An agricultural operation assistance image analysis method based on image processing technology

Abstract

The patent application relates to the technical field of intelligent agriculture, and discloses an agricultural operation assistance image analysis method based on image processing technology, which comprises the following steps: S1, collecting image data of a farmland area through a multispectral imaging device; S2, performing a pretreatment operation on the collected image data; S3, performing feature layer fusion processing on the pretreated multi-band image; S4, performing crop growth state analysis based on the fused features; S5, combining environmental sensor data to perform comprehensive evaluation on the crop health condition; S6, generating farmland operation suggestions according to the growth state and the health condition; and S7, displaying the analysis results and the operation suggestions through a visual interface. Through the cooperative collection of the multispectral imaging device, the method realizes the synchronous acquisition of visible light, near-infrared and thermal infrared data, solves the problem of insufficient collaborative utilization of multi-source data, and establishes a multi-dimensional crop information collection system.

Description

Technical Field

[0001] This invention patent relates to the field of smart agriculture technology, specifically to an image analysis method for agricultural operation assistance based on image processing technology. Background Technology

[0002] In modern agricultural operations, image analysis technology has been widely used in crop monitoring. The existing technologies used in agriculture mainly include: visible light image analysis technology, multispectral imaging technology, and thermal imaging technology.

[0003] The main problems with existing technologies are: insufficient collaborative utilization of multi-source data, independent analysis of visible light, near-infrared and thermal infrared data, single crop health assessment indicators, fragmented analysis of morphological and physiological parameters, incomplete pest and disease early warning mechanisms, lack of early identification capabilities, inaccurate agricultural decision support, insufficient quantification, unintuitive presentation of analysis results, and high application threshold for farmers.

[0004] In view of this, we propose an image analysis method for agricultural operation assistance based on image processing technology.

[0005] Invention Patent Content The purpose of this invention is to provide an image analysis method for agricultural operation assistance based on image processing technology, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this invention provides the following technical solution: An image analysis method for agricultural operations assistance based on image processing technology, the method comprising: S1. Acquire image data of farmland areas using multispectral imaging equipment, wherein the image data includes visible light, near-infrared and thermal infrared bands; S2. Perform preprocessing operations on the acquired image data, including image registration, noise reduction, and contrast enhancement; S3. Perform feature layer fusion processing on the preprocessed multi-band image; S4. Analyze crop growth status based on fusion features and extract plant morphological parameters and physiological indicators; S5. Conduct a comprehensive assessment of crop health status by combining environmental sensor data; S6. Generate agricultural operation suggestions based on growth status and health condition; S7. Display analysis results and operational suggestions through a visual interface.

[0007] Preferably, the feature layer fusion processing in step S3 adopts a multi-scale fusion strategy, which first performs pyramid decomposition on each band image, and then fuses feature information at different scales.

[0008] Preferably, the fusion processing in step S3 employs an improved weighted fusion algorithm, the specific formula of which is: ; in, Represents the eigenvalues ​​after fusion. This represents the variance value of the i-th band image in a local region. This represents the feature value of the i-th band image.

[0009] Preferably, the plant morphological parameters in step S4 include plant height, crown width, and leaf tilt angle distribution, and the physiological indicators include chlorophyll content and water stress index.

[0010] Preferably, the health status assessment in step S5 employs a multi-factor fusion algorithm, the core calculation model of which is: ; in, For the coronal health index, This represents the k-th visible vegetation feature. This represents the difference between canopy temperature and air temperature. and For feature transformation function, and These are the weighting coefficients.

[0011] Preferably, the agricultural operation suggestions in step S6 include: Fertilization recommendations are calculated based on the difference between the fertilizer requirement for the target yield and the soil nutrient content; Irrigation recommendations should be determined based on canopy temperature distribution and soil moisture gradient; The recommendations for pest and disease control are generated by combining historical trends in the health index.

[0012] Preferably, step S6 further includes constructing a crop growth trend prediction model, the model taking into input historical growth data, current growth status and environmental parameters, and outputting a future growth trend prediction curve.

[0013] Preferably, the visualization interface in step S7 includes: a spatiotemporal distribution heat map showing the risk areas of pests and diseases, a three-dimensional growth model displaying crop morphological characteristics, and an intelligent agricultural calendar marking key operation time points.

[0014] Preferably, the method further includes a mobile terminal application module that supports real-time early warning information push, voice broadcast of agricultural operation guidance, and generation of language analysis reports.

[0015] Preferably, the method further includes a data management module for storing a historical image feature database, an environmental parameter time series database, and agricultural operation effect evaluation records.

[0016] By employing the above technical solution, this invention patent provides an image analysis method for agricultural operation assistance based on image processing technology. It possesses at least the following beneficial effects: This invention patent achieves simultaneous acquisition of visible light, near-infrared, and thermal infrared data through the collaborative acquisition of multispectral imaging equipment, solving the problem of insufficient collaborative utilization of multi-source data and establishing a multi-dimensional crop information acquisition system. Through a multi-scale feature fusion strategy and an improved weighted fusion algorithm, it solves the problem of independent analysis of multi-source data, achieving the organic integration of morphological and physiological features and improving the accuracy of feature extraction. Through comprehensive analysis of plant morphological parameters and physiological indicators, it overcomes the problem of fragmented analysis of morphological and physiological parameters in traditional methods, achieving a comprehensive assessment of crop growth status and providing data support for precision agriculture. Through a multi-factor fusion health assessment model, it integrates image features and environmental parameters, solving the problem of single health assessment indicators, significantly improving the early identification capability of pests and diseases, and effectively improving the accuracy of early warning. It also provides quantitative agricultural recommendations. The generation mechanism, based on precise nutrient balance and water requirement calculations, solves the problem of inaccurate agricultural decision support, enabling quantitative recommendations for operations such as fertilization and irrigation. Through a 3D visualization interface, employing innovative display formats such as spatiotemporal distribution heatmaps and intelligent agricultural calendars, it addresses the issue of unintuitive presentation of analysis results, significantly lowering the application threshold for farmers. The mobile terminal application module supports multilingual voice broadcasts and early warning push notifications, resolving the applicability issue of technology promotion and enabling farmers with different educational backgrounds to easily access professional guidance. The data management module constructs a complete agricultural data asset system, solving the problem of low utilization of historical data and providing a data foundation for continuous model optimization. The growth trend prediction model enables dynamic simulation of crop growth processes, solving the problem of inaccurate predictions based on traditional experience and providing a scientific basis for agricultural planning. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention, form part of this application: Figure 1 This is a flowchart illustrating an image analysis method for agricultural operation assistance based on image processing technology, as per this invention patent. Detailed Implementation

[0018] The technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0019] Please see Figure 1 , Figure 1 This is a flowchart illustrating an image analysis method for agricultural operations assistance based on image processing technology, as described in this invention. Specifically, the image analysis method for agricultural operations assistance based on image processing technology includes: S1. Acquire image data of farmland areas using multispectral imaging equipment, wherein the image data includes visible light, near-infrared and thermal infrared bands; S2. Perform preprocessing operations on the acquired image data, including image registration, noise reduction, and contrast enhancement; S3. Perform feature layer fusion processing on the preprocessed multi-band image; S4. Analyze crop growth status based on fusion features and extract plant morphological parameters and physiological indicators; S5. Conduct a comprehensive assessment of crop health status by combining environmental sensor data; S6. Generate agricultural operation suggestions based on growth status and health condition; S7. Display analysis results and operational suggestions through a visual interface.

[0020] It should be noted that the multispectral imaging device in S1 adopts an integrated optical design, which realizes simultaneous imaging of visible light, near infrared and thermal infrared bands through a beam splitter prism. The acquisition system is equipped with a high-precision gimbal stabilization device to ensure imaging stability during flight. Image acquisition follows the physiological rhythm of crops and avoids reflection distortion caused by strong light periods. The acquisition area planning adopts an intelligent trajectory algorithm, which automatically generates the optimal flight path according to the farmland boundary to ensure full coverage without blind spots. Image registration in S2 employs multimodal feature matching technology. First, stable feature descriptors for each band of the image are extracted, and a correspondence is established through feature space mapping. Then, the optimal spatial transformation parameters are solved. Noise reduction utilizes a frequency-domain processing strategy: adaptive smoothing filtering is used for low-frequency components, while edge-preserving filtering is used for high-frequency components. Contrast enhancement employs multi-scale illumination compensation, separating the illumination and reflection components and performing nonlinear enhancement processing separately to effectively improve the detail recognition in shadow areas. The multi-scale fusion architecture in S3 includes three stages: pyramid decomposition, hierarchical fusion, and reconstruction. Pyramid decomposition uses Gaussian kernel convolution to achieve multi-scale spatial sampling and establish an image representation system from coarse to fine. Hierarchical fusion implements a differentiated strategy: focusing on spectral feature preservation at low-resolution levels and spatial structure preservation at high-resolution levels. The reconstruction process uses an iterative back projection algorithm to optimize the fusion result by minimizing the reconstruction error. In the weighted fusion algorithm, the local variance calculation uses a sliding window statistical method, and the window size is adaptively adjusted according to the image resolution to ensure effective capture of texture features. In S4, 3D reconstruction employs multi-view geometry principles, reduces the matching search space through epipolar constraints, and utilizes a semi-global optimization algorithm to solve the disparity map. Plant height measurement is based on the difference between the ground elevation model and the canopy surface model, while canopy width measurement is achieved through the calculation of the convex hull of projected contour feature points. Physiological parameter inversion establishes an inversion framework combining physical and statistical models: chlorophyll content is inverted through spectral characteristics at red-edge locations, and the water stress index is calculated using a canopy temperature and air saturation temperature difference model, with physiological verification based on stomatal conductance theory. The multi-factor fusion architecture in S5 includes three core modules: feature standardization, weight allocation, and decision output. Feature standardization uses a piecewise normalization method to eliminate scale differences between features of different dimensions. The weight allocation mechanism introduces feature importance assessment and quantifies the contribution of each feature through information entropy theory. The health assessment model adopts a cascaded decision structure: the primary classifier distinguishes between healthy and unhealthy states, the secondary classifier identifies specific stress types in unhealthy samples, and the model training adopts a transfer learning strategy, using a public dataset to pre-train the basic model and then fine-tuning the parameters through specific crop samples. In S6, fertilization decisions are based on a nutrient balance model that comprehensively considers crop nutrient requirements, soil nutrient supply characteristics, and environmental nutrient loss. Irrigation decisions adopt a water stress response mechanism that triggers irrigation signals through dual indicators: the magnitude of canopy temperature increase and the rate of change in stomatal conductance. Pest and disease control is based on a spatiotemporal propagation model that identifies disease centers through spatial autocorrelation analysis of health indices and predicts risk areas by combining a meteorological diffusion model. Growth prediction adopts a state-space model that integrates historical growth trajectories and environmental driving factors to construct a dynamic simulation system for the growth process. The 3D visualization engine in S7 adopts a hierarchical rendering architecture: the base layer loads geographic information data, the intermediate layer overlays multiple periods of remote sensing images, the interactive layer integrates analysis results, the heat map generation uses a kernel density estimation algorithm, and the spatial smoothness is adjusted by the bandwidth parameter. The intelligent agricultural calendar is based on crop phenology models and combines accumulated temperature theory to predict growth period nodes. It integrates meteorological early warning data to achieve dynamic adjustment. The multi-view linkage adopts a message bus mechanism and supports cross-view status synchronization and data drill-down.

[0021] The feature layer fusion process in step S3 adopts a multi-scale fusion strategy. First, the images of each band are decomposed into pyramids, and then the feature information is fused at different scales.

[0022] It should be noted that the feature layer fusion processing adopts multi-scale pyramid decomposition and reconstruction technology. First, the images of each band are decomposed into multiple scales using Gaussian pyramids. Then, feature weighted fusion is performed in different scale spaces. Finally, the fusion result is generated using the Laplacian pyramid reconstruction algorithm.

[0023] The fusion process in step S3 employs an improved weighted fusion algorithm, the specific formula of which is: ; in, Represents the eigenvalues ​​after fusion. This represents the variance value of the i-th band image in a local region. This represents the feature value of the i-th band image.

[0024] It should be noted that in the improved weighted fusion algorithm, the local region variance value is obtained by calculating through a sliding window. The window size is adaptively adjusted according to the image resolution to ensure that regions with rich textures receive higher fusion weights while maintaining the integrity of the image edge structure.

[0025] The plant morphological parameters in step S4 include plant height, crown width, and leaf tilt angle distribution, while the physiological indicators include chlorophyll content and water stress index.

[0026] It should be noted that the plant morphology parameters were extracted using three-dimensional point cloud reconstruction technology. A crop spatial point cloud model was generated by multi-view image matching, and parameters such as plant height, crown width, and leaf tilt angle distribution were calculated based on this model. Physiological indicators were calculated by inverting the correlation between multispectral reflectance and physiological parameters.

[0027] The health status assessment in step S5 uses a multi-factor fusion algorithm, the core calculation model of which is: ; in, For the coronal health index, This represents the k-th visible vegetation feature. This represents the difference between canopy temperature and air temperature. and For feature transformation function, and These are the weighting coefficients.

[0028] It should be noted that in the multi-factor fusion algorithm, the feature transformation functions f and g adopt a piecewise linear mapping method to convert input features of different dimensions into standardized evaluation values; the weight coefficients α and β are dynamically adjusted through feature importance analysis to ensure that key features receive appropriate weights in the health assessment.

[0029] The agricultural operation suggestions in step S6 include: Fertilization recommendations are calculated based on the difference between the fertilizer requirement for the target yield and the soil nutrient content; Irrigation recommendations should be determined based on canopy temperature distribution and soil moisture gradient; The recommendations for pest and disease control are generated by combining historical trends in the health index.

[0030] It should be noted that the agricultural operation suggestions are generated using a decision tree reasoning mechanism. Based on the growth models of different crops and environmental parameters, a multi-dimensional decision rule base is constructed, which includes nutrient balance, water requirements, and pest and disease risks, to achieve intelligent agricultural planning.

[0031] Step S6 also includes constructing a crop growth trend prediction model, which takes into account historical growth data, current growth status and environmental parameters, and outputs a future growth trend prediction curve.

[0032] It should be noted that the crop growth trend prediction model uses time series analysis. By integrating historical growth curves, current growth indicators, and environmental factor change trends, it constructs a dynamic prediction model based on LSTM neural networks and outputs future growth trend curves.

[0033] The visualization interface in step S7 includes: a spatiotemporal distribution heat map showing the risk areas of pests and diseases, a three-dimensional growth model displaying crop morphological characteristics, and an intelligent agricultural calendar marking key operation time points.

[0034] It should be noted that the visualization interface uses multi-layer overlay technology to achieve the linked display of spatiotemporal distribution heatmaps, 3D growth models, and agricultural calendars; it also supports switching between different data dimension visualization presentation modes through interactive operations.

[0035] It also includes a mobile terminal application module, which supports real-time early warning information push, voice broadcast of agricultural operation guidance, and generation of language analysis reports.

[0036] It should be noted that the mobile terminal application module adopts a hybrid development framework, integrating a real-time message push engine and a multilingual speech synthesis system, and supports automatically switching the information display mode and voice broadcast language according to user preferences; The early warning push uses context-aware technology to automatically filter relevant early warning information based on the user's geographical location. The voice broadcast system integrates acoustic and language models, supports dialect adaptive synthesis, and uses a template engine architecture to generate multilingual reports, separating the data layer and presentation layer, and achieving multi-format output through XSLT transformation.

[0037] It also includes a data management module for storing historical image feature databases, environmental parameter time-series databases, and agricultural operation effect evaluation records.

[0038] It should be noted that the data management module adopts a distributed storage architecture and establishes a multi-level indexing mechanism based on spatiotemporal dimensions, supporting efficient querying of historical image features, environmental parameter time-series data, and records of agricultural operation effects. The distributed storage system employs a spatiotemporal partitioning strategy, dividing spatial units into latitude and longitude grids and time units into growth cycles. The feature database establishes a multi-level indexing mechanism: the primary index is based on spatiotemporal range, the secondary index on crop type, and the tertiary index on analytical indicators. The version control system records the entire data processing workflow, supporting historical version backtracking and difference comparison.

[0039] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0040] Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing agricultural operation assistance images based on image processing technology, characterized in that, The method includes: S1. Acquire image data of farmland areas using multispectral imaging equipment, wherein the image data includes visible light, near-infrared and thermal infrared bands; S2. Perform preprocessing operations on the acquired image data, including image registration, noise reduction, and contrast enhancement; S3. Perform feature layer fusion processing on the preprocessed multi-band image; S4. Analyze crop growth status based on fusion features and extract plant morphological parameters and physiological indicators; S5. Conduct a comprehensive assessment of crop health status by combining environmental sensor data; S6. Generate agricultural operation suggestions based on growth status and health condition; S7. Display analysis results and operational suggestions through a visual interface.

2. The agricultural operation assistance image analysis method based on image processing technology according to claim 1, characterized in that, The feature layer fusion process in step S3 adopts a multi-scale fusion strategy. First, the images of each band are decomposed into pyramids, and then the feature information is fused at different scales.

3. The agricultural operation assistance image analysis method based on image processing technology according to claim 1, characterized in that, The fusion process in step S3 employs an improved weighted fusion algorithm, the specific formula of which is: ； in, Represents the eigenvalues ​​after fusion. This represents the variance value of the i-th band image in a local region. This represents the feature value of the i-th band image.

4. The agricultural operation assistance image analysis method based on image processing technology according to claim 1, characterized in that, The plant morphological parameters in step S4 include plant height, crown width, and leaf tilt angle distribution, while the physiological indicators include chlorophyll content and water stress index.

5. The agricultural operation assistance image analysis method based on image processing technology according to claim 1, characterized in that, The health status assessment in step S5 employs a multi-factor fusion algorithm, the core calculation model of which is: ； in, For the coronal health index, This represents the k-th visible vegetation feature. This represents the difference between canopy temperature and air temperature. and For feature transformation function, and These are the weighting coefficients.

6. The agricultural operations assistance image analysis method based on image processing technology according to claim 5, characterized in that, The agricultural operation suggestions in step S6 include: Fertilization recommendations are calculated based on the difference between the fertilizer requirement for the target yield and the soil nutrient content; Irrigation recommendations should be determined based on canopy temperature distribution and soil moisture gradient; The recommendations for pest and disease control are generated by combining historical trends in the health index.

7. The agricultural operations assistance image analysis method based on image processing technology according to claim 6, characterized in that, Step S6 further includes constructing a crop growth trend prediction model, which takes historical growth data, current growth status and environmental parameters as inputs and outputs a future growth trend prediction curve.

8. The agricultural operations assistance image analysis method based on image processing technology according to claim 6, characterized in that, The visualization interface in step S7 includes: a spatiotemporal distribution heat map showing the risk areas of pests and diseases, a three-dimensional growth model displaying crop morphological characteristics, and an intelligent agricultural calendar marking key operation time points.

9. The agricultural operations assistance image analysis method based on image processing technology according to claim 1, characterized in that, The method also includes a mobile terminal application module, which supports real-time early warning information push, voice broadcast of agricultural operation guidance, and generation of language analysis reports.

10. The agricultural operations assistance image analysis method based on image processing technology according to claim 1, characterized in that, The method also includes a data management module for storing a historical image feature database, an environmental parameter time series database, and agricultural operation effect evaluation records.

Images