelength Lower spatial position Color reflectance values ​​can be acquired through multispectral imaging in the visible light (400-700 nm) and near-infrared (700-1100 nm) bands. Indicates the previous time point Color reflectance value, This is a band weighting coefficient used to emphasize the sensitivity of specific bands (such as red light and near-infrared) to pests and diseases;

[0073] By calculating the time shift of color in each band, the color shift feature matrix is ​​obtained. This reflects the cumulative effect of leaf color changes at different locations and time points. The focus is on extracting the trend characteristics of gradual color changes from green to yellow, brown, and red, thus laying the foundation for subsequent index generation.

[0074] Using the color shift feature matrix and spatial shift weights, the color shift progressive reference value of the sub-region is calculated to quantify the degree of progressive shift of the leaf color in time or space. The calculation expression is as follows: ,in, This serves as a reference value for color shift, quantifying the overall gradual shift of leaf color within a sub-region. A higher value indicates a more pronounced color change. Spatial offset weights are set based on the spatial concentration of leaf density and color offset within a sub-region. A higher weight indicates a greater impact of that location on the overall index. These weights can be calculated using leaf coverage and image texture variations, but specific limitations are not specified here. As a smoothing factor, to prevent when The denominator may be zero at times;

[0075] This step takes into account both the color changes over time and the weight distribution in space. By accumulating the color shifts point by point within a sub-region, an index reflecting the overall trend of change is generated. This index is used to determine whether there are any abnormal gradual color changes in the region (such as from green to yellow or brown), thereby identifying potential pest and disease risks and helping to quickly locate high-risk areas for in-depth analysis.

[0076] As shown by the color shift progressive reference value, the larger the value generated after analyzing the progressive color shift of leaves in each sub-region, the more significant the progressive color shift of the leaves in that sub-region has occurred, such as a gradual shift from green to yellow, brown, or other abnormal colors. This progressive color shift is usually an early physiological response of plants under pest and disease stress, such as chlorophyll degradation, abnormal cell metabolism, or pigment accumulation (e.g., carotenoids, anthocyanins). Therefore, a larger color shift progressive reference value indicates a potentially higher risk of pests and diseases in that area; conversely, a smaller value indicates less color change in the leaves or a near-normal state, with a lower risk of pests and diseases. Therefore, the color shift progressive reference value can serve as an important basis for assessing the health status of an area, helping to achieve early detection and precise intervention of pests and diseases.

[0077] The vein prominence reference value and color shift reference value generated after analyzing the contrast between the leaf veins and the surrounding leaf tissues and the progressive color shift of the leaf color are input into a pre-trained convolutional neural network. Based on the convolutional neural network, a pest and disease index is generated, and the pest and disease index is used to predict the pest and disease situation in the sub-region.

[0078] A pre-trained Convolutional Neural Network (CNN) is a deep learning model that has already undergone preliminary learning and optimization based on a large amount of relevant data. This type of CNN, trained on a large number of historical plant disease and pest images and related features (such as leaf vein prominence reference values ​​and color shift gradient reference values), is already capable of accurately identifying and classifying plant health status and disease / pest types. During training, the CNN extracts local patterns from images or feature data through multiple convolutional operations, such as changes in leaf vein texture and leaf color shift trends; pooling layers reduce the dimensionality of the data to retain key information; and finally, fully connected layers synthesize these features to achieve the prediction and quantification of disease and pest risk. Pre-training means that the model already has a high generalization ability for specific tasks (such as plant disease and pest identification) and can be directly applied to the prediction of new data without needing to be trained from scratch.

[0079] In pest and disease prediction tasks, such pre-trained models play two important roles: First, through multi-level feature extraction, they integrate input information such as leaf vein prominence reference values ​​and color shift gradient reference values ​​to capture latent patterns that may be related to pests and diseases. For example, convolutional neural networks can detect the non-linear relationship between abnormal leaf vein prominence and color shift, which may indicate the risk of viral infection, fungal invasion, or insect infestation. Second, with the help of pre-trained weights, the model can be quickly applied to pest and disease detection in different cultivation areas without retraining, thus saving computational resources and time. Furthermore, since convolutional neural networks have already learned feature patterns of multiple pest and disease types during training, they can adapt to different plant varieties, disease types, and environmental conditions, improving the accuracy and robustness of pest and disease index prediction.

[0080] In this scenario, the input to the pre-trained convolutional neural network includes two main features: vein prominence reference values ​​and color shift progressive reference values. These indices reflect the contrast between the veins and surrounding leaf tissues, and the gradual change trend of leaf color, respectively, allowing the network to capture early signs of potential pests and diseases. The convolutional neural network automatically learns the complex relationships between these features and pest and disease risk through multi-dimensional analysis. For example, when the color shift progressive reference value is high and a specific pattern appears in the vein prominence reference value, the convolutional neural network may identify this as a high-risk sign of a fungal disease (such as rust). Furthermore, for different combinations of indices, the convolutional neural network can predict the possible types and severity of pests and diseases based on pre-trained weights, and generate pest and disease indices to quantify the level of pest and disease risk within a sub-region.

[0081] Specifically, the convolutional neural network uses weights learned during training to extract key patterns from leaf vein highlighting reference values ​​and color shift gradient reference values ​​through convolutional layers. Pooling layers further reduce dimensionality to remove irrelevant information while retaining salient features. In fully connected layers, the model matches the extracted features with pest and disease risk patterns, outputting a final pest and disease index. This pre-training-based approach maximizes the use of existing pest and disease images and feature data, thereby achieving efficient and accurate pest and disease prediction and providing strong support for the intelligent management of hydroponics environments.

[0082] The convolutional neural network described above is not specifically limited here, but it can highlight leaf veins as a reference value. Color gradient reference value A comprehensive analysis is conducted to generate a pest and disease index. Any convolutional neural network can be used. To achieve the technical solution of this invention, this invention provides a specific implementation method; Pest and disease index. The generated calculation formula is: In the formula, , Reference values ​​for prominent leaf veins Color gradient reference value The preset proportional coefficient, and , All are greater than 0.

[0083] As can be seen from the pest and disease index, the larger the performance value of the leaf vein prominence reference value generated after analyzing the contrast between the leaf veins and the surrounding leaf tissue in each sub-region, and the larger the performance value of the color deviation progressive reference value generated after analyzing the progressive color deviation change of the leaf color in each sub-region, the larger the performance value of the pest and disease index generated when predicting the pests and diseases in the sub-region through the pre-trained convolutional neural network, the greater the potential risk of pest and disease invasion in that sub-region, and vice versa.

[0084] Based on the prediction results of the convolutional neural network, the health status of each sub-region is classified into "potential pest and disease areas" and "normal growth areas".

[0085] The pest and disease index generated when predicting pests and diseases in a sub-region using a pre-trained convolutional neural network is compared and analyzed with a pre-set reference threshold for the pest and disease index. The health status of each sub-region is then classified, as follows:

[0086] If the pest and disease index generated in a sub-region is greater than the pest and disease index reference threshold, then the health status of the sub-region is classified as a potential pest and disease area.

[0087] Potential pest and disease infestation areas refer to sub-regions where the pest and disease index generated by convolutional neural network analysis exceeds a preset reference threshold. This classification indicates that the health of plants within this sub-region may already be threatened by early pests and diseases, such as the appearance of small lesions on leaves, gradual changes in abnormal color, prominent leaf veins, or other possible signs of pests and diseases. Although the impact of pests and diseases at this stage may not yet manifest as widespread crop damage, it already poses a potential risk of spread and requires close monitoring and timely intervention.

[0088] If the pest and disease index generated in a sub-region is less than or equal to the pest and disease index reference threshold, then the health status of that sub-region is classified as a normal growth region.

[0089] The normal growth area refers to the sub-region where the pest and disease index generated by convolutional neural network analysis is less than or equal to the reference threshold for the pest and disease index. This classification indicates that the plants in this area are in good health, with no obvious early signs or potential threats of pests and diseases, and the color and texture of the leaves are consistent with the normal state, without any abnormal progressive color deviation or enhanced contrast of leaf veins.

[0090] This step automatically distinguishes between healthy and potentially pest-infested areas based on the prediction results of convolutional neural networks, thereby reducing indiscriminate monitoring of all areas. Through intelligent classification, the system can concentrate resources on areas with potential pest risks for priority monitoring and intervention, while normal areas are subject to routine monitoring, avoiding resource waste caused by over-monitoring and improving monitoring efficiency.

[0091] For areas with normal growth, continue to monitor for pests and diseases while maintaining the initial state.

[0092] For sub-regions classified as normal growing areas, since the plants are in good health and no early signs or potential risks of pests or diseases have been detected, there is no need to adjust the monitoring methods or frequencies; the initial monitoring settings can be continued. By maintaining the initial monitoring status, basic monitoring of normal areas can be ensured while avoiding unnecessary waste of resources. This allows more system computing resources and management focus to be concentrated on high-risk areas with potential pests and diseases, thereby optimizing monitoring efficiency and enabling precise management of the cultivation environment.

[0093] For sub-areas identified as potential pest and disease areas, the camera's control module automatically adjusts the camera's focus to concentrate on that sub-area. Once focused, the camera can more accurately capture the details of the plants within that sub-area.

[0094] For sub-regions identified as potential pest and disease infestations, automatic focusing is achieved through the camera's control module. Specifically, a pest and disease index generated by a convolutional neural network first determines the location coordinates of the target sub-region, which the system then uses as the camera's target point. Next, utilizing the camera's autofocus function, combined with depth calculation algorithms (such as measuring contrast peaks through image sharpness evaluation algorithms), the camera's focal plane is dynamically adjusted to precisely focus on the target sub-region. In this way, the camera can capture high-resolution images of the target area, further refining the observation and analysis of early signs of pests and diseases, and providing more detailed visual data support for precise intervention.

[0095] After the camera focuses on the potential pest and disease area, the current focal length of the camera is shortened based on the prediction results of the convolutional neural network, so as to capture the details of the target area more clearly and identify pests and diseases.

[0096] After the camera focuses on the potential pest and disease area, the current focal length of the camera is shortened based on the prediction results of the convolutional neural network. The specific steps are as follows:

[0097] The pest and disease index generated by a convolutional neural network in this sub-region is collected. This index is then compared with a preset reference threshold. Based on the relationship between the two, a focal length difference factor is calculated. The expression for this calculation is as follows: ,in, This is the focal length difference factor, used in subsequent calculations of the adjusted focal length. This indicates the pest and disease index generated by the convolutional neural network for this sub-region, reflecting the strength of the pest and disease risk within that sub-region. This serves as a reference threshold for the pest and disease index, representing the boundary between normal growth areas and potentially pest and disease-prone areas. An adjustment factor is used to quantify the impact of the degree to which the pest and disease index exceeds the reference threshold on focal length shortening.

[0098] The degree to which the pest and disease index exceeds a reference threshold is calculated to determine the magnitude of the focal length adjustment. When the pest and disease index exceeds the reference threshold, it is determined that the current focal length needs to be shortened to capture finer image features.

[0099] Based on the current focal length and the calculated focal length difference factor, calculate the adjusted target focal length, ensuring it remains within the camera's focal length adjustment range. The expression for the target focal length adjustment is: ,in, The adjusted target focal length is used to achieve sharper image capture. The current focal length of the camera. and This refers to the physical adjustment range of the camera's focal length, ensuring that the calculated target focal length does not exceed the camera's hardware capabilities. and These are the maximum and minimum focal lengths of the camera, respectively.

[0100] This step ensures that the target focal length remains within the range allowed by the camera hardware when adjusting the focal length based on the pest and disease index. By dynamically calculating the target focal length, the system can achieve more precise focusing on the target area, improving image resolution and detail reproduction.

[0101] The camera's focal length is automatically adjusted based on the calculated target focal length. Simultaneously, the focus result is verified in real-time using an image sharpness index. If the focus effect does not meet the expected set value, it is fine-tuned according to the adjustment range. The sharpness index formula is as follows: ,in, Image sharpness index, indicating the clarity of the target area in the image. Current focal length F (This focal length is the adjusted target focal length) Next pixel Image intensity value at that location, Image gradient, used to quantize changes in pixel intensity. N The total number of pixels within the target area;

[0102] Adjust the focus to the target focus. The image details captured by the camera will be clearer. Meanwhile, based on image sharpness metrics... The focusing effect is verified in real time to ensure that image quality reaches the optimal level. If the verification results do not meet expectations, the system can further optimize the focus adjustment process, thereby achieving efficient and accurate monitoring of potential pest and disease areas.

[0103] Shortening the focal length provides higher image resolution, making subtle signs of pests and diseases (such as tiny spots or slight leaf discoloration) clearer, aiding in accurate system identification. Shortening the focal length is an effective way to improve the accuracy of pest and disease identification. With a shorter focal length, the camera can capture details of the target area more clearly, especially in the early stages when pests and diseases are not yet obvious. This operation allows the system to detect signs of pests and diseases from minute changes. Through this refined monitoring, timely intervention can be carried out in the early stages of pest and disease outbreaks, effectively preventing the spread of pests and diseases and ensuring the health of crops throughout the entire cultivation area.

[0104] The aforementioned intelligent hydroponic cultivation environment monitoring method based on big data effectively solves the problem that existing single-view cameras cannot accurately identify early signs of pests and diseases. First, by dividing the cultivation area into multiple sub-regions and collecting image data from each sub-region, the representativeness of plant characteristics within each sub-region is ensured, thereby improving the accuracy of pest and disease monitoring. Second, using convolutional neural networks to extract features and predict pests and diseases from the collected image data not only accurately assesses the health status of each sub-region but also automatically adjusts the camera's focus when potential pest and disease areas are detected, concentrating on details and shortening the focal length, allowing early signs of pests and diseases to be clearly captured. This intelligent and dynamically adjusted monitoring method significantly improves the accuracy of pest and disease identification, enabling timely detection of potential pest and disease risks and ensuring early intervention, thereby effectively preventing the spread of pests and diseases and protecting the health of crops throughout the entire cultivation environment.

[0105] This invention provides, for example Figure 2 The intelligent soilless cultivation environment monitoring system based on big data shown includes a sub-region division module, an image data acquisition module, a feature extraction and pest prediction module, a health status classification module, a normal area monitoring module, a potential pest focusing module, and a focal length optimization and fine recognition module.

[0106] The sub-region division module divides the hydroponic environment area into several sub-regions of equal size. The area of ​​each sub-region is set according to the specific cultivation environment and monitoring requirements to ensure that the plants in each sub-region are representative.

[0107] The image data acquisition module automatically acquires image data of plant leaves in each sub-region using a camera, and transmits it to the data processing system in real time for subsequent pest and disease detection and analysis.

[0108] The feature extraction and pest prediction module summarizes the image data collected from each sub-region, extracts features reflecting potential pests and diseases from the summary, analyzes the extracted features, and inputs the analyzed features into a pre-trained convolutional neural network to predict the pests and diseases in the sub-region.

[0109] The health status classification module, based on the prediction results of the convolutional neural network, classifies the health status of each sub-region into "potential pest and disease areas" and "normal growth areas".

[0110] The normal area monitoring module continues to monitor pests and diseases in normal growth areas, maintaining the initial state.

[0111] The potential pest and disease focusing module automatically adjusts the camera's focus through the camera's control module for sub-areas identified as potential pest and disease areas, so that the camera focuses on the sub-area. After the camera focuses, it will capture the details of the plants in the sub-area more accurately.

[0112] The focal length optimization and fine recognition module shortens the current focal length of the camera after the camera focuses on the potential pest and disease area, based on the prediction results of the convolutional neural network, so as to capture the details of the target area more clearly and identify pests and diseases.

[0113] The intelligent soilless cultivation environment monitoring method based on big data provided in this embodiment of the invention is implemented through the aforementioned intelligent soilless cultivation environment monitoring system based on big data. For details of the specific methods and processes of the intelligent soilless cultivation environment monitoring system based on big data, please refer to the embodiments of the intelligent soilless cultivation environment monitoring method based on big data, which will not be repeated here.

[0114] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0115] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0116] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for monitoring the intelligent hydroponics environment based on big data, characterized in that, Includes the following steps: The hydroponic environment area is divided into several sub-areas of equal size. The area of ​​each sub-area is set according to the specific cultivation environment and monitoring requirements to ensure that the plants in each sub-area are representative. Within each sub-region, images of plant leaves within that region are automatically collected by cameras and transmitted in real time to the data processing system for subsequent pest and disease detection and analysis. The image data collected from each sub-region is summarized, and features reflecting potential pests and diseases are extracted from the summarized data. After analyzing the extracted features, the analyzed features are input into a pre-trained convolutional neural network, which then predicts the pests and diseases in the sub-region. Based on the prediction results of the convolutional neural network, the health status of each sub-region is classified into "potential pest and disease areas" and "normal growth areas". For areas with normal growth, continue to monitor for pests and diseases in their initial state; for sub-areas identified as potential pest and disease areas, the camera's focus is automatically adjusted through the camera's control module so that the camera focuses on the sub-area. After the camera focuses, it will capture the details of the plants in the sub-area more accurately. After the camera focuses on the potential pest and disease area, the current focal length of the camera is shortened based on the prediction results of the convolutional neural network, so as to capture the details of the target area more clearly and identify pests and diseases. Features reflecting potential pests and diseases are extracted from the aggregated data. These features include the degree of contrast between leaf veins and surrounding leaf tissues and the gradual color shift of leaf color. After analyzing the degree of contrast between leaf veins and surrounding leaf tissues and the gradual color shift of leaf color in each sub-region, leaf vein prominence reference values ​​and color shift gradual reference values ​​are generated. The leaf vein prominence reference values ​​quantify the intensity of the contrast between leaf veins and surrounding leaf tissues, and the color shift gradual reference values ​​quantify the degree of gradual shift of leaf color in time or space. The vein prominence reference value and color shift reference value generated after analyzing the contrast between the leaf vein and the surrounding leaf tissue and the gradual color shift change of the leaf color are input into a pre-trained convolutional neural network. Based on the convolutional neural network, a pest and disease index is generated, and the pest and disease index is used to predict the pest and disease in the sub-region. The specific steps for generating vein prominence reference values ​​after analyzing the contrast between leaf veins and surrounding leaf tissue within each sub-region are as follows: The plant images of each sub-region are processed, and the spectral reflectance features of the leaves are extracted using multispectral imaging technology. A spectral contrast matrix between the leaf vein region and the surrounding leaf tissue is constructed using spectral contrast values. Based on the spectral contrast matrix, a position weighting function and a dynamic threshold function are introduced to normalize the matrix and calculate the reference value for vein prominence. The specific steps for generating a progressive color shift reference value after analyzing the progressive color shift changes of the leaves in each sub-region are as follows: By performing multi-dimensional color analysis on the leaf images in each sub-region, a color shift feature matrix is ​​extracted; By using the color offset feature matrix and spatial offset weight, the color offset progressive reference value of the sub-region is calculated to quantify the degree of progressive offset of the leaf color in time or space. Disease and pest index Pest The generated calculation formula is: , In the formula, l p l q The reference value for vein prominence V is respectively. promoinence Color shift gradient reference value GCSI index The preset proportional coefficient, and l p l q All are greater than 0; The pest and disease index generated when predicting pests and diseases in a sub-region using a pre-trained convolutional neural network is compared and analyzed with a pre-set reference threshold for the pest and disease index. The health status of each sub-region is then classified, as follows: If the pest and disease index generated in a sub-region is greater than the pest and disease index reference threshold, then the health status of the sub-region is classified as a potential pest and disease area. If the pest and disease index generated in a sub-region is less than or equal to the pest and disease index reference threshold, then the health status of the sub-region is classified as a normal growth region. After the camera focuses on the potential pest and disease area, the current focal length of the camera is shortened based on the prediction results of the convolutional neural network. The specific steps are as follows: The pest and disease index generated by a convolutional neural network in this sub-region is collected. This index is then compared with a preset reference threshold. Based on the relationship between the two, a focal length difference factor is calculated. The expression for this calculation is as follows: Δ focus =α v ·(Ther Balan -Ther ref ), Where, Δ focus This is the focal length difference factor, used in subsequent calculations of the adjusted focal length. Balan This indicates that the sub-region is generated by a convolutional neural network predicting pest and disease indices. ref As a reference threshold for the pest and disease index, α v An adjustment factor used to quantify the impact of the degree to which the pest and disease index exceeds the reference threshold on focal length shortening; Based on the current focal length and the calculated focal length difference factor, calculate the adjusted target focal length, ensuring it remains within the camera's focal length adjustment range. The expression for the target focal length adjustment is: F target =max(F min ,me(F current -Δ focus ,F max )), Among them, F target The adjusted target focal length is used to achieve sharper image capture. current F is the current focal length of the camera. min and F max This refers to the physical adjustment range of the camera's focal length, ensuring that the calculated target focal length does not exceed the camera's hardware capabilities. Here, F... min and F max These are the maximum and minimum focal lengths of the camera, respectively. The camera's focal length is automatically adjusted based on the calculated target focal length. Simultaneously, the focus result is verified in real-time using an image sharpness index. If the focus effect does not meet the expected set value, it is fine-tuned according to the adjustment range. The sharpness index formula is as follows: , Where Q(F) is the image sharpness index, representing the sharpness of the target area image, and I(F, i, j) is the image intensity value at pixel (i, j) at the current focal length F. The gradient is the image gradient used to quantify changes in pixel intensity, and N is the total number of pixels in the target region.

2. The intelligent hydroponics environment monitoring method based on big data according to claim 1, characterized in that, The hydroponics environment is divided into several sub-regions of equal size. The specific steps are as follows: It is necessary to clarify the total area of ​​the monitoring area of ​​the entire hydroponics environment and the area ratio required for dividing it into sub-areas. Based on the shape of the environment, the layout of the cultivation system, the types of plants, and the utilization needs of the cultivation area, the size of each sub-area should be determined. After determining the area of ​​the sub-region, select a region division method to divide the entire cultivation environment into several small squares according to an equal grid, with each small square being a sub-region.

3. The intelligent hydroponics environment monitoring method based on big data according to claim 1, characterized in that, The specific steps for generating vein prominence reference values ​​after analyzing the contrast between leaf veins and surrounding leaf tissue within each sub-region are as follows: The plant images of each sub-region are processed, and the spectral reflectance features of the leaves are extracted using multispectral imaging technology. A spectral contrast matrix between the vein region and the surrounding leaf tissue is constructed using spectral contrast values. The expression for constructing the spectral contrast matrix is ​​as follows: , Where: M contrast (i, j) represents the spectral contrast value between the leaf vein and the surrounding leaf tissue at pixel (i, j), R leaf_vein (i, j, λ) NIR R represents the near-infrared reflectance value of pixel (i, j) in the leaf vein region. surrounding (i, j, λ) NIR The reflectance of pixel (i, j) in the green band of the tissue surrounding the leaf vein, λ NIR For the near-infrared band, λ green It is in the green light band; Based on the spectral contrast matrix, a positional weighting function and a dynamic threshold function are introduced to normalize the matrix, and the reference value for vein prominence is calculated. The expression for the calculation is as follows: , Where: V prominence The reference value for highlighting leaf veins is used to quantify the intensity of leaf vein contrast in the entire sub-region. H and W are the height and width of the image, respectively. W(i,j) is the positional weighting function, which assigns different weights to pixels (i,j) according to the importance of the leaf region. T(i,j) is the dynamic thresholding function, used to remove background noise from normal leaf vein contrast, defined as: , Where dist(i,j) is the distance from pixel (i,j) to the center of the blade, and α and β are parameters used to control threshold decay.

4. The intelligent hydroponic cultivation environment monitoring method based on big data according to claim 3, characterized in that, The specific steps for generating a progressive color shift reference value after analyzing the progressive color shift changes of the leaves in each sub-region are as follows: By performing multi-dimensional color analysis on the leaf images within each sub-region, a color shift feature matrix is ​​extracted. This matrix is ​​calculated based on the color changes of the leaf at time t and spatial position (x, y), and the expression for the calculation is as follows: , Where C(t, x, y) is the color shift feature matrix, representing the degree of color shift at spatial location (x, y) at time t, and R λ R represents the color reflectance at a spatial location (x, y) at wavelength λ. λ (t-1, x, y) represents the color reflectance value at the previous time point t-1, W λ This refers to the band weighting coefficient; Using the color shift feature matrix and spatial shift weights, the color shift progressive reference value of the sub-region is calculated to quantify the degree of progressive shift of the leaf color in time or space. The calculation expression is as follows: , Among them, GCSI index The color shift reference value quantifies the overall gradual shift of leaf color within a sub-region. S(x, y) represents the spatial shift weight, set based on the spatial concentration of leaf density and color shift within the sub-region. ∈ is a smoothing factor to prevent the color shift from being too saturated to be too saturated. x，y When S(x, y), the denominator becomes zero.

5. The intelligent hydroponic cultivation environment monitoring method based on big data according to claim 1, characterized in that, For sub-areas identified as potential pest and disease infestations, automatic focusing is achieved through the camera's control module, specifically: First, the pest and disease index generated by the convolutional neural network determines the location coordinates of the target sub-region, and these coordinates are used as the target point of the camera. Next, by utilizing the camera's autofocus function and combining it with depth calculation algorithms, the camera's focal plane is dynamically adjusted to ensure that the camera accurately focuses on the target sub-region.

6. A big data-based intelligent soilless cultivation environment monitoring system, used to implement the big data-based intelligent soilless cultivation environment monitoring method according to any one of claims 1-5, characterized in that, It includes a sub-region division module, an image data acquisition module, a feature extraction and pest prediction module, a health status classification module, a normal area monitoring module, a potential pest focusing module, and a focal length optimization and fine recognition module. The sub-region division module divides the hydroponic environment area into several sub-regions of equal size. The area of ​​each sub-region is set according to the specific cultivation environment and monitoring requirements to ensure that the plants in each sub-region are representative. The image data acquisition module automatically acquires image data of plant leaves in each sub-region using a camera, and transmits it to the data processing system in real time for subsequent pest and disease detection and analysis. The feature extraction and pest prediction module summarizes the image data collected from each sub-region, extracts features reflecting potential pests and diseases from the summary, analyzes the extracted features, and inputs the analyzed features into a pre-trained convolutional neural network to predict the pests and diseases in the sub-region. The health status classification module, based on the prediction results of the convolutional neural network, classifies the health status of each sub-region into "potential pest and disease areas" and "normal growth areas". The normal area monitoring module continues to monitor pests and diseases in normal growth areas, maintaining the initial state. The potential pest and disease focusing module automatically adjusts the camera's focus through the camera's control module for sub-areas identified as potential pest and disease areas, so that the camera focuses on the sub-area. After the camera focuses, it will capture the details of the plants in the sub-area more accurately. The focal length optimization and fine recognition module shortens the current focal length of the camera after the camera focuses on the potential pest and disease area, based on the prediction results of the convolutional neural network, so as to capture the details of the target area more clearly and identify pests and diseases.

Images