A whole life cycle electronic information filing observation system for plants

Abstract

This invention discloses a plant lifecycle electronic information filing and observation system. It generates a unique electronic identification code for each seedling by integrating RFID and QR code dual identification systems, ensuring consistent identification throughout the entire process from seedling cultivation to harvest. It deploys intelligent camera components with multi-focus capabilities, enabling simultaneous capture of multiple plants by a single camera. It employs a unique automatic calibration algorithm for plant growth characteristics based on temporal feature fusion, continuously tracking and quantifying microscopic features such as root and stem spots and leaf morphology. The system automatically collects image data along a timeline through a backend system, generating time-lapse photographs of individual plant growth. It also incorporates an intelligent early warning mechanism based on mutation index calculation, enabling tiered alarms for abnormal diseases, thus solving the technical problems of information gaps during plant transfer across greenhouses, the inability to trace individual plants long-term, and the disconnect between growth characteristic observation and record management.

Description

Technical Field

[0001] This invention belongs to the field of smart agriculture technology, specifically relating to an electronic information filing and observation system for the entire life cycle of plants. Background Technology

[0002] The Venlo-style large-scale multi-span intelligent glass greenhouse has gradually formed an electronic and digital construction and operation mode in the long-term development process.

[0003] Under the existing Venlo-style large-scale multi-span intelligent glass greenhouse management model, key data is acquired through deployed sensors. The central control system compares the real-time data with the built-in optimal crop growth model. When a parameter exceeds the set threshold, a control command is automatically generated, which is then executed by various automated execution devices within the greenhouse.

[0004] Existing management systems mostly rely on manual, independent plant observation and recording, lacking a unified identification system. There is no integrated data system from seedbeds to seedling greenhouses, experimental greenhouses, and even actual production greenhouses. This results in each greenhouse conducting observations at the current stage independently, making it impossible to establish long-term tracking and observation of individual plants. When plants are moved across greenhouses, they need to be re-filed and registered. Paper records are fragmented from the data of each subsystem, making it impossible to trace the original growth information of seedlings, the specific characteristics of roots, stems, and leaves, and the early growth data. This makes it difficult to provide complete data support for mid-to-late stage cultivation management, disease tracing, and variety breeding. Manual records cannot automatically label characteristics such as root and stem spots, leaf morphology, and specific colors, and there is a lack of intelligent early warning and visual traceability methods. Summary of the Invention

[0005] To address the aforementioned technical problems, the present invention provides an electronic information archiving and observation system for the entire life cycle of plants, comprising:

[0006] The dual-identity binding module is used to assign a unique RFID electronic tag and QR code to each seedling, complete the initial electronic filing and identity binding, and enter basic information such as variety, seedling time and initial growth status. The cross-greenhouse identification module is used to automatically read RFID electronic tags and QR code codes during the transfer and growth of seedlings across greenhouses, ensuring seamless continuation of the electronic records of plants between different greenhouses; The multi-source visual acquisition module is used for simultaneous capture and automatic focusing of multiple plants under a single camera, continuously acquiring high-definition images and growth data of plants around the clock. The automatic feature calibration module is used to intelligently identify and automatically calibrate the personalized growth characteristics of plants based on time-series image sequences, and solidify the calibration results into the corresponding electronic file to form a unique growth ID card for the plant. The time-lapse photography generation module is used to automatically collect the time-series image data of plants according to the time axis, form a time-series image sequence, and generate a time-lapse photography video of the growth of a single plant from the seedling start point to the current time with one click when receiving instructions, and associate and highlight the marked special feature nodes in the video. The intelligent early warning module has a built-in pre-set early warning rule library that includes plant disease characteristics, abnormal growth thresholds, and environmental stress parameters. It is used to compare plant morphology, leaf condition and color changes in real time. When an abnormality is detected, it automatically triggers a graded alarm and pushes it to the management terminal. A unified data platform is used to connect and centrally store data from various greenhouse observation subsystems, including at least plant identification codes, feature calibration records, time-series image indexes, early warning logs, and greenhouse circulation trajectories, supporting the query, traceability, and analysis of data throughout the entire lifecycle.

[0007] Preferably, the personalized growth characteristics include the distribution of root and stem spots, the shape of leaf outline, the unique color of leaf surface, and the density of growth.

[0008] Preferably, the automatic feature calibration module executes a plant growth feature calibration algorithm based on temporal feature fusion, which includes the following steps: Preprocess the time-series image sequence to extract the overall plant image and color feature space; Based on inter-frame difference and morphological operations, the dynamic changes of spots in the root and stem region of the whole plant image are detected and quantified to generate spot distribution entropy to determine the possibility of lesions. Active contour model is used to extract leaf edges from the overall image of the plant, calculate the roundness and elongation of the leaves, and construct leaf morphological feature vectors to determine the possibility of growth abnormalities. Calculate the projection distribution of the overall plant image in the color feature space, determine the color overflow index, and judge the proportion of pixels on the plant whose color deviates significantly from the healthy standard. Based on spot distribution entropy, leaf morphology feature vector, and color overflow index, a comprehensive growth characteristic index is calculated to determine the plant's growth status.

[0009] Preferably, the formula for calculating the spot distribution entropy is as follows:

[0010] in, This is the rhizome region. Let x be the probability that pixel x belongs to the area of ​​blob variation.

[0011] Preferably, the formula for calculating the roundness is as follows:

[0012] in, For the leaf area, This is the circumference of the blade.

[0013] Preferably, the formula for calculating the elongation is as follows:

[0014] Among them, and These are the major and minor axes of the smallest circumscribed rectangle of the blade, respectively.

[0015] Preferably, the formula for calculating the color overflow index is as follows:

[0016] in, It is the color feature vector of pixel q. It is an indicator function. It is the total number of pixels. The mean vector of the standard color distribution of healthy plants. The standard deviation vector of the standard color distribution of healthy plants.

[0017] Preferably, the formula for calculating the comprehensive growth characteristic index is as follows:

[0018] in, For the entropy of the spot distribution, This is the leaf morphology feature vector. This represents the color overflow index. This represents the standard morphological vector of a healthy plant. α To represent the adjustable root and stem spot entropy weights for root and stem diseases, β Adjustable weights for differences in leaf and fruit morphology are used to represent the morphology of leaves and fruits. c Adjustable leaf and fruit color overflow weights are used to represent the colors of leaves and fruits.

[0019] Preferably, when the time-lapse photography generation module synthesizes the video, it generates clickable feature tags on the video progress bar based on the timestamps recorded by the feature auto-calibration module. When an administrator clicks on a tag, they can directly jump to the time point when the feature first appears or changes significantly.

[0020] This invention provides a plant lifecycle electronic information archiving and observation system. It integrates RFID and QR code dual identification systems to generate a unique electronic identification code for each seedling, ensuring consistent identification from seedling cultivation to harvest. The system deploys intelligent camera components with multi-focus capabilities, enabling simultaneous capture of multiple plants by a single camera. It employs a unique automatic calibration algorithm for plant growth characteristics based on temporal feature fusion, continuously tracking and quantifying microscopic features such as root and stem spots and leaf morphology. The system automatically collects image data along a timeline through a backend system, generating time-lapse photographs of individual plant growth. It also incorporates an intelligent early warning mechanism based on mutation index calculation, enabling tiered alarms for abnormal diseases. This addresses the technical problems of information gaps during plant transfers across greenhouses, the inability to trace individual plants long-term, and the disconnect between growth characteristic observation and record management. Detailed Implementation

[0021] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0022] This invention provides an electronic information filing and observation system for the entire life cycle of plants, applied to a Venlo-style large-scale multi-span smart glass greenhouse, covering the entire process of seedling greenhouse, experimental greenhouse, and production greenhouse. The system includes: The dual-identity binding module is configured at the seedling sorting station to assign a unique RFID electronic tag and QR code to each seedling, complete the initial electronic filing and identity binding, and enter basic information such as variety, seedling time and initial growth status.

[0023] A cross-greenhouse identification module is configured at the entrance and internal identification points of each greenhouse to automatically read the RFID electronic tags and QR code codes during the transfer and growth of seedlings across greenhouses, ensuring seamless continuation of the electronic records of plants between different greenhouses.

[0024] The multi-source visual acquisition module includes multiple intelligent camera components deployed in the greenhouse. The camera components are equipped with multi-focus technology for simultaneous capture and automatic focusing of multiple plants under a single camera, continuously acquiring high-definition images and growth data of the plants around the clock to obtain a time-series image sequence.

[0025] The automatic feature calibration module is connected to the Agricultural Eye Intelligent Detection System. It is used to intelligently identify and automatically calibrate the personalized growth characteristics of plants based on time-series image sequences, including the distribution of root and stem spots, leaf outline shape, leaf surface special color and growth density, and solidify the calibration results into the corresponding electronic file to form a unique growth ID card for the plant.

[0026] The automatic feature calibration module executes a plant growth feature calibration algorithm based on temporal feature fusion, which specifically includes the following steps: S1: For the input time-series image sequence Preprocessing was performed to extract the plant outline. With color feature space ; S2: Based on inter-frame differential and morphological operations, detect and quantify the dynamic changes of spots in the root region, and generate spot distribution entropy. ;

[0027] in, This is the rhizome region. Let x be the probability that pixel x belongs to the area of ​​blob variation. The higher the value, the more disordered the distribution of spots, and the higher the likelihood of lesions.

[0028] S3: Extract blade edges using an active contour model and calculate blade roundness. Elongation Construct leaf morphology feature vectors to comprehensively assess the likelihood of growth abnormalities: ; Roundness formula:

[0029] in, For the leaf area, This refers to the leaf circumference. Healthy leaves typically have a specific range of roundness.

[0030] Elongation formula:

[0031] Among them, and These are the major and minor axes of the smallest circumscribed rectangle of the blade, respectively. This value reflects the slenderness of the blade and is crucial for identifying deformed blades.

[0032] S4: Calculate the projection distribution of the overall plant image in a preset specific color space (such as Lab space) and determine the color overflow index. The formula is as follows:

[0033] in, It is the color feature vector of pixel q. It is an indicator function. It is the total number of pixels. The mean vector of the standard color distribution of healthy plants. The standard deviation vector of the standard color distribution of a healthy plant. This index calculates the proportion of pixels on the plant whose color significantly deviates from the healthy standard.

[0034] S5: Calculate the comprehensive growth characteristic index at time t using a weighted fusion function:

[0035] in, This represents the standard morphological vector of a healthy plant. α To represent the adjustable root and stem spot entropy weights for root and stem diseases, β Adjustable weights for differences in leaf and fruit morphology are used to represent the morphology of leaves and fruits. c Adjustable leaf and fruit color overflow weights are used to represent the colors of leaves and fruits. The higher the value, the worse the plant's growth and the more pronounced the abnormal characteristics.

[0036] The above-mentioned plant growth feature calibration algorithm based on temporal feature fusion solves the problem that existing technologies are unable to automatically quantify and track microscopic features such as root and stem spots and subtle morphological changes in leaves.

[0037] The time-lapse photography generation module is used to automatically collect time-series image data of plants according to the timeline, and when it receives an instruction, it can generate a time-lapse video of the growth of a single plant from the seedling start point to the current moment with one click, and associate and highlight the marked special feature nodes in the video.

[0038] When synthesizing videos, the time-lapse photography generation module generates clickable feature tags on the video progress bar based on the timestamps recorded by the feature auto-calibration module. Administrators can click on the tags to jump directly to the time point when the feature first appears or changes significantly.

[0039] The intelligent early warning module has a built-in pre-set early warning rule library that includes plant disease characteristics, abnormal growth thresholds, and environmental stress parameters. It is used to compare plant morphology, leaf condition and color changes in real time. When an abnormality is detected, it automatically triggers a graded alarm and pushes it to the management terminal.

[0040] The intelligent early warning module calculates the plant mutation index. For graded alarms, the calculation formula is as follows:

[0041] in, and These are the combined growth characteristic indices for the current time and the previous time, respectively. Let be the real-time normalized value of the i-th environmental sensor (e.g., temperature, humidity, light intensity). For their corresponding weights, This represents the total number of environmental sensors.

[0042] Core items This represents the rate of rapid, short-term change in the plant's condition. Taking the logarithm smooths out extreme changes while maintaining sensitivity to positive / negative shifts. This percentage increases significantly when growth conditions deteriorate drastically. Environmental weighting This represents the comprehensive evaluation value of current environmental factors. The more unsuitable the environment (the higher the comprehensive value), the stronger the "amplification" effect of this item.

[0043] The product of the core term and the environmental weighted term indicates that the warning index will only rise sharply when there is an unfavorable mutation in the plant's own characteristics and the current environmental conditions are also poor. This effectively eliminates false alarms caused by transient noise or non-critical environmental fluctuations.

[0044] When the plant mutation index When the plant mutation index exceeds the first preset threshold, the system issues a level three (minor) warning; when the plant mutation index... When the plant mutation index exceeds the second preset threshold and the second preset threshold is greater than the first preset threshold, the system issues a level two (moderate) warning; when the plant mutation index... When the third preset threshold is exceeded and the third preset threshold is greater than the second preset threshold, the system issues a Level 1 (Severe) warning and automatically associates the time-lapse video and feature labeling record of the plant.

[0045] A unified data platform is used to connect and centrally store data from various greenhouse observation subsystems, including at least plant identification codes, feature calibration records, time-series image indexes, early warning logs, and greenhouse circulation trajectories, supporting the query, traceability, and analysis of data throughout the entire lifecycle.

[0046] This application also provides a method for electronically documenting and observing the entire life cycle of a plant, using the electronically documenting and observing system for the entire life cycle of a plant as described above. The method steps are as follows: Step 1: Identity Anchoring - In the seedbed sorting process, each seedling is bound with a unique identity identifier through RFID reading and writing and QR code coding equipment, basic information is entered, and an initial electronic file is generated. Step 2: Global Deployment - Deploy intelligent camera acquisition terminals in each greenhouse zone, build a back-end management system and agricultural eye intelligent detection algorithm module, and complete the connection of all software and hardware interfaces; Step 3: Continuous Observation - When seedlings are transferred across greenhouses, their historical electronic records are automatically retrieved through identification, and new growth observation data is continuously added in the current greenhouse to achieve seamless continuation of the records; Step 4: Intelligent calibration and video synthesis - The system automatically executes the above-mentioned plant growth feature calibration algorithm based on temporal feature fusion to continuously track and calibrate plant features, archives images according to time sequence, and automatically synthesizes long-time-lapse video. Step 5: Early Warning and Intervention - The background system calculates the plant mutation index and issues early warnings in real time. When an anomaly occurs, an alarm is pushed out. Managers combine time-lapse images and feature calibration records to formulate cultivation strategy adjustments and disease control plans. Step Six: Full Lifecycle Management - All archiving data, image indexes, calibration records, and early warning logs are centrally stored in a unified data platform, supporting access, traceability, and export throughout the entire lifecycle.

[0047] The beneficial effects of the embodiments of the present invention are as follows: 1. Seamless Identity and Information: Through the dual identification system of RFID and QR code, the system has achieved the first full-cycle identity verification of a single plant from the seedling greenhouse to the production greenhouse in the industry, completely solving the problem of information gaps in cross-greenhouse circulation.

[0048] 2. Intelligent observation and feature quantification: The original algorithm based on time-series feature fusion transforms the original human experience-dependent judgment of features such as root and stem spots and leaf morphology into quantifiable indices, enabling continuous, automatic, and high-precision tracking of microscopic growth changes in plants.

[0049] 3. Intuitive traceability and efficient decision-making: Automatically generates time-lapse photography of growth with feature tags, enabling managers to "replay" the complete growth process of crops over several months in minutes, accurately pinpoint the time point when problems occur, and provide revolutionary visual data support for cultivation management, disease tracing, and variety breeding.

[0050] 4. Precise early warning and timely intervention: The early warning model, which combines plant mutations with environmental factors, significantly improves the accuracy of disease and abnormal growth warnings, and realizes the upgrade of smart agricultural management from passive response to active intervention.

Claims

1. A plant life-cycle electronic information archiving and observation system, characterized in that, include: The dual-identity binding module is used to assign a unique RFID electronic tag and QR code to each seedling, complete the initial electronic filing and identity binding, and enter basic information such as variety, seedling time and initial growth status. The cross-greenhouse identification module is used to automatically read RFID electronic tags and QR code codes during the transfer and growth of seedlings across greenhouses, ensuring seamless continuation of the electronic records of plants between different greenhouses; The multi-source visual acquisition module is used for simultaneous capture and automatic focusing of multiple plants under a single camera, continuously acquiring high-definition images and growth data of plants around the clock. The automatic feature calibration module is used to intelligently identify and automatically calibrate the personalized growth characteristics of plants based on time-series image sequences, and solidify the calibration results into the corresponding electronic file to form a unique growth ID card for the plant. The time-lapse photography generation module is used to automatically collect the time-series image data of plants according to the time axis, form a time-series image sequence, and generate a time-lapse photography video of the growth of a single plant from the seedling start point to the current time with one click when receiving instructions, and associate and highlight the marked special feature nodes in the video. The intelligent early warning module has a built-in pre-set early warning rule library that includes plant disease characteristics, abnormal growth thresholds, and environmental stress parameters. It is used to compare plant morphology, leaf condition and color changes in real time. When an abnormality is detected, it automatically triggers a graded alarm and pushes it to the management terminal. A unified data platform is used to connect and centrally store data from various greenhouse observation subsystems, including at least plant identification codes, feature calibration records, time-series image indexes, early warning logs, and greenhouse circulation trajectories, supporting the query, traceability, and analysis of data throughout the entire lifecycle.

2. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 1, characterized in that, The personalized growth characteristics include the distribution of root and stem spots, leaf outline shape, unique leaf color, and density of growth.

3. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 1, characterized in that, The automatic feature calibration module executes a plant growth feature calibration algorithm based on temporal feature fusion, which includes the following steps: Preprocess the time-series image sequence to extract the overall plant image and color feature space; Based on inter-frame difference and morphological operations, the dynamic changes of spots in the root and stem region of the whole plant image are detected and quantified to generate spot distribution entropy to determine the possibility of lesions. Active contour model is used to extract leaf edges from the overall image of the plant, calculate the roundness and elongation of the leaves, and construct leaf morphological feature vectors to determine the possibility of growth abnormalities. Calculate the projection distribution of the overall plant image in the color feature space, determine the color overflow index, and judge the proportion of pixels on the plant whose color deviates significantly from the healthy standard. Based on spot distribution entropy, leaf morphology feature vector, and color overflow index, a comprehensive growth characteristic index is calculated to determine the plant's growth status.

4. The electronic information filing and observation system for the entire life cycle of a plant as described in claim 3, characterized in that, The formula for calculating the speckle distribution entropy is as follows: in, This is the rhizome region. Let x be the probability that pixel x belongs to the area of ​​blob variation.

5. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 3, characterized in that, The formula for calculating the roundness is as follows: in, For the leaf area, This is the circumference of the blade.

6. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 3, characterized in that, The formula for calculating the elongation is as follows: Among them, and These are the major and minor axes of the smallest circumscribed rectangle of the blade, respectively.

7. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 3, characterized in that, The formula for calculating the color overflow index is as follows: in, It is the color feature vector of pixel q. It is an indicator function. It is the total number of pixels. The mean vector of the standard color distribution of healthy plants. The standard deviation vector of the standard color distribution of healthy plants.

8. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 3, characterized in that, The formula for calculating the comprehensive growth characteristic index is as follows: in, For the entropy of the spot distribution, This is the leaf morphology feature vector. This refers to the color overflow index. This represents the standard morphological vector of a healthy plant. α To represent the adjustable root and stem spot entropy weights for root and stem diseases, β Adjustable weights for differences in leaf and fruit morphology are used to represent the morphology of leaves and fruits. γ Adjustable leaf and fruit color overflow weights are used to represent the colors of leaves and fruits.

9. The electronic information archiving and observation system for the entire life cycle of a plant as described in claim 1, characterized in that, When synthesizing videos, the time-lapse photography generation module generates clickable feature tags on the video progress bar based on the timestamps recorded by the feature auto-calibration module. Administrators can click on the tags to jump directly to the time point when the feature first appears or changes significantly.

Images