A crop growth detection system and method based on machine vision

Abstract

The present invention relates to the field of intelligent agricultural monitoring technology, specifically a machine vision-based crop growth detection system and method. The system includes an image acquisition module, a morphological feature capture module, a growth trend determination module, an abnormal area marking module, and a status information output module. In the present invention, crop images are captured by multi-angle cameras and leaf contours, stem curvature, and plant spacing are extracted to achieve multidimensional morphological modeling, enhance structural change recognition, compare key morphological features over time series, improve recognition accuracy and timeliness, continuously analyze leaf and stem change trends, quantify offsets, and provide early warning of growth anomalies. Health degradation identification is combined with fluctuation zone marking to achieve dynamic monitoring, high-risk positioning of time-overlapping areas, strengthen spatiotemporal locking, and facilitate precise management. This process, from image analysis to anomaly focusing, establishes an information reasoning chain, significantly improving monitoring accuracy and response efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent agricultural monitoring, and in particular to a crop growth detection system and method based on machine vision. Background Art

[0002] The field of intelligent agricultural monitoring technology involves leveraging advanced automation and information technology to collect, analyze, and process data from agricultural production, enabling real-time monitoring and precise management of crop growth environments and status. Core elements of this technology include agricultural environmental perception, crop growth status monitoring, pest and disease prediction and control, and intelligent agricultural production decision support. By utilizing sensors, drones, satellite remote sensing, and other equipment, combined with data processing and analysis platforms, intelligent agricultural monitoring technology monitors factors such as farmland environmental changes, soil quality, crop growth, and climate, driving the development of agriculture towards precision, efficiency, and sustainable development.

[0003] The crop growth monitoring system utilizes sensors, image processing, and the Internet of Things (IoT) to monitor crop growth in real time. Field-based monitoring equipment collects data such as soil moisture, temperature, light intensity, climate conditions, and crop growth images, continuously tracking crop growth. The system's design considers data collection and transmission, and incorporates image recognition technology to accurately determine crop growth progress and health. The system utilizes automated control technology for real-time data analysis and processing, providing decision support for agricultural management.

[0004] Existing technologies mainly rely on fixed sensors to collect environmental parameters and image data. The processing paths are mostly based on static analysis after data collection. The judgment of crop status in the image lacks the ability to recognize multi-dimensional spatial structures. It relies only on changes in leaf color or single images to infer growth status, making it difficult to fully reflect the true morphological evolution of crops. In addition, the current system has a lag in processing the time relationship of data, lacks real-time capture and marking of key growth nodes, cannot accurately track growth turning points on the time axis, and is prone to missing early signals of growth anomalies. The assessment of crop trends generally uses average values ​​or thresholds to judge, which cannot form a quantitative quantification of growth change trends in consecutive stages, resulting in ambiguous judgments on growth degradation and easy loss of accuracy in management interventions. In terms of spatial recognition, common systems often use regional mean methods for problem areas in farmland, lack dynamic anomaly focusing based on morphological trend evolution, and find it difficult to achieve local risk locking in high-density crop environments. For example, if growth abnormalities occur in a crop field due to local light shading, traditional systems find it difficult to accurately mark them based on local structural changes, which can easily lead to missed abnormalities or misjudgments. Existing technologies are insufficient in multi-dimensional data recognition, time-series trend tracking, and risk focus, limiting their practicality and responsiveness in complex agricultural environments. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a crop growth detection system and method based on machine vision.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A crop growth detection system based on machine vision includes:

[0007] The image acquisition module uses multi-angle cameras set up in the field to capture crop canopy images, extract leaf edge contour changes, stem curvature, and plant spacing distribution, and then generates a crop morphological feature sequence by matching the spatial distribution characteristics of these three parameters with the overall growth status of the crop.

[0008] The morphological feature capture module reads the crop morphological feature sequence and detects whether the corresponding node position in the time axis is consistent with the critical point of the crop growth state change. If the time coincides, the corresponding morphological feature is marked as a key feature and a crop morphological key node list is generated;

[0009] The growth trend determination module uses the crop morphology key node list to collect crop image data for two consecutive cycles, compares the leaf expansion area and stem elongation length in each cycle, and evaluates the change amplitude of the two during the cycle. If the change direction changes, it is compared with the original growth benchmark to obtain crop growth trend deviation information;

[0010] The abnormal area marking module detects whether the crop health status has degraded within a corresponding time period based on the crop growth trend deviation information, marks the abnormal area, and generates a crop growth fluctuation area set.

[0011] As a further solution of the present invention, the crop morphological feature sequence includes morphological distribution characteristics, consistency matching frequency, and growth state stability; the crop morphological key node list includes key feature names, critical points of growth state changes, and time node mapping relationships; the crop growth trend offset information includes change amplitude, turning direction, and baseline deviation status; and the crop growth fluctuation area set includes turning nodes, health degradation marking areas, and key fluctuation areas within the cycle.

[0012] As a further solution of the present invention, the image acquisition module includes:

[0013] The viewing angle adjustment submodule uses a multi-angle camera set up in the field to obtain crop canopy images, adjusts the camera pitch angle and focal length, extracts leaf edge contours, stem curvature, and plant spacing distribution, and generates a crop spatial distribution image sequence;

[0014] The feature extraction submodule analyzes the changing trends of leaf edge contour changes, stem curvature, and plant spacing distribution over a continuous period of time based on the crop spatial distribution image sequence, compares the changing directions of the three with the overall growth status of the crop, calculates the frequency of consistent occurrences, and obtains a frequency sequence of consistent morphology and growth status;

[0015] The sequence screening submodule selects the morphological parameters with the best occurrence frequency according to the frequency sequence of the consistent morphology and growth state, identifies the corresponding original parameters, and establishes the crop morphological feature sequence.

[0016] As a further solution of the present invention, the morphological feature capturing module includes:

[0017] The node positioning submodule obtains the time points corresponding to the morphological parameters in the crop morphological feature sequence, extracts the parameter extreme value moments as representative nodes, sets a periodic time axis, uses three adjacent periods as sliding windows, identifies the local extreme value points in each sequence and marks the key nodes, and obtains a set of morphological key time nodes;

[0018] The overlap detection submodule extracts the growth state value corresponding to the time node based on the set of key morphological time nodes, determines that the sign of the change before and after the moment is reversed and exceeds the fluctuation threshold as the critical point, records the critical point time, analyzes the correlation between the morphological features and the area where the critical point time overlaps, identifies the morphological features with a correlation higher than the threshold, and obtains the matching morphological feature set;

[0019] The feature registration submodule marks the features and morphological parameters of the overlapping time points according to the matching morphological feature set, calculates the key feature values, records the corresponding time nodes, integrates the index, morphological name, time point number and growth status fluctuation direction value of the key features, and establishes a list of crop morphological key nodes.

[0020] As a further solution of the present invention, the growth trend determination module includes:

[0021] The area comparison submodule obtains crop image data for two consecutive cycles through the crop morphology key node list, classifies each cycle according to leaf expansion area and stem elongation length, compares the measured values ​​of each category time node, identifies the change values ​​of leaf area and stem elongation, and arranges them by cycle to generate a periodic growth change sequence;

[0022] The direction assessment submodule evaluates the direction of change between the current and previous cycle data based on the periodic growth change sequence, identifies and records the turning phenomenon, analyzes the corresponding deviation amplitude of leaves and stems, and obtains the growth trend deviation degree;

[0023] The benchmark comparison submodule extracts the growth change data of the original continuous cycle of the crop based on the growth trend deviation, identifies the growth change benchmark range, compares the deviation of the current cycle with the growth change benchmark range, and if it exceeds the benchmark range, marks the corresponding cycle as an abnormal growth fluctuation section to obtain the crop growth trend deviation information.

[0024] As a further solution of the present invention, the abnormal area marking module includes:

[0025] The turning node extraction submodule extracts the time mark of each turning node based on the direction turning node time points marked in the crop growth trend offset information, records them in chronological order, and after determining the turning behavior, regards it as a valid turn and relocates the time to establish a turning node time series;

[0026] The health degradation detection submodule identifies the continuous data of the crop health status in three adjacent cycles based on the time point corresponding to each node in the turning node time series, performs sliding analysis, extracts the health status value within the time point, compares whether there is degradation, and if so, determines it to be in a health degradation state and generates a health degradation trend sequence;

[0027] The fluctuation region screening submodule extracts the time points that exist in both sequences based on the health degradation trend sequence. Combining the index positions corresponding to the intersection of the two sequences, the module identifies the key periodic regions that meet the synchronization conditions of turning and health degradation, uniformly encodes them into a set, and generates a set of crop growth fluctuation regions.

[0028] As a further solution of the present invention, the system further includes:

[0029] The status information output module compares the crop growth fluctuation area set with the morphological key node list in time to determine whether there is overlap between adjacent cycle time windows. If so, the corresponding area is listed as a risk focus point and the crop growth status information is output;

[0030] The crop growth status information includes risk focus points, cycle time window overlap status, and abnormal trend identification.

[0031] As a further solution of the present invention, the status information output module includes:

[0032] The window overlap identification submodule extracts abnormal areas and the start and end times of the operation cycle based on the crop growth fluctuation area set and the periodic time nodes in the morphological key node list, identifies the upper and lower bounds of the standard time window in combination with the standard cycle duration, matches the area start and end time periods, analyzes the periodic overlap relationship, and generates time window overlap range data;

[0033] The overlapping interval determination submodule extracts the overlapping start and end time points based on the overlapping range data of the time window, calculates the overlapping relationship index based on the regional activity frequency, duration and the difference between the period starting points, and compares it with the threshold. If it exceeds the threshold, it is marked as a valid overlapping period, and a valid overlapping period sequence is obtained;

[0034] The status information generation submodule extracts the region index according to the valid overlapping period sequence, checks whether it is a member of the abnormal region, and if so, and is located within the period overlapping interval, marks it as a risk area, classifies it into the risk focus point set, and outputs the crop growth status information.

[0035] The crop growth detection method based on machine vision is performed based on the above-mentioned crop growth detection system based on machine vision, and includes the following steps:

[0036] S1: Using multi-angle cameras set up in the field to capture crop canopy images, extract leaf edge contour changes, stem curvature, and plant spacing distribution, determine the consistency of these changes with the overall crop growth status, screen morphological features, and generate a crop morphological feature sequence;

[0037] S2: Based on the time nodes of the crop morphological feature sequence and the crop growth state change curve, determine whether the nodes overlap, mark the overlapping features as key features and record their positions, and generate a crop morphological key node list;

[0038] S3: Acquire continuous cycle crop image data based on the crop morphology key node list, evaluate the change range of leaf expansion area and stem elongation length, determine whether the turning range exceeds the original reference, and obtain crop growth trend deviation information;

[0039] S4: judging whether the crop health status has deteriorated within the corresponding period based on the turning time points in the crop growth trend offset information, marking the time points that meet the conditions, and generating a crop growth fluctuation area set;

[0040] S5: Combining the crop growth fluctuation area set with the morphological key node list, determine whether there is a period window overlapping area, mark it as a risk focus point, and output crop growth status information.

[0041] Compared with the prior art, the advantages and positive effects of the present invention are:

[0042] In the present invention, by setting up multi-angle cameras to capture crop canopy images and extracting spatial features such as leaf contours, stem curvature, and plant spacing, multi-dimensional modeling of crop morphology is achieved, which is no longer limited to single images or single-point data collection, and the depth and breadth of recognition of crop structural changes are improved. The morphological feature sequence is compared with time nodes to detect the synchronization of growth changes and key moments, and a temporal association between morphological features and growth state evolution is established. Key morphologies are obtained in dynamic evolution, which improves the timeliness and accuracy of morphological recognition. By periodically analyzing the changing trends of leaf expansion and stem elongation in continuous time periods, the inflection point identification of crop growth potential and the quantitative assessment of trend deviation are achieved, the pre-warning capability of abnormal growth patterns is strengthened, and the degradation of health status is identified based on trend deviation information. The crop fluctuation area can be marked for time periods, breaking the static interpretation mode of the result status, introducing periodic fluctuation identification, and making health monitoring more dynamically sensitive. The time-overlapping areas of the crop status are then focused on as high-risk nodes, achieving spatiotemporal locking of local problems in the farmland and promoting the precise allocation of management resources. The processing flow, from image feature analysis to time series trend assessment and then to spatial anomaly focusing, constitutes a bottom-up information reasoning chain, which enhances the ability to respond to dynamic changes in complex growth processes and comprehensively improves monitoring accuracy, risk identification timeliness and management intervention efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a system flow chart of the present invention;

[0044] Figure 2 This is a flow chart of the image acquisition module in the present invention;

[0045] Figure 3 This is a flow chart of the morphological feature capture module in the present invention;

[0046] Figure 4 This is a flow chart of the growth trend determination module in the present invention;

[0047] Figure 5 This is a flow chart of the abnormal area marking module in the present invention;

[0048] Figure 6 This is a flow chart of the status information output module in the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0050] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0051] See also Figure 1 The present invention provides a technical solution: a crop growth detection system based on machine vision includes:

[0052] The image acquisition module uses multi-angle cameras set up in the field to capture crop canopy images, extract leaf edge contour changes, stem curvature, and plant spacing distribution, and then generates a crop morphological feature sequence by matching the spatial distribution characteristics of these three parameters with the overall growth status of the crop.

[0053] The morphological feature capture module reads the crop morphological feature sequence and detects whether the corresponding node position in the time axis is consistent with the critical point of crop growth state change. If the time coincides, the corresponding morphological feature is marked as a key feature and a list of crop morphological key nodes is generated;

[0054] The growth trend determination module uses a list of key crop morphological nodes to compile crop image data from two consecutive cycles. It compares the leaf expansion area and stem elongation length in each cycle, assessing the magnitude of their change between cycles. If the direction of change reverses, it compares the data with the original growth baseline to obtain crop growth trend deviation information.

[0055] The abnormal area marking module detects whether the crop health status has deteriorated within the corresponding time period based on the crop growth trend deviation information, marks the abnormal area, and generates a set of crop growth fluctuation areas;

[0056] The status information output module compares the crop growth fluctuation area set with the morphological key node list in time to determine whether there is overlap between adjacent cycle time windows. If so, the corresponding area is listed as a risk concern point and the crop growth status information is output.

[0057] The crop morphological feature sequence includes morphological distribution characteristics, consistency matching frequency, and growth state stability. The crop morphological key node list includes key feature names, critical points of growth state changes, and time node mapping relationships. The crop growth trend offset information includes change amplitude, turning direction, and benchmark deviation status. The crop growth fluctuation area set includes turning nodes, health degradation mark areas, and key fluctuation areas within the cycle. The crop growth state information includes risk focus points, cycle time window overlap status, and abnormal trend identification.

[0058] See also Figure 2 , the image acquisition module includes:

[0059] The viewing angle adjustment submodule uses a multi-angle camera set up in the field to obtain crop canopy images, adjusts the camera pitch angle and focal length, extracts leaf edge contours, stem curvature, and plant spacing distribution, and generates a crop spatial distribution image sequence;

[0060] By precisely adjusting the camera's pitch angle and focal length, detailed crop images can be captured from various perspectives. To capture crop canopy images at different growth levels, the camera's pitch angle is adjusted between -15° and +15°, with images captured every 5° to ensure coverage of both upper and lower layers. Adjusting the focal length determines image clarity and capture range. The focal length can be set between 500mm and 1500mm to capture a wide range of crop images from a distance, while also allowing for closer focus on details. After image acquisition, image processing techniques are used to analyze the images. First, edge detection algorithms (such as the Canny operator) are used to extract leaves from the image and accurately identify their edge contours. Image recognition and curve fitting techniques are then applied to analyze the curvature of the stems, identifying the curve of each stem and calculating the stem's bend angle and radius using a fitting algorithm. For plant spacing analysis, image segmentation techniques are used to identify individual plant locations and calculate the distances between adjacent plants, creating a plant spacing distribution map. After all image processing data is combined, a series of crop spatial distribution image sequences are formed to facilitate subsequent analysis of crop growth status and morphological characteristics. The implementation of this series of steps can be verified by setting experiments with different shooting conditions to simulate the shooting effects of the camera under different weather and lighting conditions, ensuring that accurate data can be stably obtained in actual applications.

[0061] The feature extraction submodule analyzes the changing trends of leaf edge contours, stem curvature, and plant spacing distribution over a continuous period of time based on crop spatial distribution image sequences. It compares the changing directions of the three with the overall growth status of the crop, calculates the frequency of consistent occurrences, and obtains a frequency sequence of consistent morphology and growth status.

[0062] Image difference calculation is used to analyze changes in leaf edges. By comparing images over consecutive time periods, the difference between each image and the previous one is calculated. The difference calculation uses a normalization method to calibrate the edge contours of each image, resulting in a difference value. If this value is greater than a set threshold, it is considered that a significant change has occurred in the leaf. For stem curvature analysis, curve fitting is first used to extract the morphological changes of the stem within each frame. A common method is to calculate the stem curvature angle by performing a least squares fit on the edge points in the image. As crops grow, the degree of stem curvature increases or decreases. Continuously monitoring this trend helps assess crop health. For the distribution of interplant spacing, density function estimation techniques are used to calculate the distance between plants at different time points. By analyzing the changing trends in this distance, we can understand the spatial density of crop growth. By combining the changes in leaf edges, stem curvature, and interplant spacing distribution, we can further compare the overall crop growth status to determine whether there is any consistency between morphology and growth status. This analysis process can help accurately grasp the growth dynamics of the crop, and the frequency of consistent occurrences can be calculated based on the analysis. For example, during the analysis period, if the changes in the leaf edges are consistent with the direction of the stem bending changes, and the changes in plant spacing tend to be moderate, it means that the crop is in a good growth state. Through this process, a frequency sequence consistent with the morphology and growth state can be obtained, which further provides a basis for crop management and regulation.

[0063] The sequence screening submodule selects the morphological parameters with the best frequency according to the sequence of consistent morphology and growth status, identifies the corresponding original parameters, and establishes the crop morphological feature sequence;

[0064] First, a statistical analysis is performed on the frequency sequence of morphological and growth state consistency. The frequency of occurrence of each morphological parameter is counted, and the morphological parameters with the highest frequency are identified. For example, if a morphological parameter (such as stem bend angle) frequently coincides with the crop's growth state over a period of time, this morphological parameter can be considered relatively stable and representative within that time period. To more accurately select the optimal morphological parameter, a statistical analysis method for the frequency sequence is used to calculate the mean and standard deviation of the frequency sequence. The parameter with a frequency exceeding the mean plus one standard deviation is typically selected as the optimal morphological parameter. For example, if the frequency sequence of morphological parameter A is [4, 7, 10, 15, 12, 8], with a mean of 9 and a standard deviation of 3, then the parameter with a frequency greater than 15 is considered the optimal morphological parameter. By screening the morphological parameters with the highest frequency, it is possible to effectively identify the original parameters that match the crop's growth state, such as camera angle, focal length, and capture time. These original parameters can provide key information for subsequent crop monitoring. By associating the original crop parameters with the growth status of the crops, a crop morphological feature sequence is constructed, and crops are further dynamically monitored at different growth stages, providing accurate data support for agricultural production management. By dynamically updating the morphological feature sequence, farmers can understand the growth status of crops in real time and take appropriate management measures.

[0065] See also Figure 3 , the morphological feature capture module includes:

[0066] The node positioning submodule obtains the time points corresponding to the morphological parameters in the crop morphological feature sequence, extracts the parameter extreme value moments as representative nodes, sets a periodic time axis, uses three adjacent periods as sliding windows, identifies the local extreme value points in each sequence, marks the key nodes, and obtains the set of morphological key time nodes;

[0067] Node localization requires time-series processing of crop morphological characteristics. Changes in each morphological parameter are correlated with time, generating a series of time-series points that indicate the gradual evolution of crop morphological characteristics along the time axis. For example, if the crop is in a rapid growth phase, the leaf edge changes more dramatically, resulting in shorter intervals between time-series points. However, during a slow growth phase, the intervals between time-series points are longer. After analyzing these time-series points, the extreme moments are selected, representing key moments in crop morphological change. For example, in the early stages of crop growth, the edges of the leaves will expand rapidly, but as time goes by, the expansion amplitude gradually decreases. At this time, the extreme moment of the leaf edge change is a key node. By setting a periodic time axis, for example, each cycle is 10 days, and the three adjacent cycles are used as sliding windows, local extreme points can be identified from the time points in each sliding window. Local extreme points refer to the moments when the crop morphology changes significantly within the time period. In this way, the key nodes in each cycle can be marked. By continuously sliding the window and extracting local extreme points throughout the entire crop growth process, a complete set of morphological key time nodes can be obtained. The node set can provide an important reference for subsequent analysis.

[0068] The overlap detection submodule extracts the growth state value corresponding to the time node based on the set of morphological key time nodes, determines that the sign of the change before and after the moment is reversed and exceeds the fluctuation threshold as the critical point, records the critical point time, analyzes the correlation between the morphological features and the area where the critical point time overlaps, identifies the morphological features with a correlation higher than the threshold, and obtains the matching morphological feature set;

[0069] The core of coincidence detection lies in analyzing the relationship between changes in morphological characteristics and fluctuations in growth status. For example, suppose the leaf edge at a key node undergoes a transition from expansion to contraction over a short period of time. If the corresponding growth status value also exhibits the same trend, it indicates that the crop is at a growth turning point. When this trend reverses and the magnitude of the change exceeds a preset fluctuation threshold, the node is identified as a critical point. The fluctuation threshold can be set based on statistical analysis of historical data; for example, a fluctuation exceeding ±5% is considered significant. This process first extracts the growth status value at each key node. The value is represented by changes in parameters such as plant height, leaf number, and stem thickness. By comparing the growth status at the current moment with the previous and subsequent moments, it is determined whether the sign of the change has reversed. If the sign of the change exceeds the threshold, the moment is marked as a critical point. The time of the critical point is recorded along with the corresponding morphological characteristics, allowing further analysis of morphological characteristic changes during the critical point period. This method can identify areas where morphological characteristics overlap with the critical point time, thereby assessing the importance of morphological characteristics in crop growth.

[0070] The feature registration submodule marks the features and morphological parameters of the overlapping time points according to the matching morphological feature set, using the formula:

[0071] ;

[0072] Calculate key feature values ​​and record corresponding time nodes, integrate the key feature index, morphological name, time point number and growth state fluctuation direction value, and establish a list of key nodes of crop morphology;

[0073] in, represents the key eigenvalue, represents the time node number of the i-th time point, Representative The time node number of the time point, represents the morphological feature value at the i-th time point, Represents the growth state fluctuation direction value at the i-th time point, represents the total number of time points;

[0074] Based on the previous overlap detection results, parameters whose morphological features change significantly near the critical point time are selected as matching morphological features. Morphological features include the expansion amplitude of leaves, the bending angle of stems, and the spacing between plants. Changes in features directly reflect the fluctuations in crop growth status. When a morphological feature is confirmed to coincide with a change in growth status at a specific time point, the feature can be marked as a key feature and the corresponding time node recorded. For example, assuming that the change in the leaf edge completely matches the fluctuation in crop growth status at a certain critical point, the amplitude of the leaf edge change and its corresponding time point can be recorded as the key feature of the crop. The relevant information of the key feature is integrated, including the morphological parameter name, time point number, and its corresponding growth status fluctuation direction value, to finally form a list of crop morphological key nodes. This list contains detailed information on all key nodes, facilitating subsequent data analysis and crop growth prediction. In this way, the necessary reference data can be provided for precise crop management and help optimize the planting process.

[0075] Using the formula:

[0076] ;

[0077] in, represents the key eigenvalue, represents the time node number of the i-th time point, Representative The time node number of the time point, represents the morphological feature value at the i-th time point, Represents the growth state fluctuation direction value at the i-th time point, represents the total number of time points;

[0078] The key characteristic value is a comprehensive indicator used to represent the combined effect of crop morphological characteristics and growth state fluctuation direction values ​​during the growth process. By calculating the time difference between different time nodes and the square root of the sum of the squares of the crop morphological characteristics and growth state fluctuation direction values, it reflects the changing trend of the crop during the growth cycle. Specifically, the key characteristic value can comprehensively consider the crop's morphological changes (such as leaf growth and stem development) and growth state fluctuations (such as accelerated or slowed growth rate) at various time points, thus providing a basis for analyzing key growth nodes of crops, predicting growth trends, and formulating reasonable agricultural management measures.

[0079] (Time node number):

[0080] Acquisition method: Time node numbers are obtained from monitoring data during the crop growth cycle. This is usually done using ground sensors (such as temperature, humidity, and soil moisture sensors) or remote sensing technology (such as drones and satellite imagery) to monitor the different growth stages of the crop. Each monitoring point in the crop growth stage is calibrated as a time node number (for example, the first week after sowing, the second week after sowing, etc.).

[0081] Dimensional uniformity: Time node numbers are dimensionless, meaning they do not involve units and directly represent the time sequence using integers. Therefore, the definition and acquisition method of time nodes in the data directly determine the format of the time node numbers;

[0082] (Morphological characteristic values):

[0083] Acquisition method: Morphological characteristics reflect crop growth forms, such as leaf length, stem diameter, height, and leaf area. They are typically extracted through image processing techniques (such as image recognition based on RGB or near-infrared images) or ground sensors (such as LiDAR). Deep learning algorithms are used to analyze image or LiDAR data to measure crop morphological characteristics and quantify them into numerical values.

[0084] Dimensional unification: Various morphological characteristics have different units, such as leaf length (unit: cm), stem diameter (unit: mm), etc. Therefore, the data must be normalized, usually by dividing each feature value by the maximum value of the feature during the entire monitoring period, so that all morphological feature values ​​are normalized to between 0 and 1;

[0085] (Growth state fluctuation direction value):

[0086] Acquisition method: The growth state fluctuation direction value reflects the changes in crop growth trends. It can be obtained through long-term monitoring of crop growth data. Using meteorological data, soil data, light data, etc., combined with crop growth models or machine learning algorithms, the growth state fluctuation direction of crops is predicted (for example, growth acceleration or deceleration, growth state fluctuation trend);

[0087] Dimensional unification: The unit of the growth state fluctuation direction value is a dimensionless value, so it needs to be normalized to a value between 0 and 1. A common practice is to normalize the data to be within the range of 0 to 1, so as to avoid the impact of different fluctuation direction data units;

[0088] Dimensional unification process (normalization):

[0089] Normalization of morphological characteristics: Morphological characteristic values ​​(such as leaf length, stem diameter, etc.) need to be normalized. Assume that for a specific morphological characteristic (such as leaf length), its maximum value is , then the normalized morphological feature value It can be expressed as: ,in is the original morphological eigenvalue, is the maximum value of the morphological feature. After normalization, all morphological feature values ​​will be adjusted to values ​​between 0 and 1;

[0090] Normalization of growth state fluctuation direction value: The growth state fluctuation direction value also needs to be standardized. Assuming that the minimum value of the growth state fluctuation direction value is and the maximum value is , then the normalized growth state fluctuation direction value

[0091] It can be expressed as: ,in is the original growth state fluctuation direction value, and are the minimum and maximum values ​​of the value respectively. Through standardization, the data range is adjusted to between 0 and 1, eliminating the influence of dimension;

[0092] Formula calculation derivation process (taking specific numbers as an example):

[0093] Suppose there are data at three time points:

[0094] Time node number ( ): , , ;

[0095] Morphological characteristic values ​​( ): , , ;

[0096] Growth state fluctuation direction value ( ): , , ;

[0097] Calculate the time difference ( ):

[0098] ;

[0099] ;

[0100] Calculate the sum of squares of the fluctuation direction values ​​of morphological characteristics and growth status ( ):

[0101] ;

[0102] ;

[0103] ;

[0104] Compute the square root of a sum of squares: , , ;

[0105] Calculate the weighted value for each time point ( ):

[0106] for : ;

[0107] for : ;

[0108] The summation gives the key eigenvalues ​​( ): ;

[0109] The calculated key eigenvalue K represents a comprehensive picture of the crop's morphological characteristics and growth status fluctuation direction during its growth process. The results can reveal the changing patterns of crops in different growth stages and the fluctuation trends of their growth status, helping to assess the overall health of the crops and the stability of their growth process. Continuous monitoring of key eigenvalues ​​can provide important references for crop management, such as optimizing crop growth by adjusting agricultural management measures to ensure that crops reach their potential in the best growth state. At the same time, the changing trends of key eigenvalues ​​can also provide a scientific basis for crop production decisions, guiding the timing and strategies for implementing agricultural activities such as fertilization and irrigation.

[0110] See also Figure 4 , the growth trend determination module includes:

[0111] The area comparison submodule obtains crop image data from two consecutive cycles using a list of key crop morphological nodes. It then categorizes each cycle by leaf expansion area and stem elongation length. It compares the measured values ​​at each time node, identifies the changes in leaf area and stem elongation, and arranges them by cycle to generate a cyclical growth change sequence.

[0112] By obtaining a list of key morphological nodes of the crop, the key growth points of the crop in each time period can be determined, such as the leaf expansion area and stem elongation length. During the image data processing process, the image data of each cycle will be classified according to the changes in leaf area and stem elongation. For example, within a growth cycle, the leaf expansion area will increase over time, and the stem elongation length will also vary with the seasons. Therefore, by measuring the changes in these two parameters within each cycle, the growth status of the crop can be clearly judged. For example, in the first cycle, the leaf expansion area is 50cm² and the stem length is 15cm, while in the next cycle, the leaf expansion area is 70cm² and the stem elongates to 20cm. By comparing the changes in these two values, the growth progress of the crop can be identified and a cyclical growth change sequence can be generated. The data can serve as the basis for subsequent crop growth trend analysis.

[0113] The direction assessment submodule evaluates the direction of change between the current and previous cycle data based on the periodic growth change sequence, identifies and records the turning phenomenon, and analyzes the corresponding deviation amplitude of leaves and stems using the formula:

[0114] ;

[0115] Get the growth trend deviation;

[0116] in, Indicates the growth trend deviation, represents the blade length of the jth cycle, Representative The blade length of the cycle, represents the stem diameter in the jth period, Representative The stem diameter of the cycle, represents the total number of cycles;

[0117] By comparing growth data from the current cycle with the previous cycle, the direction of change can be identified. For example, if leaf area increases from 50 cm² to 70 cm² and stem length increases from 15 cm to 20 cm in a given cycle, if these two indicators increase in the same direction, the crop's growth remains stable. However, if leaf area increases while stem length decreases, this indicates a shift in growth direction. To assess the degree of shift, the magnitude of the change between the two cycles is calculated. Then, a determination is made as to whether the direction of change has reversed. If so, this is recorded and labeled as a reversal. The magnitude of the shift is calculated by comparing the change in leaf and stem area. For example, if the change in leaf area is 20 cm², the change in stem length is -5 cm. By calculating the relative difference between these two changes, the degree of growth trend shift is determined. If the shift exceeds a certain threshold, a significant directional shift in crop growth is considered. For example, if the threshold is set at 10%, a shift greater than 10% indicates a shift in growth direction.

[0118] Growth trend deviation is a comprehensive indicator used to measure the differences in changes between leaves and stems of crops during different growth cycles. Specifically, it reflects the relative deviation between leaf length and stem diameter. When the deviation is large, it means that the changes in leaves and stems are more significant, indicating that the growth of crops within a certain cycle fluctuates greatly, or that the growth state is significantly affected by external factors (such as climate change and soil conditions). Conversely, when the deviation is small, it indicates that the growth changes of leaves and stems are relatively consistent and the crop growth is relatively stable. By analyzing the deviation, we can better understand the growth trends of crops and provide an important basis for agricultural production decision-making.

[0119] (Leaf length in the jth cycle):

[0120] Acquisition method: Leaf length data is collected through field observations, remote sensing images, or ground sensors. Remote sensing technologies (such as drones or satellite images) can measure leaf length through image recognition technology. In addition, laser scanning or image processing technology can also accurately determine leaf length.

[0121] Quantification: To ensure that data can be compared in the same dimension, leaf length is recorded in centimeters (cm). If leaf length data comes from image recognition or sensor measurement, it needs to be normalized based on the longest leaf. For example, by measuring the maximum leaf length of the crop, all leaf lengths are divided by the maximum leaf length to normalize the data so that it remains in the range of 0 to 1.

[0122] (Blade length in the last cycle):

[0123] Acquisition method: Leaf length data from the previous cycle is collected using the same technical means and recorded at the end of each monitoring cycle. The cycle definition is based on the growth stage of the crop, for example, data is recorded every 7 or 10 days.

[0124] Quantification: To ensure data comparability, the blade length of the previous cycle is also normalized to ensure that its unit is the same as the blade length of the current cycle;

[0125] (stem diameter in period j):

[0126] Acquisition method: The stem diameter data is obtained using specialized measuring tools (e.g., calipers, digital calipers) or remote sensing technology. In the field, high-precision calipers are used for physical measurement. In remote sensing images, the maximum stem diameter can be extracted and calculated using image processing algorithms.

[0127] Quantification: Stem diameters are recorded in millimeters (mm). After being captured through image recognition or sensors, they are normalized to a value between 0 and 1. For example, the maximum value method is used to divide each stem diameter by the maximum stem diameter of the crop to normalize the data.

[0128] (Stem diameter in the previous cycle):

[0129] Acquisition method: The stem diameter data of the previous cycle are obtained using the same acquisition method, which is consistent with the periodic monitoring time of leaf length. The definition of the cycle affects the data recording and acquisition timing of each measurement;

[0130] Quantification: The normalization process is the same as the stem diameter of the current cycle to ensure unit consistency and data comparability;

[0131] Dimensional unification process (normalization):

[0132] Blade length ( and ):

[0133] The unit of leaf length is centimeters (cm). Due to the different growth stages of crops, direct comparison of the original data of leaf length data in different periods will be affected by the difference in units and scales. Therefore, in actual application, the leaf length will be normalized. Assuming that the maximum leaf length in the period is , then the normalized leaf length for: , the normalized leaf length data are in the range of 0 to 1, eliminating the differences between units;

[0134] Stem diameter ( and ):

[0135] The unit of stem diameter is millimeter (mm). Since the stem diameter varies at different time points and in different crop varieties, it is necessary to perform normalization. The same maximum value normalization method is used. Assuming that the maximum stem diameter in the period is , then the normalized stem diameter for: ,In this way, all stem diameter data will be normalized to the range of 0 to 1, eliminating the influence of size differences;

[0136] Formula calculation derivation process (taking specific numbers as an example):

[0137] Assume that the raw data during the monitoring period are as follows:

[0138] Data for cycle 1: blade length cm, stem diameter mm;

[0139] Data for cycle 2: blade length cm, stem diameter mm;

[0140] Data for cycle 3: blade length cm, stem diameter mm;

[0141] By normalization:

[0142] Normalized blade length: Assuming the maximum blade length cm, the normalized leaf length is:

[0143] , , ;

[0144] Normalized stem diameter: Assuming the maximum stem diameter mm, the normalized stem diameter is:

[0145] , , ;

[0146] Calculation process:

[0147] Calculate the blade length difference:

[0148] For period 2, ;

[0149] For cycle 3, ;

[0150] Calculate the stem diameter difference:

[0151] For period 2, ;

[0152] For cycle 3, ;

[0153] Calculate the leaf-to-stem deviation ratio:

[0154] For period 2, the deviation ratio is: ;

[0155] For period 3, the deviation ratio is: ;

[0156] Summing this gives the offset:

[0157] Total deviation for: ;

[0158] : The calculated offset is 1.75, which means that the difference in changes between leaves and stems of crops in these cycles is relatively large, and as the cycle progresses, the difference in changes between leaves and stems is decreasing.

[0159] The benchmark comparison submodule extracts the growth change data of the original continuous cycle of the crop based on the growth trend deviation, identifies the growth change benchmark range, and compares the deviation of the current cycle with the growth change benchmark range. If it exceeds the benchmark range, the corresponding cycle is marked as an abnormal growth fluctuation segment, thereby obtaining the crop growth trend deviation information;

[0160] Based on historical data, a baseline range for normal crop growth is determined. For example, if leaf area and stem length fluctuate regularly over certain cycles, the baseline range can be set based on this regularity, using the mean ± one standard deviation as the baseline range. For example, if the standard deviation of leaf area for a particular crop over the past few cycles is 5 cm², the baseline range would be ±5 cm². If the leaf area change in the current cycle is 10 cm², exceeding the baseline range threshold, the cycle is marked as an abnormal fluctuation segment. By comparing the deviation of the current cycle with the baseline range, abnormal fluctuations in crop growth can be detected promptly. This method allows for accurate detection of abnormal fluctuations in crop growth and timely response measures. For example, if stem length is significantly below the baseline range in a given cycle, this indicates that the crop is experiencing environmental stress or disease during that cycle, and appropriate management measures are needed.

[0161] See also Figure 5, the abnormal area marking module includes:

[0162] The turning node extraction submodule extracts the time mark of each turning node based on the marked turning node time points in the crop growth trend offset information, records them in chronological order, and after determining the turning behavior, regards it as a valid turn and relocates the time to establish a turning node time series;

[0163] Growth trend offset data is used to identify turning points in crop growth. Turning points refer to time points during crop growth when the direction of crop growth changes significantly, such as when the direction of leaf growth changes from expansion to contraction, or when the direction of stem elongation changes. By extracting the time stamp of each turning point and sorting them in chronological order, the time when the turning behavior occurred can be determined, and whether the turning point is a valid turning point can be further verified. For example, when the growth of the stem suddenly slows down and the expansion of the leaves no longer increases, it is considered that the growth direction of the crop has turned, and this time point is recorded as a turning point. To further improve accuracy, the time of the turning point can be repositioned according to the time period before and after the turning point occurs, ensuring that the turning point accurately reflects the actual state of crop growth. Ultimately, a turning point time series can be generated, which records in detail the moment when the crop growth direction changes, providing an important reference for subsequent health status assessment and growth trend analysis.

[0164] The health degradation detection submodule identifies the continuous data of crop health status in three adjacent cycles based on the time point corresponding to each node in the turning node time series, performs sliding analysis, extracts the health status value within the time point, and compares whether there is degradation. If so, it is determined to be in a health degradation state and generates a health degradation trend sequence;

[0165] Based on the time point corresponding to each turning point, crop health data for three consecutive cycles is extracted. Health values ​​are measured using growth parameters such as leaf condition, stem thickness, and plant height. Sliding analysis of health data across these three cycles reveals trends in crop health. The analysis first compares changes in health over consecutive time periods. If a significant decline in health is observed within a given cycle, and the magnitude of the decline exceeds a set threshold (for example, a decrease of more than 10%), the crop is considered to have experienced health degradation. Degradation is determined based on a set threshold. For example, if a crop's health value decreases by more than 10% compared to its baseline level over three consecutive cycles, it is considered to be in a state of health degradation. A health degradation trend sequence is generated, documenting fluctuations in crop health within each cycle and identifying periods of health degradation, providing a basis for subsequent intervention measures.

[0166] The fluctuation region screening submodule extracts time points that exist in both sequences based on the health degradation trend sequence. Combining the index positions corresponding to the intersection of the two sequences, the module identifies key periodic regions that meet the synchronization conditions of turning and health degradation, uniformly encodes them into a set, and generates a set of crop growth fluctuation regions.

[0167] By comparing the health degradation trend series with the turning node time series, we can identify the time points that appear in both series. These time points are the key moments when crop turning occurs and health degradation is synchronized. For example, in a certain cycle, when the crop turns to a node, there is a significant decline in the health status, indicating that this cycle is an important time point for crop growth fluctuations. By identifying the synchronization phenomenon, we can further analyze the abnormal fluctuation areas in the crop growth process. Combined with the intersection index position of the turning node and health degradation, we can accurately locate the key cycle areas that show synchronous changes, and then generate a collection of crop growth fluctuation areas. This collection will help monitor and identify abnormal fluctuation sections in the crop growth process, such as the impact of disease or environmental changes on crop health, and provide data support for precision agriculture management.

[0168] Table 1: Crop health status change data

[0169]

[0170] Table 1 shows the change in crop health over three consecutive cycles. The "Health Change" column shows the change in health value for each cycle. For example, at T1, the health value dropped from 85 to 80, a decrease of 5 units, indicating that the crop health deteriorated during this cycle.

[0171] See also Figure 6 , the status information output module includes:

[0172] The window overlap identification submodule extracts abnormal areas and the start and end times of the operation cycle based on the crop growth fluctuation area set and the periodic time nodes in the morphological key node list. It then identifies the upper and lower bounds of the standard time window based on the standard cycle duration, matches the regional start and end time periods, analyzes the periodic overlap relationship, and generates time window overlap range data.

[0173] Classify the cycle time nodes, identify the start and end time periods of the cycle, and combine the time nodes with the growth fluctuation data of the region to locate the abnormal area. The abnormal area refers to the area where the fluctuation amplitude of the crop growth process exceeds the normal range. It can be defined by calculating the standard deviation of the growth curve. For example, when the fluctuation exceeds ±2 times the standard deviation, it can be regarded as an abnormal fluctuation. On this basis, the start and end time periods of the abnormal area are extracted. For example, if the fluctuation of a certain area increases significantly between the 5th and 8th days, the time period can be regarded as an abnormal period. The time period is further verified by the standard cycle length to ensure that it is consistent with the duration of the standard cycle. The standard cycle length needs to be set according to the normal growth cycle of the crop. For example, for a certain crop, its normal growth cycle is 30 days, so the length of the standard cycle is 30 days. If the abnormal cycle deviates significantly from the length, it needs to be marked. Combined with this information, we further identify the upper and lower bounds of the time window that conforms to the standard period, and analyze the overlap by matching the start and end time periods of the region to generate the time window overlap range data. This process requires the use of a time series data matching algorithm, such as dynamic time warping (DTW), which is used to measure the similarity of growth fluctuations in different time periods and generate accurate overlap range data.

[0174] The overlapping interval determination submodule extracts the overlapping start and end time points based on the overlapping range data of the time window, calculates the overlapping relationship index based on the regional activity frequency, duration and the difference between the period starting points, and compares it with the threshold. If it exceeds the threshold, it is marked as a valid overlapping period, and a valid overlapping period sequence is obtained;

[0175] The threshold is a judgment limit for the overlap index. Only when the overlap index exceeds the threshold does it indicate that the overlap between the two cycles is significant and will it be identified as an "effective overlap cycle";

[0176] The overlapping start and end time points are extracted and further calculated based on the regional activity frequency, duration, and cycle start point difference. The activity frequency can be calculated by counting the number of frequent fluctuations in crop growth within each time period. For example, if there are three significant fluctuations in a week, the activity frequency is 3 times / week. The duration refers to the duration of the fluctuation. For example, if a fluctuation lasts from the 5th to the 8th day, the duration is 3 days. The cycle start point difference refers to the difference between the start time of the overlapping period and the start time of the standard period. The larger the difference, the greater the deviation of the overlapping period from the standard period. The parameters can be used to calculate the overlap relationship index, which is a key indicator of the degree of overlap between two cycles. If the overlap index exceeds the set threshold, for example, the threshold is set to 0.8, the cycle is considered to be a valid overlap cycle. The threshold setting needs to be adjusted based on experience and actual conditions. A reasonable percentage can be set by statistically analyzing the normal overlap of crop growth cycles. For example, 0.8 means that 80% of the overlap periods meet the standard cycle requirements. In actual applications, if the overlap index is higher than the threshold, the cycle is marked as a valid overlap cycle and included in the valid overlap cycle sequence.

[0177] The status information generation submodule extracts the region index based on the valid overlapping cycle sequence and checks whether it is a member of the abnormal region. If so, and if it is located within the period overlapping interval, it is marked as a risk area and included in the risk focus set, and the crop growth status information is output;

[0178] Check whether the area is a member of the abnormal area. If it is an abnormal area, further check whether it is located in the period overlap interval. If the area is located in the overlap interval, the area is marked as a risk area and included in the risk focus set. The calibration process of the risk area depends on the valid overlapping period identified previously. If the overlapping time period of the valid period happens to be within the growth fluctuation interval of the abnormal area, and the overlapping relationship index is high, then this area is considered to be a high-risk area. For example, suppose an area is in the overlapping period range from the 10th to the 12th day, and the area has a large abnormal fluctuation during the growth cycle, then it will be marked as a risk area, the risk area will be included in the risk focus set, and the crop growth status information will be output. The information provides data support for subsequent crop monitoring and management.

[0179] The crop growth detection method based on machine vision is performed based on the above-mentioned crop growth detection system based on machine vision, and includes the following steps:

[0180] S1: Using multi-angle cameras set up in the field to capture crop canopy images, extract leaf edge contour changes, stem curvature, and plant spacing distribution, determine the consistency of these changes with the overall crop growth status, screen morphological features, and generate a crop morphological feature sequence;

[0181] S2: Based on the time nodes of the crop morphological feature sequence and the crop growth state change curve, determine whether the nodes overlap, mark the overlapping features as key features and record their positions, and generate a list of crop morphological key nodes;

[0182] S3: Based on the list of key nodes of crop morphology, continuous cycle crop image data is obtained to evaluate the change in leaf expansion area and stem elongation length, determine whether the turning amplitude exceeds the original benchmark, and obtain crop growth trend deviation information;

[0183] S4: Based on the turning time points in the crop growth trend offset information, determine whether the crop health status has deteriorated within the corresponding period, mark the time points that meet the conditions, and generate a crop growth fluctuation area set;

[0184] S5: Combine the crop growth fluctuation area set and the morphological key node list to determine whether there is a period window overlapping area, mark it as a risk focus point, and output the crop growth status information.

[0185] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A crop growth detection system based on machine vision, characterized in that: The system comprises: The image acquisition module uses multi-angle cameras set up in the field to capture crop canopy images, extract leaf edge contour changes, stem curvature, and plant spacing distribution, and then generates a crop morphological feature sequence by matching the spatial distribution characteristics of these three parameters with the overall growth status of the crop. The morphological feature capture module reads the crop morphological feature sequence and detects whether the corresponding node position in the time axis is consistent with the critical point of the crop growth state change. If the time coincides, the corresponding morphological feature is marked as a key feature and a crop morphological key node list is generated; The growth trend determination module uses the crop morphology key node list to collect crop image data for two consecutive cycles, compares the leaf expansion area and stem elongation length in each cycle, and evaluates the change amplitude of the two during the cycle. If the change direction changes, it is compared with the original growth benchmark to obtain crop growth trend deviation information; The abnormal area marking module detects whether the crop health status has degraded within a corresponding time period based on the crop growth trend deviation information, marks the abnormal area, and generates a crop growth fluctuation area set.

2. The crop growth detection system based on machine vision according to claim 1, characterized in that: The crop morphological feature sequence includes morphological distribution characteristics, consistency matching frequency, and growth state stability. The crop morphological key node list includes key feature names, critical points of growth state changes, and time node mapping relationships. The crop growth trend offset information includes change amplitude, turning direction, and baseline deviation status. The crop growth fluctuation area set includes turning nodes, health degradation marking areas, and key fluctuation areas within the cycle.

3. The crop growth detection system based on machine vision according to claim 1, characterized in that: The image acquisition module includes: The viewing angle adjustment submodule uses a multi-angle camera set up in the field to obtain crop canopy images, adjusts the camera pitch angle and focal length, extracts leaf edge contours, stem curvature, and plant spacing distribution, and generates a crop spatial distribution image sequence; The feature extraction submodule analyzes the changing trends of leaf edge contour changes, stem curvature, and plant spacing distribution over a continuous period of time based on the crop spatial distribution image sequence, compares the changing directions of the three with the overall growth status of the crop, calculates the frequency of consistent occurrences, and obtains a frequency sequence of consistent morphology and growth status; The sequence screening submodule selects the morphological parameters with the best occurrence frequency according to the frequency sequence of the consistent morphology and growth state, identifies the corresponding original parameters, and establishes the crop morphological feature sequence.

4. The machine vision-based crop growth detection system according to claim 3, characterized in that: The morphological feature capturing module includes: The node positioning submodule obtains the time points corresponding to the morphological parameters in the crop morphological feature sequence, extracts the parameter extreme value moments as representative nodes, sets a periodic time axis, uses three adjacent periods as sliding windows, identifies the local extreme value points in each sequence and marks the key nodes, and obtains a set of morphological key time nodes; The overlap detection submodule extracts the growth state value corresponding to the time node based on the set of key morphological time nodes, determines that the sign of the change before and after the moment is reversed and exceeds the fluctuation threshold as the critical point, records the critical point time, analyzes the correlation between the morphological features and the area where the critical point time overlaps, identifies the morphological features with a correlation higher than the threshold, and obtains the matching morphological feature set; The feature registration submodule marks the features and morphological parameters of the overlapping time points according to the matching morphological feature set, calculates the key feature values, records the corresponding time nodes, integrates the index, morphological name, time point number and growth status fluctuation direction value of the key features, and establishes a list of crop morphological key nodes.

5. The machine vision-based crop growth detection system according to claim 4, characterized in that: The growth trend determination module includes: The area comparison submodule obtains crop image data for two consecutive cycles through the crop morphology key node list, classifies each cycle according to leaf expansion area and stem elongation length, compares the measured values ​​of each category time node, identifies the change values ​​of leaf area and stem elongation, and arranges them by cycle to generate a periodic growth change sequence; The direction assessment submodule evaluates the direction of change between the current and previous cycle data based on the periodic growth change sequence, identifies and records the turning phenomenon, analyzes the corresponding deviation amplitude of leaves and stems, and obtains the growth trend deviation degree; The benchmark comparison submodule extracts the growth change data of the original continuous cycle of the crop based on the growth trend deviation, identifies the growth change benchmark range, compares the deviation of the current cycle with the growth change benchmark range, and if it exceeds the benchmark range, marks the corresponding cycle as an abnormal growth fluctuation section to obtain the crop growth trend deviation information.

6. The machine vision-based crop growth detection system according to claim 5, characterized in that: The abnormal area marking module includes: The turning node extraction submodule extracts the time mark of each turning node based on the direction turning node time points marked in the crop growth trend offset information, records them in chronological order, and after determining the turning behavior, regards it as a valid turn and relocates the time to establish a turning node time series; The health degradation detection submodule identifies the continuous data of the crop health status in three adjacent cycles based on the time point corresponding to each node in the turning node time series, performs sliding analysis, extracts the health status value within the time point, compares whether there is degradation, and if so, determines it to be in a health degradation state and generates a health degradation trend sequence; The fluctuation region screening submodule extracts the time points that exist in both sequences based on the health degradation trend sequence. Combining the index positions corresponding to the intersection of the two sequences, the module identifies the key periodic regions that meet the synchronization conditions of turning and health degradation, uniformly encodes them into a set, and generates a set of crop growth fluctuation regions.

7. The machine vision-based crop growth detection system according to claim 1, characterized in that: The system also includes: The status information output module compares the crop growth fluctuation area set with the morphological key node list in time to determine whether there is overlap between adjacent cycle time windows. If so, the corresponding area is listed as a risk focus point and the crop growth status information is output; The crop growth status information includes risk focus points, cycle time window overlap status, and abnormal trend identification.

8. The machine vision-based crop growth detection system according to claim 7, characterized in that: The status information output module includes: The window overlap identification submodule extracts abnormal areas and the start and end times of the operation cycle based on the crop growth fluctuation area set and the periodic time nodes in the morphological key node list, identifies the upper and lower bounds of the standard time window in combination with the standard cycle duration, matches the area start and end time periods, analyzes the periodic overlap relationship, and generates time window overlap range data; The overlapping interval determination submodule extracts the overlapping start and end time points based on the overlapping range data of the time window, calculates the overlapping relationship index based on the regional activity frequency, duration and the difference between the period starting points, and compares it with the threshold. If it exceeds the threshold, it is marked as a valid overlapping period, and a valid overlapping period sequence is obtained; The status information generation submodule extracts the region index according to the valid overlapping period sequence, checks whether it is a member of the abnormal region, and if so, and is located within the period overlapping interval, marks it as a risk area, classifies it into the risk focus point set, and outputs the crop growth status information.

9. A crop growth detection method based on machine vision, characterized in that: The method is used to implement the machine vision-based crop growth detection system according to any one of claims 1 to 8, comprising the following steps: S1: Using multi-angle cameras set up in the field to capture crop canopy images, extract leaf edge contour changes, stem curvature, and plant spacing distribution, determine the consistency of these changes with the overall crop growth status, screen morphological features, and generate a crop morphological feature sequence; S2: Based on the time nodes of the crop morphological feature sequence and the crop growth state change curve, determine whether the nodes overlap, mark the overlapping features as key features and record their positions, and generate a crop morphological key node list; S3: Acquire continuous cycle crop image data based on the crop morphology key node list, evaluate the change range of leaf expansion area and stem elongation length, determine whether the turning range exceeds the original reference, and obtain crop growth trend deviation information; S4: judging whether the crop health status has deteriorated within the corresponding period based on the turning time points in the crop growth trend offset information, marking the time points that meet the conditions, and generating a crop growth fluctuation area set; S5: Combining the crop growth fluctuation area set with the morphological key node list, determine whether there is a period window overlapping area, mark it as a risk focus point, and output crop growth status information.

Images