Cotton field growth state evaluation and prediction system based on unmanned aerial vehicle RGB image

Abstract

The invention provides a cotton field growth state evaluation and prediction system based on an unmanned aerial vehicle RGB image, and the system comprises a data collection module which collects initial image data; the data processing module is used for denoising, distortion correction, orthographic splicing and illumination correction to obtain standard image data; the image mining module is used for carrying out image segmentation and target detection and then extracting cotton profile features and growth indexes; the comprehensive evaluation module is used for constructing a growth state evaluation model to obtain a comprehensive growth condition score and an evaluation index; the growth prediction module introduces yield related factors to carry out yield prediction and growth abnormity early warning; and the real-time monitoring and decision support module visually displays the evaluation prediction result and automatically generates decision support. The cotton field growth vigor can be comprehensively analyzed, the comprehensiveness, the accuracy and the high efficiency of evaluation and prediction are improved, real-time monitoring and intelligent decision making can be carried out on the whole cotton growth process, and the efficiency and the precision of agricultural management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of plant growth assessment, and particularly to a cotton field growth status assessment and prediction system based on UAV RGB images. Background Art

[0002] Cotton plants are important economic crops globally and are the main raw materials for the textile industry. They are widely used in the production of products such as clothing and home textiles, supporting the global textile industry chain. By studying the comprehensive growth status of cotton, planting management measures can be optimized, the yield per unit area and fiber quality of cotton can be improved, and understanding the spatial and temporal variation characteristics of cotton growth helps with precise fertilization, irrigation, and pest control, reducing agricultural input costs.

[0003] With the development of science and technology, existing methods and technologies for evaluating the comprehensive growth profile of cotton plants include, but are not limited to, hyperspectral imaging, Internet of Things sensors, etc. However, most of these technologies are costly and have limitations in data processing, making it difficult to be widely applied. Evaluating cotton growth using UAV RGB images can reduce the usage cost. At the same time, UAVs are highly flexible and can quickly cover large planting areas. However, existing related technologies for evaluating cotton growth using UAV RGB images still have the following problems: 1) High complexity in data processing, large amounts of RGB image data, low processing efficiency, and image processing algorithms being sensitive to lighting conditions, which may lead to unstable results; 2) Lack of comprehensive analysis of multiple factors: Existing research mostly focuses on a single indicator (such as canopy coverage rate) and lacks comprehensive analysis of multiple factors (such as soil and meteorology); 3) Insufficient real-time performance: There is a time delay in data collection and analysis, making it difficult to achieve real-time monitoring and dynamic assessment. Summary of the Invention

[0004] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a cotton field growth status assessment and prediction system based on UAV RGB images, which can comprehensively analyze cotton growth by combining multiple factors, improve the comprehensiveness, accuracy, and efficiency of assessment and prediction, and can also conduct real-time monitoring and intelligent decision-making on the entire process of cotton growth, enhancing the efficiency and precision of agricultural management and promoting the new stage of digitalization and intelligentization of cotton precision agriculture.

[0005] To achieve the above purpose, the present invention provides the following solution: A cotton field growth status assessment and prediction system based on UAV RGB images, including:

[0006] A data acquisition module, which is used to select a UAV, mount an RGB camera on the selected UAV, preset the UAV flight path and image acquisition parameters according to the terrain of the plot, and then conduct data acquisition to obtain initial image data;

[0007] A data processing module for denoising, distortion correction, orthorectification mosaicking, and illumination correction of the initial image data to obtain standard image data;

[0008] An image mining module for image segmentation and target detection of the standard image data to obtain cotton target information, and then extracting the canopy coverage, plant height, color features, vegetation indices, and texture and spatial structure features of the cotton target information to obtain cotton profile features and growth indicators;

[0009] A comprehensive evaluation module for introducing environmental data, training a machine learning model using the environmental data and the growth indicators to obtain a growth status evaluation model, and then using the growth status evaluation model to obtain a comprehensive growth status score and evaluation indicators;

[0010] A growth prediction module for based on the growth status evaluation model, introducing yield-related factors for yield prediction, and presetting a growth anomaly threshold for growth anomaly warning;

[0011] A real-time monitoring and decision support module for using a cloud service platform for distributed storage and management of data, using WebGIS technology to visually display the comprehensive evaluation results and growth prediction results of each cotton plot, and automatically generating decision support based on the comprehensive evaluation results and growth prediction results, and performing automated management according to the decision support;

[0012] Wherein, the data acquisition module, the data processing module, the image mining module, the comprehensive evaluation module, the growth prediction module, and the real-time monitoring and decision support module are interconnected with each other.

[0013] Optionally, the data acquisition module includes:

[0014] A hardware configuration unit for selecting a drone and mounting an RGB camera and a three-axis gimbal adapted to the RGB camera on the selected drone to achieve the anti-shake performance of the camera;

[0015] A flight path planning unit for automatically generating a flight path using drone flight path planning software and setting the forward overlap and side overlap of the images;

[0016] A parameter setting unit for setting the flight height and resolution of the drone according to the cotton growth stage, and then performing data acquisition to obtain initial image data.

[0017] Optionally, the data processing module includes:

[0018] A denoising and distortion correction unit for denoising the initial image data using an automatic batch processing tool and a denoising algorithm, and correcting the image edge distortion using camera calibration parameters;

[0019] An orthorectification stitching unit, which is used to use aerial survey processing software to perform feature point matching and image stitching on the overlapping images in the initial image data, and output an orthophoto with geographic coordinate system information;

[0020] A lighting correction unit, which is used to deploy a gray scale board inside the flight area of the UAV, record the standard reflection value, perform color correction on the initial image data based on the standard reflection value, and integrate a radiation correction algorithm in the aerial survey processing software to standardize the image brightness and image color.

[0021] Optionally, the image mining module includes:

[0022] A data annotation unit, which is used to perform pixel-level annotation based on the standard image data by using an annotation tool, generate segmentation labels for cotton plants and soil background, and perform bounding box annotation on single cotton plants and pest and disease areas;

[0023] An image segmentation unit, which is used to select a convolutional neural network, introduce an attention mechanism, a multi-scale feature extraction module and depthwise separable convolution into the selected convolutional neural network to obtain a first initial model, and then use a deep learning framework to train the first initial model to obtain a segmentation model, and use the segmentation model to segment the standard image data to output a binary image; wherein, the binary image includes the overall distribution information of the cotton canopy;

[0024] A target detection unit, which is used to combine YOLOv5 and multi-task learning to obtain a target detection model, and use the target detection model to detect the binary image to output position information; wherein, the position information includes the positions of single cotton plants or pest and disease areas;

[0025] A result fusion unit, which is used to combine the binary image and the position information to obtain cotton target information;

[0026] A profile feature and growth index extraction unit, which is used to extract the canopy coverage, plant height, color features, vegetation index, and texture and spatial structure features of the cotton target information to obtain the profile features and growth indexes of the cotton target information.

[0027] Optionally, the profile feature and growth index extraction unit includes:

[0028] A canopy coverage calculation sub-unit, which is used to count the proportion of cotton canopy pixels in the binary image to obtain the canopy coverage, and evaluate whether the growth process is normal according to the canopy coverage in combination with the cotton growth period;

[0029] A plant height estimation subunit, which is used to estimate the plant height by combining the orthoimage and the digital surface model, and output a spatial distribution map of the plant height to identify areas with abnormal growth;

[0030] A texture and spatial structure feature extraction subunit, which is used to extract texture features based on the standard image data by using a gray-level co-occurrence matrix, and calculate the population density and plant spacing of cotton plants to obtain spatial structure features;

[0031] A color feature and vegetation index calculation subunit, which is used to extract the color features of the RGB bands in the standard image data, design an RGB vegetation index by combining the texture features and the color features, and then evaluate the health status of cotton and identify pest and disease areas according to the spatial distribution of the RGB vegetation index.

[0032] Optionally, the comprehensive evaluation module includes:

[0033] A data fusion unit, which is used to obtain soil data, meteorological data and geographical information data of a cotton field to obtain environmental data, perform spatial alignment and temporal alignment on the environmental data and the standard image data, and then fuse the environmental data with the extracted growth indicators to obtain a comprehensive data set;

[0034] A model construction unit, which is used to construct an initial model by using an ensemble learning method, set the input variables and output variables of the initial model, and use the comprehensive data set to train the model to obtain a growth status evaluation model; the input variables include canopy coverage, plant height, vegetation index, texture features and environmental data, and the output variables include a comprehensive growth status score, a health index and a stress index;

[0035] An index quantification unit, which is used to generate a spatial distribution map according to the comprehensive growth status score, and combine the health index and the stress index to obtain evaluation indicators for comprehensive evaluation.

[0036] Optionally, the expression of the comprehensive growth status score is:

[0037] S = ω1·C + ω2·H + ω3·VI + ω4·T

[0038] Where S is the comprehensive growth status score, C is the canopy coverage, H is the plant height, VI is the vegetation index, T is the texture feature, and ω1, ω2, ω3, ω4 are the weights of each index;

[0039] The expression of the health index is:

[0040] HI = α·VI + β·T

[0041] Where HI is the health index, and α and β are both weight parameters;

[0042] The expression of the adversity index is as follows:

[0043] AI = γ1·(1 - E soil ) + γ2·|E temp - T opt |

[0044] where AI is the adversity index, E soil is the soil humidity, E temp is the temperature, T opt is the optimal growth temperature of cotton, and γ1 and γ2 are both weight parameters.

[0045] Optionally, the growth prediction module includes:

[0046] A yield prediction unit, which is used to fuse the canopy coverage, the plant height, the vegetation index, the texture feature, the environmental data and the yield-related factors to obtain a yield prediction data set, use the yield prediction data set for model training to obtain a yield prediction model, and use the yield prediction model to output the single-plant yield prediction value and the unit-area yield prediction value; wherein, the yield-related factors include plant density and boll opening amount, and an environmental correction factor for adjusting the output value is introduced into the yield prediction model;

[0047] A growth anomaly unit, which is used to set the comprehensive growth condition score threshold, the health index threshold and the adversity index threshold according to the output value of the growth state evaluation model, and then judge whether the cotton plot is an abnormal area according to the set thresholds. If so, an early warning information is automatically generated.

[0048] Optionally, the expression of the single-plant yield prediction value is:

[0049] Y single = B·W boll

[0050] where Y single is the single-plant yield prediction value, B is the boll opening amount, and W boll is the weight of a single boll;

[0051] The expression of the unit-area yield prediction value is:

[0052] Y yield = D·Y single ·(1 - k1·|E soil - E opt |- k2·|E temp - T opt |)

[0053] where Y yield is the unit-area yield prediction value, D is the plant density, Eopt is the optimal soil moisture, T opt is the optimal growth temperature, and k1 and k2 are environmental correction coefficients.

[0054] Optionally, the real-time monitoring and decision support module includes:

[0055] A data storage unit for storing structured data based on a cloud service platform using a relational database, managing unstructured data using object storage, and performing real-time data access through an API interface;

[0056] A monitoring and visualization unit for developing a responsive interface that supports multi-terminal access using a front-end framework, and using WebGIS technology to display the processing results of the image mining module, the comprehensive evaluation module, and the growth prediction module in real time on the interface in the form of map overlay, plot view, and heat map;

[0057] A decision support unit for generating fertilization and irrigation suggestions, pest control supervision suggestions, climate change trend analysis, and pest pattern analysis based on the processing results of the image mining module, the comprehensive evaluation module, and the growth prediction module.

[0058] The present invention provides a cotton field growth status evaluation and prediction system based on UAV RGB images, and discloses the following technical effects:

[0059] 1. Efficient data processing: Through automated data processing and sufficient image mining, the present invention enables multi-features such as canopy coverage, plant height, color features, vegetation indices, and texture and spatial structure features to truly reflect near-infrared or hyperspectral results, comprehensively and accurately reflecting the growth status of cotton, providing a reliable data basis for subsequent cotton growth evaluation and prediction.

[0060] 2. High comprehensive evaluation accuracy: By fusing UAV image data with external environmental data such as soil and meteorology in the spatial and temporal dimensions, the present invention obtains multi-dimensional comprehensive evaluation data to construct a growth status evaluation model, obtaining a quantified comprehensive growth status score and health index, which not only comprehensively reveals the comprehensive growth status of cotton but also improves the scientificity and reliability of the evaluation.

[0061] 3. Real-time monitoring and intelligent decision-making: By building a cloud service platform, the present invention can achieve efficient data storage and management, ensure real-time data upload, and through a visualization interface, can display the comprehensive evaluation results and growth prediction results in real time, realize real-time monitoring of data and multi-terminal access, and can automatically generate fertilization and irrigation suggestions, pest control supervision suggestions, analyze climate change trends and pest patterns, etc., improving the intelligent level of agricultural management.

[0062] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings required for use in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0064] Figure 1 System architecture diagram provided for the embodiment of the present invention;

[0065] Figure 2 Flow diagram of image mining provided for the embodiment of the present invention;

[0066] Figure 3 Flow diagram of evaluation prediction provided for the embodiment of the present invention. Detailed Embodiments

[0067] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0068] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] As Figure 1 shown, the present invention provides a cotton field growth status evaluation and prediction system based on UAV RGB images, including a data acquisition module, a data processing module, an image mining module, a comprehensive evaluation module, a growth prediction module, and a real-time monitoring and decision support module that are interconnected.

[0070] 1. Data Acquisition Module

[0071] Used to select a UAV, mount an RGB camera on the selected UAV, preset the UAV flight path and image acquisition parameters according to the terrain of the plot, and then perform data acquisition to obtain initial image data; the data acquisition module includes:

[0072] 1.1 Hardware Configuration Unit

[0073] For selecting a drone, mounting an RGB camera and a three-axis gimbal adapted to the RGB camera on the selected drone to achieve the anti-shake performance of the camera and reduce image blurring caused by vibration.

[0074] 1.2 Route Planning Unit

[0075] For using drone route planning software (such as DJIPilot, Pix4Dcapture, UGCS) to carry out mission design. Input cotton field boundary and terrain data, and automatically generate a flight path covering the entire area.

[0076] Set the forward overlap degree and side overlap degree of the images. The forward overlap degree is set to 70%-80% to ensure there is enough overlap area between images for stitching. The side overlap degree is set to 60%-70% to avoid blank areas when stitching images.

[0077] 1.3 Parameter Setting Unit

[0078] For setting the flight altitude and resolution of the drone according to the cotton growth stage, and then carrying out data acquisition to obtain initial image data.

[0079] 2. Data Processing Module

[0080] For denoising, distortion correction, ortho-mosaic stitching and illumination correction of the initial image data to obtain standard image data; the data processing module includes:

[0081] 2.1 Denoising and Distortion Correction Unit

[0082] For using an automatic batch processing tool (such as OpenCV or ImageJ) and a denoising algorithm (such as mean filtering, Gaussian filtering) to perform denoising processing on the initial image data and reduce the interference of random noise on image features.

[0083] Use camera calibration parameters (such as focal length, distortion coefficient) to correct the distortion of the image edges and correct the bending of the image edges caused by the optical characteristics of the lens. An open-source tool (such as the `cv2.undistort()` function in OpenCV) can be used to achieve automatic correction.

[0084] 2.2 Ortho-mosaic Stitching Unit

[0085] For using photogrammetry processing software (such as Pix4D, Agisoft Metashape, etc.) to perform feature point matching and image stitching on the overlapping images in the initial image data, and output an ortho-image with geographic coordinate system information.

[0086] 2.3 Illumination Correction Unit

[0087] Used to deploy gray scale plates within the flight area of the drone, record standard reflection values, perform color correction on the initial image data based on the standard reflection values, reduce lighting differences, and integrate a radiometric correction algorithm (such as Empirical Line Calibration, ELC) in the aerial survey processing software to standardize image brightness and image color.

[0088] 3. Image Mining Module

[0089] Such as Figure 2 As shown, it is used to perform image segmentation and target detection on the standard image data to obtain cotton target information, and then extract the canopy coverage, plant height, color characteristics, vegetation index, and texture and spatial structure characteristics of the cotton target information to obtain cotton profile characteristics and growth indicators; the image mining module includes:

[0090] 3.1 Data Annotation Unit

[0091] Used to perform pixel-level annotation on the standard image data using annotation tools (such as Labelme, LabelImg), generate segmentation labels for cotton plants and soil backgrounds, and perform bounding box annotation on single cotton plants and pest and disease areas.

[0092] 3.2 Image Segmentation Unit

[0093] Used to select a convolutional neural network (such as MobileNet U-Net), introduce an attention mechanism (such as SE module or CBAM module), a multi-scale feature extraction module, and depthwise separable convolution (to reduce the number of parameters and computational complexity) in the selected convolutional neural network to obtain a first initial model, and then use a deep learning framework to train the first initial model to obtain a segmentation model, and use the segmentation model to segment the standard image data and output a binary image (the cotton plant area is 1 and the soil background is 0), providing a basis for subsequent feature extraction; among them, the binary image includes the overall distribution information of the cotton canopy.

[0094] 3.3 Target Detection Unit

[0095] Used to combine YOLOv5 and multi-task learning (simultaneously output the positions of cotton plants and pest and disease areas in target detection to improve detection efficiency) to obtain a target detection model, and use the target detection model to detect the binary image and output position information; among them, the position information includes the positions of single cotton plants or pest and disease areas.

[0096] 3.4 Result Fusion Unit

[0097] Used to combine the binary image and the position information to obtain cotton target information.

[0098] 3.5 Profile Feature and Growth Index Extraction Unit

[0099] It is used to extract the canopy coverage, plant height, color features, vegetation indices, and texture and spatial structure features of the cotton target information, and obtain the profile features and growth indices of the cotton target information.

[0100] The profile feature and growth index extraction unit includes:[[]]

[0101] 3.51 Canopy Coverage Calculation Sub-unit

[0102] It is used to count the proportion of cotton canopy pixels in the binary image to obtain the canopy coverage, and evaluate whether the growth process is normal according to the canopy coverage in combination with the cotton growth period (seedling stage, budding stage, flowering and boll stage, boll opening stage).

[0103] 3.52 Plant Height Estimation Sub-unit

[0104] It is used to estimate the plant height by combining the orthoimage and the digital surface model, and output the spatial distribution map of the plant height to identify the growth abnormal areas (such as too high or too low plants).

[0105] 3.53 Texture and Spatial Structure Feature Extraction Sub-unit

[0106] It is used to extract texture features based on the standard image data by using the gray-level co-occurrence matrix, and calculate the population density of cotton plants (count the number of cotton plants per unit area) and the plant spacing (analyze the distribution uniformity of cotton plants) to obtain the spatial structure features.

[0107] 3.54 Color Feature and Vegetation Index Calculation Sub-unit

[0108] It is used to extract the color features (such as average brightness, green component) of the RGB band in the standard image data, design the RGB vegetation index by combining the texture features and the color features, and then evaluate the health status of the cotton and identify possible pest and disease areas according to the spatial distribution of the RGB vegetation index.

[0109] 4. Comprehensive Evaluation Module

[0110] Such as Figure 3 shown, it is used to introduce environmental data, train a machine learning model by using the environmental data and the growth indices to obtain a growth status evaluation model, and then use the growth status evaluation model to obtain a comprehensive growth status score and evaluation indices; the comprehensive evaluation module includes:[[]]

[0111] 4.1 Data Fusion Unit

[0112] Used to obtain soil data, meteorological data, and geographic information data of cotton fields to obtain environmental data, spatially and temporally align the environmental data with the standard image data, and then fuse the environmental data with the extracted growth indicators to obtain a comprehensive dataset.

[0113] Spatial alignment: Align the environmental data with the geographic coordinate system (such as WGS84) of the UAV images. Use a GIS platform (such as ArcGIS, QGIS) for spatial overlay, map the soil and meteorological data to the cotton field area, and generate a multi-layer spatial information map.

[0114] Temporal alignment: Unify the timestamps, synchronize the environmental data with the UAV flight time, and ensure the timeliness of the data.

[0115] When fusing data, fuse at the plot or pixel level in the spatial dimension. Aggregate at time windows such as daily, weekly, and monthly in the temporal dimension.

[0116] 4.2 Model construction unit

[0117] Used to construct an initial model using an ensemble learning method (such as Random Forest, XGBoost), set the input variables and output variables of the initial model, and use the comprehensive dataset for model training to obtain a growth status evaluation model; the input variables include canopy coverage, plant height, vegetation index, texture features, and environmental data (soil moisture, nutrients, meteorological data), and the output variables include a comprehensive growth status score (0 - 100 points), a health index (reflecting the health degree of plants), and a stress index (reflecting the influence degree of stresses such as drought and pests).

[0118] The expression for the comprehensive growth status score is:

[0119] S = ω1·C + ω2·H + ω3·VI + ω4·T

[0120] Where, S is the comprehensive growth status score, C is the canopy coverage, H is the plant height, VI is the vegetation index, T is the texture feature, and ω1, ω2, ω3, ω4 are the weights of each index.

[0121] The expression for the health index is:

[0122] HI = α·VI + β·T

[0123] Where, HI is the health index, and α and β are both weight parameters.

[0124] The expression for the stress index is:

[0125] AI = γ1·(1 - E soil ) + γ2·|E temp - Topt |

[0126] Among them, AI is the adversity index, E soil is the soil humidity, E temp is the temperature, T opt is the optimal growth temperature of cotton, and γ1 and γ2 are both weight parameters.

[0127] 4.33 Index Quantification Unit

[0128] It is used to generate a spatial distribution map according to the comprehensive growth status score, and combine the health index and the adversity index to obtain evaluation indicators for comprehensive evaluation.

[0129] 5. Growth Prediction Module

[0130] As Figure 3 shown, it is used to introduce yield-related factors based on the growth status evaluation model to predict yield, and preset a growth anomaly threshold for growth anomaly warning; the growth prediction module includes:

[0131] 5.1 Yield Prediction Unit

[0132] As Figure 3 shown, it is used to fuse the canopy coverage, the plant height, the vegetation index, the texture feature, the environmental data and the yield-related factors to obtain a yield prediction data set, use the yield prediction data set for model training to obtain a yield prediction model, and use the yield prediction model to output the single-plant yield prediction value and the unit area yield prediction value; among them, the yield-related factors include plant density and boll opening amount, and an environmental correction factor for adjusting the output value is introduced into the yield prediction model.

[0133] The expression of the single-plant yield prediction value is:

[0134] Y single = B·W boll

[0135] Among them, Y single is the single-plant yield prediction value, B is the boll opening amount, and W boll is the weight of a single boll.

[0136] The expression of the unit area yield prediction value is:

[0137] Y yield = D·Y single ·(1 - k1·|E soil - E opt |- k2·|E temp - T opt |)

[0138] Among them, Yyield is the predicted value of yield per unit area, D is the plant density, and E opt is the optimal soil moisture, and T opt is the optimal growth temperature, and k1 and k2 are environmental correction factors.

[0139] 5.2 Growth Abnormal Unit

[0140] It is used to set the comprehensive growth condition score threshold, health index threshold, and adversity index threshold according to the output value of the growth state evaluation model, and then judge whether the cotton plot is an abnormal area according to the set thresholds. If so, warning information will be automatically generated. The warning information includes the location of the abnormal area (latitude and longitude), analysis of the abnormal cause (such as drought, pests and diseases), and prevention and control suggestions (such as irrigation, fertilization, spraying pesticides).

[0141] 6. Real-time Monitoring and Decision Support Module

[0142] It is used to use the cloud service platform for distributed storage and management of data, use WebGIS technology to visually display the comprehensive evaluation results and growth prediction results of each cotton plot, and automatically generate decision support according to the comprehensive evaluation results and growth prediction results, and perform automated management according to the decision support; the real-time monitoring and decision support module includes:

[0143] 6.1 Data Storage Unit

[0144] It is used to store structured data (such as health index, yield prediction results) based on a cloud service platform (such as AWS, Google Cloud, Alibaba Cloud) using a relational database, manage unstructured data (such as drone images, ortho-mosaics) using object storage, and perform real-time data access through an API interface.

[0145] 6.2 Monitoring and Visualization Unit

[0146] It is used to develop a responsive interface that supports multi-terminal access using a front-end framework, and use WebGIS technology to display the processing results of the image mining module, the comprehensive evaluation module, and the growth prediction module in real time on the interface in the form of map overlay, plot view, and heat map.

[0147] Map Overlay: Use an open-source GIS framework (such as Leaflet, OpenLayers) or a commercial platform (such as ArcGISOnline) to implement the map overlay function. Display multiple layers of information on the map (such as health index, canopy coverage, soil moisture, pest and disease distribution).

[0148] Plot View: Display growth data by plot, and click on the plot to view detailed information (such as historical growth trends, predicted yields).

[0149] Heat map: Generate spatial heat maps of health index and adversity index to visually display abnormal growth areas.

[0150] 6.3 Decision Support Unit

[0151] It is used to generate fertilization and irrigation suggestions, pest and disease control supervision suggestions, climate change trend analysis (analyze the long-term impact of climate change on cotton growth, such as the impact of increased drought frequency on yield), and pest and disease pattern analysis (use data mining algorithms, such as association rule mining, to discover the time and environmental condition patterns of pest and disease occurrence) according to the processing results of the image mining module, the comprehensive evaluation module, and the growth prediction module.

[0152] Therefore, by providing a cotton field growth status evaluation and prediction system based on UAV RGB images, the present invention can conduct a comprehensive analysis of cotton growth by combining multiple factors, improve the comprehensiveness, accuracy, and efficiency of evaluation and prediction, and can also conduct real-time monitoring and intelligent decision-making on the entire process of cotton growth, improving the efficiency and precision of agricultural management, and promoting the cotton precision agriculture to enter a new stage of digitalization and intelligentization.

[0153] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0154] Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, there will be changes in the specific implementation manner and application scope according to the idea of the present invention. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A cotton field growth status assessment and prediction system based on UAV RGB images, characterized in that: include: The data acquisition module is used to select a drone, equip the selected drone with an RGB camera, preset the drone route and image acquisition parameters according to the terrain of the plot, and then perform data acquisition to obtain initial image data; A data processing module, used for performing denoising, distortion correction, orthophoto stitching and illumination correction on the initial image data to obtain standard image data; An image mining module is used to perform image segmentation and target detection on the standard image data to obtain cotton target information, and then extract the canopy coverage, plant height, color characteristics and vegetation index, as well as texture and spatial structure characteristics of the cotton target information to obtain cotton profile characteristics and growth indicators; A comprehensive evaluation module is used to introduce environmental data, use the environmental data and the growth indicators to train a machine learning model to obtain a growth status evaluation model, and then use the growth status evaluation model to obtain a comprehensive growth status score and evaluation indicators; A growth prediction module is used to introduce yield-related factors based on the growth status evaluation model to make yield predictions, and to preset growth abnormality thresholds to provide growth abnormality warnings; The real-time monitoring and decision support module is used to use the cloud service platform to perform distributed storage and management of data, and use WebGIS technology to visualize the comprehensive evaluation results and growth prediction results of each cotton plot, and automatically generate decision support based on the comprehensive evaluation results and growth prediction results, and perform automated management based on the decision support; Wherein, the data acquisition module, the data processing module, the image mining module, the comprehensive evaluation module, the growth prediction module and the real-time monitoring and decision support module are interconnected.

2. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 1 is characterized in that: The data acquisition module comprises: A hardware configuration unit, used for selecting a drone, and equipping the selected drone with an RGB camera and a three-axis gimbal compatible with the RGB camera to achieve anti-shake performance of the camera; A route planning unit, used to automatically generate a flight path using the drone route planning software and set the forward overlap and lateral overlap of the image; The parameter setting unit is used to set the flight altitude and resolution of the UAV according to the growth stage of cotton, and then collect data to obtain initial image data.

3. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 2 is characterized in that: The data processing module comprises: A denoising and distortion correction unit, configured to perform denoising on the initial image data using an automatic batch processing tool and a denoising algorithm, and perform distortion correction on the image edge using camera calibration parameters; An orthophoto stitching unit is used to perform feature point matching and image stitching on overlapping images in the initial image data using aerial survey processing software, and output an orthophoto image with geographic coordinate system information; The illumination correction unit is used to deploy grayscale plates within the flight area of ​​the UAV, record standard reflection values, perform color correction on the initial image data based on the standard reflection values, and integrate a radiation correction algorithm in the aerial survey processing software to standardize image brightness and image color.

4. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 3 is characterized in that: The image mining module comprises: A data annotation unit, for performing pixel-level annotation based on the standard image data using an annotation tool, generating segmentation labels for cotton plants and soil background, and performing bounding box annotation for individual cotton plants and pest and disease areas; An image segmentation unit is used to select a convolutional neural network, introduce an attention mechanism, a multi-scale feature extraction module and a depth-separable convolution into the selected convolutional neural network to obtain a first initial model, then train the first initial model using a deep learning framework to obtain a segmentation model, use the segmentation model to segment the standard image data, and output a binary image; wherein the binary image includes the overall distribution information of the cotton canopy; A target detection unit, used to combine YOLOv5 and multi-task learning to obtain a target detection model, use the target detection model to detect the binary image, and output position information; wherein the position information includes the position of a single cotton plant or a pest and disease area; A result fusion unit, used for combining the binary image and the position information to obtain cotton target information; The profile feature and growth index extraction unit is used to extract the canopy coverage, plant height, color features and vegetation index as well as texture and spatial structure features of the cotton target information to obtain the profile features and growth index of the cotton target information.

5. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 4 is characterized in that: The profile feature and growth index extraction unit comprises: A canopy coverage calculation subunit is used to count the percentage of cotton canopy pixels in the binary image to obtain canopy coverage, and to evaluate whether the growth process is normal based on the canopy coverage and the cotton growth period; A plant height estimation subunit, used to estimate plant height by combining the orthophoto and the digital surface model, and output a spatial distribution map of plant height to identify abnormal growth areas; A texture and spatial structure feature extraction subunit is used to extract texture features based on the standard image data using a gray-level co-occurrence matrix, and calculate the population density and plant spacing of cotton plants to obtain spatial structure features; The color feature and vegetation index calculation subunit is used to extract the color features of the RGB bands in the standard image data, combine the texture features and the color features, design the RGB vegetation index, and then evaluate the health status of cotton and identify the pest and disease areas based on the spatial distribution of the RGB vegetation index.

6. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 5 is characterized in that: The comprehensive assessment module includes: A data fusion unit, used for acquiring soil data, meteorological data and geographic information data of the cotton plot to obtain environmental data, spatially aligning and temporally aligning the environmental data with the standard image data, and then fusing the environmental data with the extracted growth index to obtain a comprehensive data set; A model building unit, used to build an initial model using an integrated learning method, set input variables and output variables of the initial model, and use the comprehensive data set to perform model training to obtain a growth status assessment model; the input variables include canopy coverage, plant height, vegetation index, texture characteristics and environmental data, and the output variables include a comprehensive growth status score, a health index and an adversity index; The index quantification unit is used to generate a spatial distribution map according to the comprehensive growth status score, and to obtain evaluation indicators in combination with the health index and adversity index to conduct a comprehensive evaluation.

7. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 6 is characterized by: The expression of the comprehensive growth status score is: S=ω1·C+ω2·H+ω3·VI+ω4·T Among them, S is the comprehensive growth status score, C is the canopy coverage, H is the plant height, VI is the vegetation index, T is the texture feature, ω1, ω2, ω3, and ω4 are the weights of each index; The expression of the health index is: HI=α·VI+β·T Among them, HI is the health index, α and β are weight parameters; The expression of the adversity index is: AI=γ1·(1-E soil )+γ2·|E temp -T opt | Among them, AI is the adversity index, E soil is soil moisture, E temp is the temperature, T opt is the optimal growth temperature of cotton, γ1 and γ2 are weight parameters.

8. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 7 is characterized in that: The growth prediction module comprises: A yield prediction unit, for fusing the canopy coverage, the plant height, the vegetation index, the texture features, the environmental data and yield-related factors to obtain a yield prediction data set, using the yield prediction data set to perform model training to obtain a yield prediction model, and using the yield prediction model to output a single plant yield prediction value and a unit area yield prediction value; wherein the yield-related factors include plant density and fluff opening amount, and an environmental correction factor for adjusting the output value is introduced into the yield prediction model; The growth abnormality unit is used to set the comprehensive growth status score threshold, health index threshold and adversity index threshold according to the output value of the growth status evaluation model, and then judge whether the cotton plot is an abnormal area according to the set threshold. If so, early warning information is automatically generated.

9. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 8, characterized in that: The expression of the predicted value of single plant yield is: Y single =B·W boll Among them, Y single is the predicted value of single plant yield, B is the boll opening amount, W is boll is the weight of a single floc; The expression of the predicted value of unit area yield is: Y yield =D·Y single ·(1-k1·|E soil -E opt |-k2·|E temp -T opt |) Among them, Y yield is the predicted yield per unit area, D is the plant density, E opt is the optimal soil moisture, T opt is the optimal growth temperature, k1 and k2 are environmental correction coefficients.

10. The cotton field growth status assessment and prediction system based on drone RGB images according to claim 9, characterized in that: The real-time monitoring and decision support module includes: The data storage unit is used to store structured data using a relational database based on a cloud service platform, manage unstructured data using object storage, and access data in real time through an API interface; A monitoring and visualization unit, which is used to develop a responsive interface supporting multi-terminal access using a front-end framework, and to display the processing results of the image mining module, the comprehensive evaluation module and the growth prediction module in real time on the interface in the form of map overlay, plot view and heat map using WebGIS technology; A decision support unit is used to generate fertilization and irrigation suggestions, pest and disease prevention and control supervision suggestions, climate change trend analysis and pest and disease law analysis based on the processing results of the image mining module, the comprehensive evaluation module and the growth prediction module.

Images