A rapid method and system for determining the activity of plant probiotics

Abstract

This application relates to the field of image data processing technology, and in particular to a rapid method and system for determining the activity of plant probiotics. The method includes: acquiring a microscopic video stream containing plant probiotics and organic matrix particles, and performing trajectory tracking on moving targets to obtain the motion trajectory of each moving target; for any motion trajectory, calculating an adaptive diffusion threshold for activity determination; and determining the activity of plant probiotics based on the adaptive diffusion threshold. This application improves the accuracy of probiotic activity detection.

Description

Technical Field

[0001] This application relates to the field of image data processing technology, and in particular to a rapid method and system for determining the activity of plant probiotics. Background Technology

[0002] With the rapid development of modern ecological agriculture, microbial fertilizers have become an important component of green agricultural inputs, and their core competitiveness lies in the bioactivity of the functional plant probiotics they contain. This activity directly determines the colonization ability and metabolic level of the bacteria after they are applied to the soil. Because traditional plate count methods are cumbersome and time-consuming, and cannot directly characterize the immediate vitality of the bacteria, machine vision-based microscopic detection technology is increasingly being applied to the rapid identification of bacterial activity. Among these technologies, trajectory classification algorithms based on mean square displacement (MSD) are currently the mainstream technique for distinguishing the motion patterns of microscopic particles.

[0003] However, existing MSD-based trajectory classification algorithms have shortcomings in practical detection scenarios for plant probiotics. These algorithms typically employ a global statistical strategy, determining whether a particle is undergoing random Brownian motion or directional active swimming by calculating the slope of the mean square displacement curve along the entire trajectory. This global averaging of motion patterns masks the unique dynamic characteristics of organisms. Typical plant probiotics (such as Bacillus subtilis) exhibit a unique motion mechanism that alternates between rapid linear swimming and disordered tumbling in place, and existing algorithms cannot accurately capture this instantaneous modal change. Furthermore, fermentation broth contains a large amount of suspended organic residue, and these background particles also undergo displacement in liquid environments (such as thermal convection or sampling vibration). Existing algorithms struggle to effectively distinguish between the organism's autonomous driving and the passive drift carried by the fluid, making activity determination in complex solid-liquid mixtures highly susceptible to interference from environmental noise. Summary of the Invention

[0004] To reduce the interference from the movement of organic residues during the application of relevant trajectory classification algorithms and to improve the accuracy and robustness of probiotic activity detection, this application provides a rapid method and system for identifying the activity of plant probiotics.

[0005] Firstly, this application provides a rapid method for determining the activity of plant probiotics, employing the following technical solution:

[0006] A rapid method for determining the activity of plant probiotics includes: acquiring a microscopic video stream containing plant probiotics and organic matrix particles, and performing trajectory tracking on moving targets to obtain the motion trajectory of each moving target;

[0007] For any motion trajectory, calculate the adaptive diffusion threshold for activity determination of the motion trajectory;

[0008] The activity of plant probiotics is determined based on an adaptive diffusion threshold.

[0009] The step of calculating the adaptive diffusion threshold for activity determination of the motion trajectory includes: for any motion trajectory, analyzing its short-term mean square displacement variation characteristics to obtain the instantaneous modal volatility of the motion trajectory, which is used to characterize the variation characteristics of the short-term mean square displacement along the motion trajectory.

[0010] The correlation between the motion vectors of the moving target and the local flow field during the mode switching period of the motion trajectory is analyzed to obtain the biological mode confidence score. The biological mode confidence score is used to characterize the possibility that the switching of motion modes in the motion trajectory is biologically driven.

[0011] The product of the biomodal confidence and instantaneous modal volatility of the motion trajectory is used as the biofeature value, and the biofeature value is used to adjust the preset initial threshold to obtain the adaptive diffusion threshold.

[0012] Instantaneous modal volatility derived from short-term mean square displacement analysis can characterize the unique rapid linear motion-random tumbling switching features of organisms, enabling the algorithm to identify the unique instantaneous dynamic changes of live bacteria. Meanwhile, the biomodal confidence score constructed through flow field consistency analysis can effectively distinguish between autonomously driven probiotics and passively drifting background particles, fundamentally reducing misjudgments caused by environmental disturbances. Finally, by combining instantaneous modal volatility and biomodal confidence scores to form biometric values ​​and adjusting the threshold, the classification criteria are automatically optimized for each trajectory, improving the accuracy, robustness, and applicability to solid-liquid mixed environments in activity determination.

[0013] Optionally, the steps to analyze the short-term mean square displacement variation characteristics and obtain the instantaneous modal volatility of the motion trajectory include: dividing the motion trajectory into sliding windows along the time axis; calculating the instantaneous diffusion index within each sliding window; and taking the product of the standard deviation and the sum of the absolute values ​​of the differences of the instantaneous diffusion index as the instantaneous modal volatility.

[0014] By dividing the trajectory into sliding windows and combining the sum of the standard deviation and absolute value of the instantaneous diffusion index, the severity of the deviation of the motion pattern from the average state and the switching frequency between adjacent time windows are reflected. This allows for a more accurate identification of the biological characteristics of active bacteria actively adjusting their motion posture due to life activities, thereby improving the robustness and accuracy of instantaneous modal fluctuation calculation.

[0015] Optionally, the step of calculating the instantaneous diffusion index within each sliding window includes: using the slope obtained by linearly fitting the mean square displacement of different time intervals under the window to the logarithmic relationship of the time interval as the instantaneous diffusion index.

[0016] Optionally, the step of analyzing the correlation between the motion vectors of the moving target and the local flow field during the mode switching period of the motion trajectory to obtain the biomodal confidence score includes: obtaining the set of key time windows in the motion trajectory where mode switching occurs; for any time window in the set of key time windows, obtaining the velocity vector of the moving target and the flow field vector within the local range of the moving target in that time window;

[0017] Obtain the cosine of the angle between the velocity vector of the moving target and the local flow field vector; the confidence level of the biological modality is positively correlated with the sum of the cosine values ​​of the angle at each moment in the key time window set.

[0018] Analyzing the cosine relationship between the velocity vectors of the moving target and the local flow field within a key time window can effectively identify the true driving force source of the motion pattern. If the motion direction is consistent with the background flow field, it may indicate passive drift; if the direction difference is large, it is more likely to be driven autonomously by active bacteria.

[0019] Optionally, the step of obtaining the set of key time windows in the motion trajectory where mode switching occurs includes: taking the average instantaneous diffusion index of each sliding window in the motion trajectory as the diffusion benchmark, taking the sliding window whose instantaneous diffusion index differs from the diffusion benchmark by more than a preset diffusion threshold as the key time window, and selecting multiple key time windows to form a set of key time windows.

[0020] By selecting key time windows based on the difference between the instantaneous diffusion index and the diffusion benchmark corresponding to the entire motion trajectory, mode switching detection has a clearer physical threshold basis.

[0021] Optionally, the step of obtaining the flow field vector within a local area of ​​the moving target includes: extracting the velocity vectors of other background particles within a preset radius around the moving target in the time window, and using the average value of the velocity vectors of the other background particles as the local flow field vector.

[0022] Optionally, the step of adjusting the preset initial threshold using biometric values ​​to obtain an adaptive diffusion threshold includes: normalizing the biometric values ​​using a hyperbolic tangent function to obtain a standard feature value; and using the difference between the preset initial threshold and the standard feature value as the adaptive diffusion threshold.

[0023] The hyperbolic tangent function is used to normalize the biometric values, making the mapping of biometric values ​​smooth and the range stable, which can reduce the impact of outliers or noise on subsequent threshold adjustments.

[0024] Optionally, the steps for tracking moving targets to obtain the motion trajectory of each moving target include: extracting moving targets from the video stream using a Gaussian mixture model; performing morphological opening operations on the moving targets to remove noise, and extracting the centroids of the connected components corresponding to the moving targets as the coordinates of the moving targets; and generating motion trajectories by associating the coordinates of moving targets in adjacent frames based on Kalman filter prediction and Hungarian algorithm matching.

[0025] Optionally, after acquiring the microscopic video stream containing plant probiotics and organic matrix particles, the acquired image sequence is preprocessed using a contrast-limited adaptive histogram equalization algorithm.

[0026] Preprocessing with a contrast-limited adaptive histogram equalization algorithm effectively enhances the contrast between the bacterial cells and the turbid fermentation broth background, overcoming the imaging blurring problem caused by the poor light transmittance of the fermentation broth, and providing a high-quality image foundation for subsequent target extraction and feature analysis.

[0027] Secondly, this application provides a rapid identification system for the activity of plant probiotics, which adopts the following technical solution:

[0028] A rapid identification system for the activity of plant probiotics includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a rapid identification method for the activity of plant probiotics as described above is implemented.

[0029] The above-mentioned method for rapid identification of the activity of plant probiotics is generated into a computer program and stored in a memory so that it can be loaded and executed by a processor. Thus, a system can be made based on the memory and processor for convenient use.

[0030] This application has the following technical advantages:

[0031] The short-term instantaneous modal fluctuation index of the trajectory is used to capture the "swimming-tumbling" switching characteristics of the bacterial cells, and the correlation with the motion vector of the local flow field is combined to eliminate interference from passively drifting organic residues. By dynamically adjusting the judgment threshold using these two biological characteristics, an adaptive diffusion threshold is constructed, which effectively solves the problem of misjudgment caused by microscopic turbulence and background noise in complex fermentation broth environments, and improves the accuracy and robustness of activity detection. Attached Figure Description

[0032] Figure 1 This is a flowchart of a method for rapidly determining the activity of plant probiotics according to an embodiment of this application.

[0033] Figure 2 This is a flowchart of step S2 in a rapid method for determining the activity of plant probiotics according to an embodiment of this application.

[0034] Figure 3 This is a diagram showing the effect of this application in judging the activity of plant probiotics based on an adaptive diffusion threshold. Detailed Implementation

[0035] This application discloses a rapid method for determining the activity of plant probiotics. It involves acquiring a video stream of the bacterial culture using microscopic vision technology and obtaining trajectory data using a multi-target tracking algorithm. Based on this, by analyzing the instantaneous diffusion mode fluctuation characteristics of the trajectory and its coupling relationship with the local flow field, a bio-driven mode switching confidence level is constructed. This allows for the dynamic generation of an adaptive diffusion threshold for each motion trajectory, achieving accurate determination of the activity of plant probiotics in complex fermentation broth environments.

[0036] Reference Figure 1 A rapid method for determining the activity of plant probiotics includes steps S1 to S3.

[0037] S1: Collect microscopic video streams containing plant probiotics and organic matrix particles, and perform trajectory tracking on moving targets to obtain the motion trajectory of each moving target.

[0038] First, a detection terminal is deployed on the bio-fertilizer fermentation production line, utilizing a high-resolution industrial microscope camera, for example equipped with... or oil immersion The camera performs real-time imaging of the bacterial culture slides after sampling from the liquid fermenter. The acquisition parameters can be set to frame rate. The duration of a single data collection session is seconds, resolution not lower than Pixels are used to ensure that the complete linear swimming-tumbling cycle of the bacteria can be recorded.

[0039] After acquiring the video stream, the image sequence is preprocessed. Specifically, contrast-limited adaptive histogram equalization can be used. The algorithm enhances the contrast between the bacterial cells and the background to overcome the problem of poor light transmittance in the fermentation broth. Subsequently, a Gaussian mixture background modeling algorithm (Gaussian mixture background modeling algorithm) is used. Extract the foreground moving target, which is biological bacteria. For ease of description, it will be referred to as the moving target below. Perform morphological opening operation on the binarized image to remove small noise and extract the centroid coordinates of all connected components.

[0040] Finally, using the Hungarian matching algorithm based on Kalman filter (KF) prediction, the centroids of moving targets in adjacent frames are correlated to construct the centroids of each moving target. Time series trajectory set ,in Total number of frames Indicates the target of motion No. The horizontal coordinate of the centroid in a frame image, or it can also be understood as the time. The x-coordinate of the centroid; Indicates the target of motion No. The centroid ordinate in the frame image is used as the motion trajectory of the time series trajectory set.

[0041] S2: For any motion trajectory, calculate the adaptive diffusion threshold for activity determination of the motion trajectory.

[0042] Reference Figure 2 Step S2 includes steps S21-S23.

[0043] S21: For any motion trajectory, analyze its short-term mean square displacement variation characteristics to obtain the instantaneous modal variability of the motion trajectory. The instantaneous modal variability is used to characterize the variation characteristics of the short-term mean square displacement along the motion trajectory.

[0044] Plant probiotics similar to Bacillus subtilis are driven by flagella, and their movement mechanism causes their diffusion properties to change drastically in a short period of time. Their movement is close to ballistic motion, and when they tumble, they are close to restricted diffusion, while the diffusion properties of organic residues in the background are relatively stable.

[0045] The motion trajectory is divided into sliding windows along the time axis; the instantaneous diffusion index within each sliding window is calculated; and the product of the standard deviation and the sum of the absolute values ​​of the differences of the instantaneous diffusion index is taken as the instantaneous modal variability.

[0046] Specifically, a sliding time window is constructed. In this embodiment, the length of the sliding time window is [value missing]. Frames, with a step size of Frame. For the first For any sliding time window of a motion trajectory, the slope obtained by linearly fitting the mean square displacement of different time intervals within that window to the logarithm of the time interval is used as the instantaneous diffusion index.

[0047] The instantaneous modal variability of any motion trajectory can be calculated using the following formula: In the formula, Indicates the first Instantaneous modal variability of a motion trajectory; This represents the total number of sliding windows segmented from the trajectory; Indicates the first The moving target in the first Instantaneous diffusion index within a sliding window; Indicates the first The arithmetic mean of the instantaneous diffusion index within all windows of a motion trajectory.

[0048] In the formula, This reflects the standard deviation of the diffusion index; the larger the value, the more pronounced the diffusion index. The more drastically a movement trajectory deviates from the average state during the entire observation period (i.e., the existence of alternating fast and slow movements), the more consistent it is with the characteristics of active bacteria.

[0049] This represents the sum of the absolute differences of the instantaneous diffusion index over the time series, i.e., the cumulative volatility; the larger this value, the greater the volatility. The more frequently and significantly a movement trajectory switches between adjacent time windows, the more it aligns with the characteristics of active bacteria actively regulating their movement posture. The product of these two factors can effectively amplify the bioactivity characteristics, thereby improving the accuracy and robustness of probiotic activity assessment.

[0050] S22: Analyze the correlation between the motion vectors of the moving target and the local flow field during the mode switching period of the motion trajectory to obtain the biological mode confidence score. The biological mode confidence score is used to characterize the possibility that the switching of motion modes in the motion trajectory is biologically driven.

[0051] To address potential micro-turbulence or thermal convection in the fermentation broth, passively moving organic residues must be removed. The movement of organic residues is passive, and their velocity vector is highly consistent with the surrounding fluid (other particles in the neighborhood); while active probiotics have autonomous obstacle avoidance or chemotaxis capabilities, and their movement direction is less consistent with the flow field direction.

[0052] Obtain the set of key time windows in the motion trajectory where mode switching occurs; for any time window in the set of key time windows, obtain the velocity vector of the moving target and the flow field vector within the local range of the moving target in that time window.

[0053] Obtain the cosine of the angle between the velocity vector of the moving target and the local flow field vector; the confidence level of the biological modality is positively correlated with the sum of the cosine values ​​of the angle at each moment in the key time window set.

[0054] Specifically, the biomodal confidence score for each motion trajectory is calculated using the following formula:

[0055] In the formula, Indicates the first Biological modality confidence of a motion trajectory; Indicates the first The moving target is in the critical time window The velocity vector in the equation can be the ratio of the displacement vector of the moving target within the critical time window to the length of the critical time window. Indicates the first The set of critical time windows in which mode switching occurs within a trajectory; Indicates the first The window moment, the first The velocity vector of a moving target Its local flow field vector The angle between them This is obtained by calculating the average velocity of all other background particles within a preset radius around the moving target. In this embodiment, the preset radius is set to... 1 pixel; For a linear rectified function, when Time output Otherwise, the key time window set in this embodiment is defined as the set of time points where the instantaneous diffusion index is greater than a preset threshold. To prevent the minimum value where the denominator is zero, the preferred method is... ; Indicates An exponential function with base 0.

[0056] The steps for constructing the set of critical time windows for mode switching include: through step S21 above, for any sliding window, there is a corresponding instantaneous diffusion index. If the difference between the instantaneous diffusion index and the mean of all instantaneous diffusion indices in the motion trajectory is greater than a preset diffusion threshold, then the sliding window is considered to be a critical time window for the switching or change of the motion mode of the moving target. All critical time windows with a value greater than the preset diffusion threshold are counted to form a set of critical time windows. In this embodiment, the diffusion threshold is set to 0.2 based on the experience of those skilled in the art.

[0057] In the formula, the exponent term It has an inhibitory effect. If The larger the value, the more likely it is that at the critical moment of mode switching, the first... The more the direction of motion of a moving target is consistent with the direction of the flow field (i.e., the more consistent the direction of motion of the target is with the direction of the flow field) If the angle is large or the direction is opposite, the exponential term is closer to... The greater the confidence level of the biological modality, the higher the confidence level.

[0058] S23: The product of the biomodal confidence and instantaneous modal volatility of the motion trajectory is used as the biofeature value, and the biofeature value is used to adjust the preset initial threshold to obtain the adaptive diffusion threshold.

[0059] The final decision threshold is dynamically adjusted based on the calculated biometric values. If the biometric value is high, the threshold is lowered to prevent the missed detection of slow-moving bacteria with pattern-switching characteristics; if the biometric value is low, the threshold is raised to eliminate interference.

[0060] In this embodiment, the biometric value is normalized using the hyperbolic tangent function to obtain the standard value; the difference between the preset initial threshold and the standard value is used as the adaptive diffusion threshold.

[0061] Specifically, the formula for calculating the adaptive diffusion threshold can be expressed as:

[0062] In the formula, Indicates the first An adaptive diffusion threshold for each motion trajectory; This represents a preset initial threshold, which is preferably specified in this embodiment. . The hyperbolic tangent function is used to convert biometric values... Mapped to interval; Represents the trajectory of motion Biometric values.

[0063] Therefore, it can be concluded that the larger the biometric value, Represents the standard eigenvalue. The closer ,but The smaller the value, the more lenient the judgment criteria; the smaller the biometric value, the more lenient the judgment criteria. The closer ,but The higher the value, the stricter the judgment criteria.

[0064] S3: Determine the activity of plant probiotics based on adaptive diffusion threshold.

[0065] Using the adaptive parameters calculated in step S2, real-time and quantitative determination of the activity of bio-fertilizer is achieved on the production site. The specific operation is as follows: For each moving target in the video stream... First, calculate the global average diffusion index of the entire trajectory, i.e., the first... The global slope is obtained by fitting the mean square displacement (MSD) of the entire trajectory of the moving target. Simultaneously, its specific adaptive threshold is calculated according to step S23. Perform binary classification: if If the target of the movement is determined to be highly active plant probiotics; if If so, the moving target is determined to be an inactive particle.

[0066] Combination Figure 3 This graph distinguishes different types of targets using two physical dimensions. The horizontal axis represents the biomodal confidence of the moving target. A low value indicates that the moving target follows the water flow and is usually inactive impurities. A high value indicates that the target's movement direction is inconsistent with the water flow, demonstrating the active nature of the organism. The vertical axis represents the instantaneous modal fluctuation of the moving target. A low value indicates slow movement. A high value indicates fast movement, which could be swimming bacteria or debris washed away by a strong water flow. True active bacteria, represented by pentagrams in the graph, are usually distributed in the upper right corner. Organic debris, represented by triangles in the graph, is usually distributed in the upper left corner. Traditional fixed threshold lines are a one-size-fits-all line, only considering the vertical axis. Anything above this line is considered active, thus potentially misidentifying organic debris as active bacteria. In this application, the adaptive threshold is a dynamically changing curve, thereby reducing the occurrence of organic debris being identified as active bacteria and improving the accuracy of active bacteria activity detection.

[0067] The system calculates the density (unit: cells / mL) and activity rate (number of active targets / total number of targets) of viable bacteria within the current field of view. This data is transmitted to the central controller in real time. If the activity rate falls below the set standard, the system automatically triggers an alarm and issues adjustment commands to maintain the stability of the fermentation process.

[0068] This application also discloses a rapid identification system for the activity of plant probiotics, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a rapid identification method for the activity of plant probiotics according to this application is implemented.

[0069] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0070] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A rapid method for determining the activity of plant probiotics, characterized in that, Microscopic video streams containing plant probiotics and organic matrix particles are collected, and the motion trajectory of each moving target is obtained by trajectory tracking. For any motion trajectory, calculate the adaptive diffusion threshold for activity determination of the motion trajectory; The activity of plant probiotics is determined based on an adaptive diffusion threshold. The steps for calculating the adaptive diffusion threshold for activity determination of the motion trajectory include: for any motion trajectory, analyzing its short-term mean square displacement variation characteristics to obtain the instantaneous modal volatility of the motion trajectory, which is used to characterize the variation characteristics of the short-term mean square displacement along the motion trajectory; the calculation steps for the instantaneous modal volatility include: dividing the motion trajectory into sliding windows along the time axis; calculating the instantaneous diffusion index within each sliding window; taking the product of the standard deviation and the sum of the absolute values ​​of the differences of the instantaneous diffusion index as the instantaneous modal volatility; calculating the instantaneous diffusion index within each sliding window includes: taking the slope obtained by linearly fitting the mean square displacement of different time intervals under the window to the logarithmic relationship of the time interval as the instantaneous diffusion index; The correlation between the motion vectors of the moving target and the local flow field during the mode switching period of the motion trajectory is analyzed to obtain the biological mode confidence score. The biological mode confidence score is used to characterize the possibility that the switching of motion modes in the motion trajectory is biologically driven. The calculation of the biological mode confidence score includes: obtaining the set of key time windows in the motion trajectory where mode switching occurs; for any time window in the set of key time windows, obtaining the velocity vector of the moving target and the flow field vector within the local range of the moving target in that time window. Obtain the cosine of the angle between the velocity vector of the moving target and the local flow field vector; the confidence level of the biological modality is positively correlated with the sum of the cosine values ​​of the angle at each moment in the key time window set. The product of the biomodal confidence and instantaneous modal volatility of the motion trajectory is used as the biofeature value. The biofeature value is used to adjust the preset initial threshold to obtain the adaptive diffusion threshold. The adaptive diffusion threshold is calculated by: normalizing the biofeature value using the hyperbolic tangent function to obtain the standard feature value; and using the difference between the preset initial threshold and the standard feature value as the adaptive diffusion threshold.

2. The rapid method for determining the activity of plant probiotics according to claim 1, characterized in that, The steps for obtaining the set of key time windows in the motion trajectory where mode switching occurs include: taking the average instantaneous diffusion index of each sliding window in the motion trajectory as the diffusion benchmark, taking the sliding window whose instantaneous diffusion index differs from the diffusion benchmark by more than a preset diffusion threshold as the key time window, and selecting multiple key time windows to form a set of key time windows.

3. The rapid method for determining the activity of plant probiotics according to claim 1, characterized in that, The steps for obtaining the flow field vector within a local area of ​​the moving target include: extracting the velocity vectors of other background particles within a preset radius around the moving target during the time window, and using the average value of the velocity vectors of the other background particles as the local flow field vector.

4. The rapid method for determining the activity of plant probiotics according to claim 1, characterized in that, The steps for tracking moving targets and obtaining their motion trajectories include: extracting moving targets from the video stream using a Gaussian mixture model; performing morphological opening operations on the moving targets to remove noise and extracting the centroids of the connected components corresponding to the moving targets as their coordinates; and generating motion trajectories by associating the coordinates of moving targets in adjacent frames based on Kalman filter prediction and Hungarian algorithm matching.

5. The rapid method for determining the activity of plant probiotics according to claim 1, characterized in that, After acquiring the microscopic video stream containing plant probiotics and organic matrix particles, the acquired image sequence was preprocessed using a contrast-limited adaptive histogram equalization algorithm.

6. A rapid identification system for the activity of plant probiotics, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a rapid method for determining the activity of plant probiotics according to any one of claims 1-5.

Images