A method and system for screening a high-temperature-resistant lactic acid bacteria

Abstract

The application discloses a high-temperature-resistant lactic acid bacteria screening method and system, and particularly relates to the field of image processing and image recognition, and is used for solving the problem that the source of pixel color change cannot be traced due to the overlapping of multi-colony color halos during high-throughput screening; the application obtains multi-time sequence images of candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture, extracts the color change curve of each pixel, compares the starting change time and change direction of the pixel and its neighborhood pixels, marks passive diffusion pixels and autonomous acid-producing pixels, clusters the autonomous acid-producing pixels to form a candidate colony area, traces the passive diffusion pixels along the earliest discoloration direction to form a propagation streamline and distributes the propagation streamline to the corresponding candidate colony area, combines to obtain an independent acid-producing coloration area, and screens the high-temperature-resistant lactic acid bacteria according to the area change rate.

Description

Technical Field

[0001] This invention relates to the field of image processing and image recognition technology, and more specifically, to a method and system for screening thermoresistant lactic acid bacteria. Background Technology

[0002] In the field of microbial strain screening, using machine vision and image processing technology to replace traditional manual biochemical testing can improve screening throughput and objectivity. For screening thermotolerant lactic acid bacteria, the conventional approach is to spread the strain to be screened onto a petri dish containing an indicator, place it in a high-temperature incubation environment, and periodically acquire images of the petri dish using an industrial camera. Image analysis algorithms are then used to extract the morphological characteristics of each candidate colony and the color changes of the surrounding culture medium. This allows for the assessment of the acid production rate and growth activity of the strain under high-temperature stress, thereby screening out the target strain. The core of this process lies in establishing a one-to-one correspondence between individual colonies and the color areas they cause on the culture medium, and characterizing the strength of the strain's acid production capacity by quantifying the expansion rate of the color areas.

[0003] However, in actual high-throughput screening, to improve screening efficiency, a large number of candidate strains often need to be densely arranged on the culture plate. Since the target screening object is a rapidly acid-producing strain, the acidic metabolites produced by it under high temperature will quickly diffuse in all directions in the agar medium, forming a color halo much larger than the size of the colony itself. When the diffusion halos of adjacent colonies overlap, the spatial segmentation algorithm based on single-frame images can only rigidly divide the fused color region according to the geometric boundary, and cannot trace the actual source of color change at each pixel position. As a result, the independent contribution of the strongly heat-resistant and highly acid-producing strain is shared by the neighboring weakly heat-resistant strain, which makes the acid production performance of the latter falsely inflated, while the true performance of the former is relatively underestimated. This causes the judgment of the heat resistance and acid production ability of the candidate strain to deviate from the actual situation, which restricts the accuracy and reliability of the image-based screening results. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for screening thermoresistant lactic acid bacteria to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for screening thermoresistant lactic acid bacteria, comprising: S1: Acquire multiple time-series images of candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture conditions; S2: For each pixel in multiple time-series images, extract the color value of the pixel in each time-series image to form the color change curve of the pixel; S3: For each pixel, compare the starting time and direction of the color change curve of the pixel with those of each pixel in its spatial neighborhood. If the starting time of the pixel's change lags behind that of a certain pixel in its neighborhood and the two change directions are consistent, then mark it as a passively diffused pixel; otherwise, mark it as a self-generating acid pixel. S4: Cluster the autonomous acid-producing pixels according to their spatial location to form multiple candidate colony regions; S5: For each passively diffused pixel, trace it pixel by pixel along the direction of the pixel with the earliest color change start time in its neighborhood to form a propagation streamline until it reaches a candidate colony area, and assign the passively diffused pixel to the candidate colony area. S6: Merge the autonomous acid-producing pixels and passive diffusion pixels corresponding to each candidate colony region to obtain independent acid-producing color regions, and screen thermotolerant lactic acid bacteria based on the area change rate of the independent acid-producing color regions in multiple time-series images.

[0006] Furthermore, S1 includes: Place the culture medium containing the indicator, inoculated with multiple candidate lactic acid bacteria colonies, in a high-temperature incubation environment; Set a preset time interval and use an industrial camera to take pictures of culture medium plates at multiple acquisition time points to obtain single-frame plate images corresponding to each acquisition time point; The single-frame plate images corresponding to each acquisition time point are arranged in the order of acquisition time to form a multi-time sequence image.

[0007] Furthermore, S2 includes: Determine the color value corresponding to the same pixel position in each time series image from multiple time series images; Arrange the color values ​​corresponding to the same pixel location in each time series image according to the order of acquisition time; The arranged color value sequence is used as the color change curve of the corresponding pixel.

[0008] Furthermore, S3 includes: For each pixel, identify the moment when the color value first breaks through the preset stable fluctuation range from the color change curve of the corresponding pixel, and take it as the starting moment of change; For each pixel, determine the neighboring pixel with the earliest start change time in the neighborhood. If the start change time of the pixel lags behind the start change time of the earliest neighboring pixel, and the lag time is within the preset neighborhood propagation lag range, then the pixel is pre-judged as a candidate passive diffusion pixel. For each candidate passive diffusion pixel, compare the color change curve of the pixel after the initial change time with the color change curve of the earliest neighboring pixel after the initial change time at multiple consecutive time points to see if the change direction is consistent. Candidate passive diffusion pixels with consistent change direction are marked as passive diffusion pixels, and candidate passive diffusion pixels with inconsistent change direction are marked as autonomous acid-producing pixels. Pixels that were not pre-judged as candidate passive diffusion pixels are marked as autonomous acid-producing pixels.

[0009] Furthermore, candidate passive diffusion pixels with consistent change directions are marked as passive diffusion pixels. This includes: comparing the color change curve of the candidate passive diffusion pixel after the initial change time with the color change curve of the earliest neighboring pixel after the initial change time, and comparing the number of consecutive time points in which the color value moves along the same trajectory in the color space point by point. If the proportion of the number of consecutive time points to the total number of compared time points is greater than or equal to a preset proportion threshold, the change direction is determined to be consistent. If the proportion of the number of consecutive time points to the total number of compared time points is less than the preset proportion threshold, the change direction is determined to be inconsistent.

[0010] Furthermore, S4 includes: Identify autonomous acid-producing pixels that exhibit local density maxima in space from all autonomous acid-producing pixels and use them as cluster center pixels; Starting from each cluster center pixel, autonomous acid-producing pixels that are spatially located within the preset neighborhood of the cluster center pixel and have the same direction of change on the color change curve as the cluster center pixel are assigned to the cluster corresponding to the cluster center pixel. The spatial region covered by autonomous acid-producing pixels belonging to the same cluster is considered as a candidate colony region.

[0011] Furthermore, autonomous acid-producing pixels that are spatially located within a preset neighborhood of the cluster center pixel and have the same direction of change as the cluster center pixel on the color change curve are assigned to the cluster corresponding to the cluster center pixel. This includes: comparing the direction of change of each autonomous acid-producing pixel within the preset neighborhood of the cluster center pixel with the direction of change of the cluster center pixel; if the direction of change of the autonomous acid-producing pixel is the same as the direction of change of the cluster center pixel, then the autonomous acid-producing pixel is assigned to the cluster corresponding to the cluster center pixel; otherwise, it is not assigned to the cluster corresponding to the cluster center pixel.

[0012] Furthermore, S5 includes: For each passively diffused pixel, the neighboring pixel whose color change started earliest in its spatial neighborhood is determined as the tracking direction pixel. Starting from the passive diffusion pixel, the tracking direction pixel is taken as the current tracking pixel. The earliest neighboring pixel in the spatial neighborhood of the current tracking pixel is repeatedly determined as the next tracking pixel, and the propagation streamline is gradually formed. When the propagation stream advances to the point where the currently tracked pixel is an autonomous acid-producing pixel, the advancement of the propagation stream is terminated; the passive diffusion pixel at the beginning of the propagation stream is assigned to the candidate colony region where the autonomous acid-producing pixel at the end is located.

[0013] Furthermore, S6 includes: For each candidate colony region, all autonomous acid-producing pixels and all passive diffusion pixels belonging to the candidate colony region are merged to form an independent acid-producing color development region. In each frame of the multi-time series images, the area of ​​the independent acid-producing colorimetric region is calculated, and the area change rate over time is obtained. When the area change rate of the independent acid-producing colorimetric region of the candidate colony area is greater than or equal to the preset area change rate threshold, the candidate lactic acid bacteria colony corresponding to the candidate colony area is determined to be thermostable lactic acid bacteria.

[0014] On the other hand, the present invention provides a screening system for thermoresistant lactic acid bacteria, comprising: The image acquisition module is used to acquire multi-time-series images of multiple candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture conditions; The curve forming module is used to extract the color value of each pixel in multiple time-series images and form the color change curve of the pixel. The pixel marking module is used to compare the starting time and direction of the color change curve of each pixel with those of each pixel in its spatial neighborhood. If the starting time of the pixel's change lags behind that of a neighboring pixel and the two change directions are consistent, it is marked as a passively diffusing pixel; otherwise, it is marked as a self-generating acid pixel. The region formation module is used to cluster autonomous acid-producing pixels according to their spatial location to form multiple candidate colony regions; The pixel allocation module is used to track each passively diffused pixel pixel by pixel along the direction of the pixel with the earliest color change start time in its neighborhood, forming a propagation streamline until it reaches a candidate colony area, and then allocate the passively diffused pixel to the candidate colony area. The lactic acid screening module is used to merge the autonomous acid-producing pixels and passive diffusion pixels corresponding to each candidate colony region to obtain an independent acid-producing colorimetric region. The heat-resistant lactic acid bacteria are screened based on the area change rate of the independent acid-producing colorimetric region in multiple time-series images.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By analyzing the spatiotemporal characteristics of pixel color changes in multi-temporal images, pixel-level tracing of the acid production and diffusion behavior of candidate lactic acid bacteria colonies under high-temperature stress was achieved. The starting time and direction of color change curve of each pixel were used as the basis for judgment. By comparing the temporal relationship of color change of a pixel with that of its spatial neighbors, autonomous acid-producing pixels and passively diffusing pixels were distinguished. This allows the driving source of color change to be directly identified at the image level without the need for additional biochemical markers or physical isolation methods. This discrimination method based on the causal direction between pixels transforms the problem of overlapping color halos, which is difficult to handle by rigid geometric segmentation in traditional methods, into a diffusion attribution problem that can be solved pixel by pixel.

[0016] 2. By tracking passively diffusing pixels pixel by pixel along the decreasing direction at the start of color change, a propagation flow line was constructed from the diffusion receiver back to the acid-producing source. This enabled the directional allocation of each passively diffusing pixel to its corresponding candidate colony region. After merging autonomously acid-producing pixels with the traced passively diffusing pixels, the independent acid-producing colorimetric range corresponding to each candidate colony region was restored. This spatially decoupled the originally overlapping diffusion halos. Screening was performed based on the area change rate of the independent acid-producing colorimetric region, ensuring that the acid-producing capacity assessment of each candidate colony relied solely on its own metabolically driven colorimetric expansion rate. This avoided interference from metabolic crosstalk between neighboring colonies on the assessment results and improved the accuracy of image-based screening of thermostable lactic acid bacteria. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for screening thermoresistant lactic acid bacteria according to the present invention; Figure 2 This is a schematic diagram of the structure of a screening system for heat-resistant lactic acid bacteria according to the present invention. Detailed Implementation

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

[0019] Example 1: Figure 1 The present invention provides a method for screening thermoresistant lactic acid bacteria, comprising: S1: Acquire multiple time-series images of candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture conditions; S2: For each pixel in multiple time-series images, extract the color value of the pixel in each time-series image to form the color change curve of the pixel; S3: For each pixel, compare the starting time and direction of the color change curve of the pixel with those of each pixel in its spatial neighborhood. If the starting time of the pixel's change lags behind that of a certain pixel in its neighborhood and the two change directions are consistent, then mark it as a passively diffused pixel; otherwise, mark it as a self-generating acid pixel. S4: Cluster the autonomous acid-producing pixels according to their spatial location to form multiple candidate colony regions; S5: For each passively diffused pixel, trace it pixel by pixel along the direction of the pixel with the earliest color change start time in its neighborhood to form a propagation streamline until it reaches a candidate colony area, and assign the passively diffused pixel to the candidate colony area. S6: Merge the autonomous acid-producing pixels and passive diffusion pixels corresponding to each candidate colony region to obtain independent acid-producing color regions, and screen thermotolerant lactic acid bacteria based on the area change rate of the independent acid-producing color regions in multiple time-series images.

[0020] When performing S1, provide Petri dishes containing indicators inoculated with multiple candidate lactic acid bacteria colonies. The candidate lactic acid bacteria colonies are derived from lactic acid bacteria strains isolated from different natural environments or obtained from strain collections. The lactic acid bacteria strains are activated and cultured separately in MRS liquid medium to the logarithmic growth phase, washed with sterile physiological saline, and the bacterial suspension concentration is adjusted to approximately 10⁻⁶ / mL. 6 Up to 10 8 Each colony-forming unit (CFU) is used. An appropriate amount of bacterial suspension is evenly spread onto a Petri dish containing an indicator, specifically MRS medium supplemented with bromocresol purple indicator and agar. The final concentration of the bromocresol purple indicator is, for example, 0.04 g / L to 0.06 g / L. Bromocresol purple indicator is purple at pH above 5.2 and yellow at pH below 5.2. The acid produced by lactic acid bacteria metabolism lowers the pH of the medium, causing the color change. After spreading, the Petri dishes are air-dried in a laminar flow hood to obtain Petri dishes containing multiple candidate lactic acid bacteria colonies and the indicator.

[0021] Petri dishes containing indicators and inoculated with multiple candidate lactic acid bacteria colonies are placed in a high-temperature incubation environment. This environment is provided by a constant-temperature incubator, with the internal temperature pre-set to a fixed value within the range of 45°C to 55°C, for example, 50°C. The petri dishes are placed horizontally on a partition within the incubator, with the lids facing upwards, ensuring air circulation and uniform heating of the petri dishes. The incubator maintains the set temperature throughout the incubation process, with temperature fluctuations not exceeding ±0.5°C. This high-temperature condition is used to stress the candidate lactic acid bacteria colonies; only those colonies exhibiting heat resistance and capable of normal acid production can survive and produce a noticeable colorimetric reaction under these conditions.

[0022] A preset time interval is set, and an industrial camera is used to capture images of culture plates containing multiple candidate lactic acid bacteria colonies and indicators at multiple acquisition time points, obtaining a single-frame image of the plate corresponding to each acquisition time point. The preset time interval is set according to the growth rate and acid production rate of lactic acid bacteria at high temperature, and the preset time interval ranges from 10 minutes to 60 minutes, for example, set to 30 minutes. The industrial camera is a digital camera with color image acquisition capability, and the image sensor type is complementary metal-oxide-semiconductor CMOS or charge-coupled device CCD. The resolution of the industrial camera is set to, for example, 12 megapixels or higher, and the focal length of the industrial camera lens is fixed. The shooting distance and field of view of the industrial camera cover the entire culture plate containing multiple candidate lactic acid bacteria colonies and indicators. The industrial camera is fixed to the outside of the constant temperature incubator by a bracket, and the lens of the industrial camera is aimed at the transparent observation window of the constant temperature incubator, or the industrial camera is placed directly inside the constant temperature incubator to eliminate the influence of glass refraction. Before the actual shooting began, the exposure time, aperture, and gain parameters of the industrial camera were adjusted to ensure that the brightness of each single-frame plate image was moderate and the contrast between the candidate lactic acid bacteria colonies and the culture medium background was clear. These parameters were maintained constant throughout all subsequent acquisition points. During shooting, a ring-shaped LED white light source surrounding the industrial camera lens was used to evenly illuminate the surface of the culture medium plates containing multiple candidate lactic acid bacteria colonies and indicators, avoiding shadows and reflections. At multiple pre-defined acquisition points, the industrial camera automatically triggered shooting, capturing one color digital image (single-frame plate image) each time. All single-frame plate images recorded the morphological and color changes of the candidate lactic acid bacteria colonies on the same culture medium plate containing multiple candidate lactic acid bacteria colonies and indicators throughout the entire culture cycle.

[0023] The single-frame petri dish images corresponding to each acquisition time point are arranged in chronological order to form a multi-time-series image. Specifically, each single-frame petri dish image file is named with the timestamp of the acquisition time point and numbered sequentially from earliest to latest timestamp, generating an ordered image sequence, which is the multi-time-series image. Each frame in the multi-time-series image corresponds to an acquisition time point, and the entire multi-time-series image contains complete visual information on the evolution of all candidate lactic acid bacteria colonies over time from the start to the end of cultivation. The multi-time-series image serves as the basis for subsequent color change analysis of each pixel, where the color value of each pixel in each time-series image is extracted to form a color change curve. Through the above steps, multi-time-series images of multiple candidate lactic acid bacteria colonies on culture plates containing indicators under high-temperature cultivation conditions are obtained.

[0024] In S2, for each pixel in the multi-time series images, the color value of the pixel in each time series image is extracted to form a color change curve of the pixel. The multi-time series images are obtained from S1. Each frame of the single-frame plate image in the multi-time series images has the same pixel resolution. Pixels at the same coordinate position in each frame of the single-frame plate image in the multi-time series images correspond to the same physical position on a culture medium plate containing an indicator and inoculated with multiple candidate lactic acid bacteria colonies. The color value corresponding to the same pixel position in each time series image is determined from the multi-time series images. For each pixel position in the multi-time series images, all frames of the multi-time series images are traversed, and the color value of the pixel position is read from each frame of the single-frame plate image. The color value is represented by the red component value, green component value, and blue component value in the RGB color space. The red component value ranges from, for example, 0 to 255, the green component value ranges from, for example, 0 to 255, and the blue component value ranges from, for example, 0 to 255. The color values ​​are read by extracting the red, green, and blue component values ​​at the same pixel location coordinates in the multi-time series images frame by frame. The extracted red, green, and blue component values ​​for each frame are then associated with and recorded along with the corresponding acquisition time point. During the traversal, the pixel location coordinates remain unchanged, and each frame in the multi-time series images arranged in chronological order of acquisition time is traversed sequentially to ensure a one-to-one correspondence between the acquired color values ​​and the acquisition time points. Since a preset time interval is set in S1 and the multi-time series images are arranged in chronological order of acquisition time, the acquisition time point of each frame is known. The acquisition time point is the cumulative time calculated from the start of cultivation, for example, starting from 0 minutes. An acquisition time point is generated every preset time interval, with minutes as the unit of time. A correspondence is established between the acquisition time point of each frame and the extracted color values, forming time-stamped color value data.

[0025] The color values ​​corresponding to the same pixel location in each time series image are arranged sequentially according to the acquisition time. For the extracted time-labeled color value data corresponding to the same pixel location in each time series image, the color value data is arranged sequentially from earliest to latest acquisition time. The arrangement is based on the acquisition time, with the color value data corresponding to the earliest acquisition time frame ranked first, the color value data corresponding to the second earliest acquisition time frame ranked second, and so on, until the color value data corresponding to the latest acquisition time frame ranked last. The arranged data is a sequence of color value data arranged in chronological order of acquisition time. Each element in the color value data sequence contains an acquisition time point and the corresponding red, green, and blue component values ​​at the pixel location at that acquisition time point. During the arrangement process, the color value data is not modified, interpolated, or smoothed to maintain the original correspondence between the color value data and the acquisition time points, ensuring the authenticity of the color change curve.

[0026] The arranged color value sequence is used as the color change curve for the corresponding pixel. For each pixel location in a multi-time series image, the arranged color value data sequence is treated as a whole and assigned to the pixel location; this data object is the pixel's color change curve. The color change curve records the complete trajectory of the red, green, and blue component values ​​of the pixel location over all covered acquisition time points. The red component value sequence in the color change curve is used to analyze the evolution of the red component in the pixel color over time, the green component value sequence is used to analyze the evolution of the green component in the pixel color over time, and the blue component value sequence is used to analyze the evolution of the blue component in the pixel color over time. On a culture medium plate containing bromocresol purple indicator, the medium is purple when no acid is produced, corresponding to higher red and blue component values ​​and lower green component values ​​in the RGB color space. When lactic acid bacteria metabolize and produce acid, causing the pH to drop below 5.2, the medium color gradually changes from purple to yellow, corresponding to a gradual increase in red and green component values ​​and a gradual decrease in blue component values ​​in the RGB color space. The decreasing trend of the blue component value and the increasing trend of the red and green component values ​​in the color change curve together reflect the progress of the culture medium acidification process at the pixel location.

[0027] Color change curves serve as the basis for subsequent determination of whether each pixel is an autonomously acid-producing pixel or a passively diffusing pixel. The color change curves of autonomously acid-producing pixels exhibit color change characteristics driven by autonomous metabolism, while the color change curves of passively diffusing pixels exhibit color change characteristics caused by diffusion and conduction. A complete process is executed for each pixel location in the multi-temporal images, from determining the corresponding color value of the same pixel location in each temporal image to using the arranged color value sequence as the color change curve for the corresponding pixel. This results in a set of color change curves equal in number to the number of pixel locations in the multi-temporal images, with each pixel location having an independent color change curve.

[0028] In S3, for each pixel, the starting time and direction of color change curves of the pixel and its spatial neighbors are compared to distinguish passively diffusing pixels from autonomously acid-producing pixels. The color change curves are obtained from S2, and each curve contains the red, green, and blue component values ​​of the corresponding pixel at all acquisition time points. For each pixel, the moment when the color value first breaks through the preset stable fluctuation range from the corresponding pixel's color change curve is identified as the starting time of change. The preset stable fluctuation range is determined as follows: several acquisition time points are selected within a time window during the initial culture period before significant color changes occur in the culture medium, for example, the first three acquisition time points from 0 minutes to 60 minutes. The color values ​​of the pixel at these acquisition time points are read. Taking the blue component value in the RGB color space as an example, the average and standard deviation of the blue component value are calculated. The average reflects the initial color baseline, and the standard deviation reflects the noise fluctuation amplitude. The preset stable fluctuation range is set to the blue component value between the average minus, for example, twice the standard deviation, and the average plus, for example, twice the standard deviation. If the blue component value of a pixel at a certain acquisition time point exceeds the preset stable fluctuation range, the color value is determined to have broken through the preset stable fluctuation range. Since the blue component value is most sensitive to acidification, it continuously decreases during the process of the bromocresol purple indicator changing from purple to yellow. Breaking through the preset stable fluctuation range means that the pixel has begun to undergo a detectable color change. The acquisition time point at which the preset stable fluctuation range is first broken is recorded as the pixel's initial change time. The initial change time is a specific minute value counted from the start of cultivation. For pixels that have not broken through the preset stable fluctuation range, the initial change time is marked as a null value, indicating that the pixel has not undergone a significant color change. For each pixel, the direction of the color change curve after the initial change time is also recorded. The direction of change is defined as the trend of the color value movement in the RGB color space, that is, the three-dimensional direction vector formed by the changes in the red, green, and blue component values ​​between two adjacent acquisition time points. For example, if the red component value remains unchanged, the green component value increases, and the blue component value decreases between two adjacent acquisition time points, the direction of change is represented by the red component change sign 0, the green component change sign +, and the blue component change sign -, where 0 indicates no change, + indicates an increase, and - indicates a decrease. At each acquisition time point after the initial change moment, the direction of change between adjacent time points is calculated sequentially, resulting in a sequence of direction changes, which serves as the pixel's direction of change feature.

[0029] For each pixel, determine the neighboring pixel with the earliest starting change time within its spatial neighborhood. The spatial neighborhood consists of pixels within a square window with a preset edge length around the pixel. For example, if the preset edge length is 5 pixels, the spatial neighborhood includes the pixel and a set of pixels within two pixels above, below, to the left, and to the right. Iterate through the starting change times of all pixels within the spatial neighborhood and find the neighboring pixel with the smallest starting change time value, i.e., the neighboring pixel that first underwent a color change. If the starting change time of a pixel lags behind the starting change time of its earliest neighboring pixel (i.e., the lag time obtained by subtracting the starting change time of the earliest neighboring pixel from the starting change time of the pixel is a positive number, and the lag time is within the preset neighborhood propagation lag range), then the pixel is pre-judged as a candidate passive diffusion pixel. The preset neighborhood propagation lag range is determined based on the physical diffusion rate of acidic metabolites in the culture medium and the actual physical size represented by the pixel. The lactic acid diffusion coefficient in the culture medium is approximately 1 × 10⁻⁶. -5 In a petri dish imaging resolution, each pixel corresponds to an actual physical size of 0.01 cm on the culture medium. The theoretical time required for acidic metabolites to diffuse across one pixel is approximately the square of 0.01 cm divided by the diffusion coefficient, which is on the order of 10 seconds. Multiplying the theoretical time by a safety factor, such as 5 to 10 times, yields the upper limit of the preset neighborhood propagation lag range, for example, set to 2 minutes. The lower limit of the preset neighborhood propagation lag range is set to 0 minutes. Pixels with lag times falling between 0 minutes and the upper limit are considered to be potentially affected by the diffusion of the earliest neighboring pixel and are pre-judged as candidate passive diffusion pixels. If the lag time exceeds the upper limit, it indicates that although the initial change time is delayed, the lag time is too long and cannot be caused by the same diffusion event; the pixel is not pre-judged as a candidate passive diffusion pixel. If the initial change time of a pixel is not delayed compared to the initial change time of the earliest neighboring pixel, meaning the pixel itself is the earliest changing pixel in its spatial neighborhood, the pixel is not pre-judged as a candidate passive diffusion pixel.

[0030] Pixels not initially identified as candidate passively diffusing pixels are directly labeled as autonomously acid-producing pixels. These pixels are either the earliest to change color in their spatial neighborhood, or, although delayed, their delay exceeds a preset neighborhood propagation time lag range, indicating that the pixel's color change is caused by its own acid production, rather than by receiving acidic metabolites diffused from neighboring pixels. The candidate lactic acid bacteria colonies corresponding to autonomously acid-producing pixels possess autonomous metabolic acid-producing capabilities.

[0031] For each candidate passively diffused pixel, the consistency of the change direction is further verified to ultimately determine the properties of the candidate passively diffused pixel. The color change curve of the candidate passively diffused pixel after the initial change time is compared with the color change curve of the earliest neighboring pixel after the initial change time at multiple consecutive time points to see if the change direction is consistent. The change direction sequences of the candidate passively diffused pixel and the earliest neighboring pixel from their respective initial change time to the most recent acquisition time point are extracted. Since the initial change time of the candidate passively diffused pixel and the initial change time of the earliest neighboring pixel may be different, the time window needs to be aligned during comparison: either select an acquisition time point range commonly covered by the two pixels, or start from after the initial change time of the earliest neighboring pixel and continue for a preset number of acquisition time points after the initial change time of the candidate passively diffused pixel. At each time point within the time window, the change direction of the candidate passively diffused pixel from that time point to the next time point is compared to see if it is completely consistent with the change direction of the earliest neighboring pixel from that time point to the next time point. If the red, green, and blue component change signs are all the same, the candidate passive diffusion pixel is considered to have the same change direction as its earliest neighbor pixel at that time point. A comparison is made point-by-point, counting the number of consecutive time points with the same change direction, and the total number of time points within the time window. The ratio of the number of consecutive time points with the same change direction to the total number of compared time points is calculated. If the ratio is greater than or equal to a preset threshold, the candidate passive diffusion pixel is determined to have the same change direction as its earliest neighbor pixel, and is marked as a passive diffusion pixel. If the ratio of the number of consecutive time points with the same change direction to the total number of compared time points is less than the preset threshold, it indicates that although the candidate passive diffusion pixel lags behind in the initial change time, its color change evolution process is not sufficiently consistent with that of the earliest neighbor pixel. This may be the result of independent acid production or interference from other factors, and the candidate passive diffusion pixel is marked as an autonomously acid-producing pixel.

[0032] The preset ratio threshold is set as follows: Before formal screening, prepare a set of control culture plates, inoculating only a single standard strain of lactic acid bacteria with known acid-producing ability on each plate. Acquire multiple time-series images under the same temperature and imaging conditions as the formal screening. Process the multiple time-series images of the control culture plates according to the steps described above to obtain sample sets of known passively diffusing pixels and known autonomously acid-producing pixels. For known passively diffusing pixel samples, record the ratio of the number of consecutive time-series points where the change direction is consistent between each known passively diffusing pixel and its earliest neighboring pixel to the total number of time-series points. Calculate the distribution of this ratio for all known passively diffusing pixels, and take the 5th percentile of this distribution as the preset ratio threshold. Alternatively, calculate the distribution of this ratio for all known autonomously acid-producing pixel samples, and take the 95th percentile of this ratio distribution as the preset ratio threshold. This preset ratio threshold can reliably distinguish between passively diffusing pixels and autonomously acid-producing pixels because the change direction of autonomously acid-producing pixels is basically unrelated to neighboring pixels, resulting in a lower ratio; passively diffusing pixels are dominated by the diffusion source, resulting in a higher ratio. Experiments have shown that, in the bromocresol purple indicator and MRS medium system, the preset ratio threshold typically falls between 0.7 and 0.9. Once this preset ratio threshold is calibrated, it can be used for subsequent image processing of petri dishes with the same type of medium without the need for repeated calibration.

[0033] Through the above process, each pixel is ultimately labeled as a passive diffusion pixel or an autonomous acid-producing pixel, providing an accurate pixel type basis for the S4 cluster candidate colony region and the S5 propagation streamline tracing.

[0034] In S4, autonomous acid-producing pixels are clustered according to their spatial location to form multiple candidate colony regions. Autonomous acid-producing pixels are identified from S3, and each pixel has corresponding spatial coordinates and a direction of change extracted from the color change curve. Autonomous acid-producing pixels exhibiting local density maxima in space are identified from all autonomous acid-producing pixels and used as cluster center pixels. The identification method for local density maxima is as follows: for each autonomous acid-producing pixel, a preset local density calculation neighborhood is defined centered on the spatial coordinates of the autonomous acid-producing pixel; the preset local density calculation neighborhood is a square window. The side length of the preset local density calculation neighborhood is set based on the expected diameter of candidate lactic acid bacteria colonies on a culture medium plate containing an indicator. The expected diameter of candidate lactic acid bacteria colonies typically occupies, for example, a range of 15 to 25 pixels. The side length of the preset local density calculation neighborhood is calculated as: L = round(D × 0.5); where L represents the side length of the preset local density calculation neighborhood in pixels; D represents the number of pixels corresponding to the expected diameter of the candidate lactic acid bacteria colony; and round represents the rounding function. The number of autonomously acid-producing pixels is counted within the preset local density calculation neighborhood; this number is the local density value of the autonomously acid-producing pixels. All autonomously acid-producing pixels are traversed to obtain the local density value of each autonomously acid-producing pixel. The local density value of each autonomous acid-producing pixel is compared one by one with the local density values ​​of all other autonomous acid-producing pixels in its preset local density calculation neighborhood. If the local density value of an autonomous acid-producing pixel is greater than the local density values ​​of all other autonomous acid-producing pixels in its preset local density calculation neighborhood, then the autonomous acid-producing pixel is identified as an autonomous acid-producing pixel with a spatially local density maxima and is used as a cluster center pixel. If two or more autonomous acid-producing pixels have equal local density maxima and are spatially adjacent, the autonomous acid-producing pixel whose spatial coordinates are closer to the geometric center of the preset local density calculation neighborhood is selected as the cluster center pixel, and the remaining autonomous acid-producing pixels are not selected as cluster center pixels to avoid generating multiple cluster center pixels for the same candidate lactic acid bacteria colony.

[0035] Starting from the central pixel of each cluster, autonomously acid-producing pixels that are spatially located within a preset neighborhood of the central pixel and exhibit the same direction of color change as the central pixel are grouped into the cluster corresponding to the central pixel. The preset neighborhood of the central pixel is determined based on the minimum possible distance between candidate lactic acid bacteria colonies, which refers to the minimum center-to-center distance between two independent candidate lactic acid bacteria colonies without edge overlap under conventional coating conditions. The inoculum concentration is approximately 10... 6 Up to 10 8When the number of colony-forming units per milliliter and the coating amount is, for example, 0.1 milliliters, the average spacing between candidate lactic acid bacteria colonies is approximately 50 to 100 pixels. The radius of the preset neighborhood range of the cluster center pixel is calculated as: R = S / 2; where R represents the radius of the preset neighborhood range of the cluster center pixel in pixels; and S represents the minimum possible spacing between candidate lactic acid bacteria colonies in pixels. The value of R is, for example, between 25 and 50 pixels, to ensure that the preset neighborhood range of the cluster center pixel can completely cover the distribution area of ​​the autonomous acid-producing pixels of the corresponding candidate lactic acid bacteria colony, without extending into the distribution area of ​​adjacent candidate lactic acid bacteria colonies. For each cluster center pixel, each autonomous acid-producing pixel within the preset neighborhood range of the cluster center pixel is traversed. For each autonomous acid-producing pixel within the preset neighborhood range of the cluster center pixel, the direction of change of the autonomous acid-producing pixel is compared with the direction of change of the cluster center pixel. The direction of change is represented by the sequence of changes recorded in S3. The direction of change of the autonomous acid-producing pixel at the first acquisition time point after the initial change is taken and compared with the direction of change of the cluster center pixel at the first acquisition time point after the initial change. The comparison method is as follows: the red, green, and blue component change symbols of the autonomous acid-producing pixel are compared item by item with the red, green, and blue component change symbols of the cluster center pixel, respectively. If the red component change symbol of the autonomous acid-producing pixel is the same as the red component change symbol of the cluster center pixel, the green component change symbol of the autonomous acid-producing pixel is the same as the green component change symbol of the cluster center pixel, and the blue component change symbol of the autonomous acid-producing pixel is the same as the blue component change symbol of the cluster center pixel, then the direction of change of the autonomous acid-producing pixel is determined to be in the same direction as the direction of change of the cluster center pixel, and the autonomous acid-producing pixel is assigned to the cluster corresponding to the cluster center pixel. If at least one of the red, green, or blue component change symbols of a self-generating acid pixel differs from the component change symbol corresponding to the cluster center pixel, then the change direction of the self-generating acid pixel is determined to be different from that of the cluster center pixel, and the self-generating acid pixel is not assigned to the cluster corresponding to the cluster center pixel. A self-generating acid pixel may be within the preset neighborhood of multiple cluster center pixels and satisfy the condition that its change direction is in the same direction as multiple cluster center pixels. In this case, the self-generating acid pixel is assigned to the cluster corresponding to the cluster center pixel with the closest spatial distance. The spatial distance is calculated as follows: Where d represents the spatial distance in pixels; x1 and y1 represent the x and y coordinates of the autonomous acid-producing pixel, respectively; and x2 and y2 represent the x and y coordinates of the cluster center pixel, respectively. After traversing the preset neighborhood of a cluster center pixel, the process continues with the next cluster center pixel until all cluster center pixels have been included in their preset neighborhoods. Autonomous acid-producing pixels belonging to the cluster center pixel itself are automatically assigned to the cluster corresponding to that cluster center pixel.

[0036] The spatial region covered by autonomous acid-producing pixels belonging to the same cluster is considered as a candidate colony region. For each cluster corresponding to a cluster center pixel, the cluster contains the cluster center pixel itself and several autonomous acid-producing pixels belonging to it, which are spatially distributed around the cluster center pixel. The union of the spatial coordinates of all autonomous acid-producing pixels within a cluster constitutes a connected spatial region, which is the candidate colony region. A candidate colony region corresponds to a candidate lactic acid bacteria colony. The boundary of the candidate colony region is determined by the spatial coordinates of the outermost autonomous acid-producing pixel within the cluster, and the bounding rectangle of the outermost autonomous acid-producing pixel is used as the spatial range of the candidate colony region. Multiple clusters corresponding to multiple cluster center pixels form multiple candidate colony regions, each of which is spatially independent and does not overlap with each other. If two candidate colony regions partially overlap, the autonomous acid-producing pixels in the overlapping area are reassigned to the cluster corresponding to the nearest cluster center pixel according to the principle of proximity, to ensure the uniqueness and spatial independence of each candidate colony region. Through the above process, multiple candidate colony regions corresponding one-to-one with candidate lactic acid bacteria colonies are obtained from the pixel set marked as autonomous acid-producing pixels. Each candidate colony region defines the spatial range of the autonomous acid-producing function of the corresponding candidate lactic acid bacteria colony, providing a spatial basis and attribution basis for the subsequent passive diffusion pixel allocation and the formation of independent acid-producing color development regions.

[0037] In S5, for each passively diffusing pixel, it is traced pixel by pixel along the direction of the pixel with the earliest color change start time in its spatial neighborhood, forming a propagation streamline until it reaches a candidate colony region, and the passively diffusing pixel is assigned to the candidate colony region. The passively diffusing pixels are obtained by S3, and the candidate colony regions are obtained by S4 clustering. Each candidate colony region corresponds to a cluster center pixel and a set of autonomous acid-producing pixels belonging to the corresponding cluster.

[0038] For each passively diffusing pixel, the neighboring pixel with the earliest color change start time within its spatial neighborhood is determined as the tracking direction pixel. The range of the spatial neighborhood is consistent with the range used in S3, i.e., a square window with a preset edge length centered on the passively diffusing pixel, for example, a preset edge length of 5 pixels. The spatial neighborhood includes the passively diffusing pixel and a set of pixels within a range of two pixels above, below, to the left, and to the right. All neighboring pixels within the spatial neighborhood are traversed to obtain the start change time of each neighboring pixel as determined in S3. The start change time is a specific minute value counted from the start of cultivation; the start change time of pixels that have not undergone significant color change is marked as null. The earliest non-null start change time is selected from the start change times of all neighboring pixels, and the neighboring pixel with the earliest non-null start change time is determined as the tracking direction pixel. The tracking direction pixel represents the upstream source direction of diffusion propagation, i.e., the color change wavefront propagates from the tracking direction pixel to the passively diffusing pixel. If multiple neighboring pixels in the spatial neighborhood have the same initial change time and are all the earliest, the one with the closest spatial distance to the passively diffusing pixel is selected as the tracking direction pixel. The spatial distance is calculated as the Euclidean distance between the spatial coordinates of the passively diffusing pixel and the spatial coordinates of its neighboring pixels. If the initial change time of the passively diffusing pixel itself is the earliest in the spatial neighborhood, i.e., there are no neighboring pixels with an initial change time earlier than the passively diffusing pixel, then the passively diffusing pixel is not included in the passively diffusing pixel allocation process, is re-marked as an autonomous acid-producing pixel, and is included in the candidate colony region formation process of S4.

[0039] Starting from the passively diffusing pixel, the tracking direction pixel is taken as the current tracking pixel. The neighboring pixel with the earliest color change start time within the spatial neighborhood of the current tracking pixel is repeatedly identified as the next tracking pixel, gradually advancing to form a propagation streamline. Specifically, the current tracking pixel is used as the new tracking starting point. A spatial neighborhood is defined centered on the current tracking pixel, with the same range as the previously defined spatial neighborhood. Within the spatial neighborhood of the current tracking pixel, all neighboring pixels except the previous tracking pixel are traversed. The start time of change for each neighboring pixel is obtained, and the neighboring pixel with the earliest start time of change is identified as the next tracking pixel. The previous tracking pixel refers to the pixel that was just identified as a tracking pixel before the current tracking pixel; excluding the previous tracking pixel prevents the propagation streamline from oscillating in place. The next tracking pixel is taken as the new current tracking pixel, and the process of determining the next tracking pixel is repeated. Each step advances and records one tracking pixel. The sequence of these sequentially connected tracking pixels is the propagation streamline. The propagation streamline records the complete path tracing back to the diffusion source from the initial passively diffusing pixel along a direction that gradually decreases in the color change start time. Each segment in the propagation flow connects a tracking pixel to an earlier tracking pixel at the next starting moment of change, and the whole flow points in the direction of the source of the color change.

[0040] When the propagation flow reaches a pixel that is autonomously acid-producing, its advancement is terminated. Autonomously acid-producing pixels, already marked in S3, are pixels whose color changes are driven by their own metabolic acid production rather than diffusion; therefore, they constitute the natural endpoint of the propagation flow. Once the propagation flow touches any autonomously acid-producing pixel, it indicates that the origin of the color change has been traced, and the flow terminates. If the propagation flow touches pixels at the image boundary or the edge of the culture plate during its advancement, and the boundary pixel is not an autonomously acid-producing pixel, the flow terminates. The passively diffusing pixel at the beginning of the flow is not assigned to any candidate colony area and is marked as an unassigned passively diffusing pixel, not used for subsequent screening. If the propagation flow loops during its advancement, i.e., the spatial coordinates of the next tracked pixel are the same as the spatial coordinates of a tracked pixel already recorded in the flow, it indicates that the tracking has entered a closed loop, and the assignment of the passively diffusing pixel is ambiguous. The propagation flow terminates, and the passively diffusing pixel is not assigned to any candidate colony area.

[0041] The passively diffusing pixels at the beginning of the propagation streamline are assigned to the candidate colony region where the autonomously acid-producing pixels at the end of the propagation streamline are located. The autonomously acid-producing pixels at the end of the propagation streamline are the autonomously acid-producing pixels reached by the propagation streamline. In S4, each autonomously acid-producing pixel has already been assigned to its corresponding candidate colony region, therefore the candidate colony region where the autonomously acid-producing pixel at the end of the propagation streamline is located is known. An attribution relationship is established between the passively diffusing pixels at the beginning of the propagation streamline and the candidate colony regions where the autonomously acid-producing pixels at the end of the propagation streamline are located, and the passively diffusing pixels are assigned to the candidate colony regions. A candidate colony region may receive multiple passively diffusing pixels from different propagation streamlines, indicating that the diffusion range of acidic metabolites produced by the candidate lactic acid bacteria colony corresponding to the candidate colony region covers multiple passively diffusing pixels. If the autonomously acid-producing pixels at the end of the propagation streamline have undergone overlapping region processing in S4, and have been assigned to a single candidate colony region in S4 according to the nearest-distance principle, the passively diffusing pixels are directly assigned to the single candidate colony region. The above tracking and allocation process is performed on all passively diffusing pixels. Each passively diffusing pixel is either assigned to a candidate colony region or marked as an unassigned passively diffusing pixel due to being tracked to a boundary or a cycle. Through the above process, all traceable passively diffusing pixels are correctly assigned to the candidate colony regions corresponding to the candidate lactic acid bacteria colonies that produce diffusion, providing a complete pixel attribution relationship for the merging into independent acid-producing colorimetric regions in S6.

[0042] In S6, the autonomous acid-producing pixels and passively diffusing pixels corresponding to each candidate colony region are merged to obtain independent acid-producing colorimetric regions. Thermotolerant lactic acid bacteria are screened based on the area change rate of the independent acid-producing colorimetric regions in multi-time series images. The candidate colony regions are obtained by clustering in S4. Each candidate colony region corresponds to a cluster center pixel and a set of autonomous acid-producing pixels belonging to the cluster corresponding to the cluster center pixel. Passively diffusing pixels are assigned to each candidate colony region in S5.

[0043] For each candidate colony region, all autonomously acid-producing pixels and all passively diffusing pixels belonging to that region are merged to form an independent acid-producing colorimetric region. For a given candidate colony region, the set of autonomously acid-producing pixels was determined in S4, and the set of passively diffusing pixels was assigned to the region in S5 through propagation streamline tracing. The merging operation involves taking the union of these two pixel sets to obtain all colorimetric pixels corresponding to the candidate colony region. Colorimetric pixels refer to pixels that change color due to acidification on a culture medium plate containing an indicator, including both autonomously acid-producing and passively diffusing pixels. The resulting independent acid-producing colorimetric region is a spatially complete connected region, reflecting the range of color changes in the culture medium caused by the candidate lactic acid bacteria colony's own metabolic acid production and the diffusion of its acidic metabolites. The boundary of the independent acid-producing colorimetric region is determined by the spatial coordinates of the outermost pixels in the merged pixel set, and can be obtained by calculating the convex hull or circumscribed rectangle. The independent acid-producing chromogenic region differs from the candidate colony region. The candidate colony region only contains autonomously acid-producing pixels, reflecting the colony's own location and acid-producing center. The independent acid-producing chromogenic region, however, extends to passively diffused pixels, reflecting the actual influence range of acidic metabolites in the culture medium. Each candidate colony region yields one independent acid-producing chromogenic region, and multiple candidate colony regions each yield multiple independent acid-producing chromogenic regions. These independent acid-producing chromogenic regions may partially overlap spatially due to overlapping diffusion halos, but this does not affect their independently calculated area and area change rate.

[0044] In each frame of the multi-time series images, the area of ​​the independent acid-producing chromogenic region is calculated, and the area change rate over time is obtained. The multi-time series images are obtained by S1 and contain a set of single-frame plate images arranged in chronological order of acquisition time. For an independent acid-producing chromogenic region of a candidate colony region, its area in each single-frame plate image needs to be calculated. The area is measured in pixels, i.e., the number of pixels contained in the independent acid-producing chromogenic region in the current frame is counted. For a time series, starting from the first frame of the multi-time series images, the pixels contained in the independent acid-producing chromogenic region are extracted frame by frame, and the number of pixels is counted as the area value corresponding to that frame. The area value is paired with the corresponding acquisition time point to form an area-time series. The area change rate reflects how quickly the area of ​​the independent acid-producing chromogenic region increases over time. The area change rate is calculated by linearly fitting the area-time series and obtaining the slope of the fitted line. The slope of the fitted line is the area change rate. The slope represents the number of pixels added to the independent acid-producing chromogenic region per unit time, with the time unit being minutes. Linear fitting can be performed using the least squares method. During the cultivation period, there is typically a phase where the area grows approximately linearly. Several consecutive acquisition time points corresponding to this phase are selected for fitting. The approximately linear growth phase is determined as follows: The area increment between adjacent frames is calculated point-by-point along the area-time series. When the area increment remains positive across multiple acquisition time points and the fluctuation range is within a preset range, the approximately linear growth phase is considered complete. When the area increment approaches zero or becomes negative, the approximately linear growth phase is considered over. The preset fluctuation range is set based on image noise and colony growth characteristics, taking the average area increment within the approximately linear growth phase and fluctuating by, for example, 20%. The area at each acquisition time point within the approximately linear growth phase is used as the dependent variable, and the acquisition time point as the independent variable. Least squares linear fitting is performed to obtain the linear regression coefficients, where the slope represents the area change rate. A larger area change rate indicates stronger acid production and diffusion capabilities of the candidate lactic acid bacteria colonies under high-temperature conditions, and more outstanding high-temperature resistance.

[0045] When the area change rate of the independent acid-producing chromogenic region of a candidate colony is greater than or equal to a preset area change rate threshold, the candidate lactic acid bacteria colony corresponding to that region is classified as thermotolerant lactic acid bacteria. The preset area change rate threshold is set as follows: Prepare a group of lactic acid bacteria strains with known thermotolerant characteristics as control strains. These control strains include thermotolerant lactic acid bacteria strains known to grow normally and produce acid at temperatures between 45℃ and 55℃, and thermosensitive lactic acid bacteria strains known to have inhibited growth at high temperatures. The control strains are cultured and imaged under the same conditions as the formal screening. Following methods S1 to S6, the area change rate of the independent acid-producing chromogenic region for each control strain is calculated. The area change rate distributions of the thermotolerant and thermosensitive lactic acid bacteria strains are statistically analyzed to find the area change rate value that reliably separates the two distributions. For example, the 5th percentile of the area change rate distribution of the thermotolerant lactic acid bacteria strains is taken as the preset area change rate threshold. This ensures that candidate lactic acid bacteria colonies with an area change rate greater than or equal to the preset area change rate threshold have a greater than 95% probability of being thermotolerant. Alternatively, the 95th percentile of the area change rate distribution of thermosensitive lactic acid bacteria strains can be used as the preset area change rate threshold. This ensures that candidate lactic acid bacteria colonies with an area change rate below the preset threshold have a greater than 95% probability of being thermosensitive. Once the preset area change rate threshold is set, it can be directly used for subsequent screening of the same type. For each candidate colony region, the area change rate of the independent acid-producing colorimetric region is compared with the preset area change rate threshold. If the area change rate is greater than or equal to the preset area change rate threshold, the candidate lactic acid bacteria colony is classified as thermoresistant lactic acid bacteria; if the area change rate is less than the preset area change rate threshold, the candidate lactic acid bacteria colony is classified as non-thermoresistant lactic acid bacteria. Finally, all candidate lactic acid bacteria colonies classified as thermoresistant lactic acid bacteria are screened out, completing the image-based screening process for thermoresistant lactic acid bacteria.

[0046] Example 2: Figure 2 A schematic diagram of a screening system for thermoresistant lactic acid bacteria according to the present invention is provided. The screening system for thermoresistant lactic acid bacteria includes: The image acquisition module is used to acquire multi-time-series images of multiple candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture conditions; The curve forming module is used to extract the color value of each pixel in multiple time-series images and form the color change curve of the pixel. The pixel marking module is used to compare the starting time and direction of the color change curve of each pixel with those of each pixel in its spatial neighborhood. If the starting time of the pixel's change lags behind that of a neighboring pixel and the two change directions are consistent, it is marked as a passively diffusing pixel; otherwise, it is marked as a self-generating acid pixel. The region formation module is used to cluster autonomous acid-producing pixels according to their spatial location to form multiple candidate colony regions; The pixel allocation module is used to track each passively diffused pixel pixel by pixel along the direction of the pixel with the earliest color change start time in its neighborhood, forming a propagation streamline until it reaches a candidate colony area, and then allocate the passively diffused pixel to the candidate colony area. The lactic acid screening module is used to merge the autonomous acid-producing pixels and passive diffusion pixels corresponding to each candidate colony region to obtain an independent acid-producing colorimetric region. The heat-resistant lactic acid bacteria are screened based on the area change rate of the independent acid-producing colorimetric region in multiple time-series images.

[0047] All calculations involved in the embodiments are performed using dimensionless numerical values, and the preset parameters and thresholds in the calculations can be set by those skilled in the art according to actual conditions.

[0048] This technical solution can be flexibly deployed, for example, as embedded software running on device hardware, or installed on personal computers or other smart terminals with user interfaces, thus adapting to various hardware environments and usage requirements.

[0049] The above solutions can be implemented in software, hardware, firmware, or a combination thereof. When implemented in software, they are presented, in whole or in part, as a computer program product, containing one or more computer instructions or programs. When these instructions or programs are loaded and executed on a computer, results are produced corresponding to the processes or functions of the embodiments of this application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one medium to another via wired or wireless means, such as from a website, server, or data center via wired means such as fiber optic cables, twisted-pair cables, or coaxial cables, or wireless means such as infrared or microwaves to another site. A computer-readable storage medium refers to any usable medium that a computer can access or a data storage device such as a server or data center that contains one or more usable media, including magnetic media such as floppy disks, hard disks, and magnetic tapes, optical media such as DVDs, and semiconductor media such as solid-state drives.

[0050] The specific working process of the system, device and module can be found in the method embodiment, and will not be repeated here.

[0051] The disclosed systems, devices, and methods can be implemented in other ways. The device embodiments are for illustrative purposes only, and the module division is only a logical division. In practice, different divisions can be implemented, such as merging or integrating multiple modules or components, or omitting some features. Coupling, direct coupling, or communication connections between the components can be achieved through interfaces, while indirect coupling or communication connections can take electrical, mechanical, or other forms.

[0052] The modules described as separate components may or may not be physically separated. The components shown as modules can be physical hardware or software, and can be deployed centrally or distributed across multiple network nodes. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0053] The functional modules in each embodiment can be integrated into one processing module, or they can exist independently, or two or more modules can be integrated into one.

[0054] If the functionality is implemented as a software module and used as an independent product, it can be stored in a computer-readable storage medium. Under this understanding, the substantial contribution of the technical solution of this application can be embodied in the form of a software product. The computer software product is stored in the storage medium and contains instructions to cause a computer device, such as a personal computer, server, or network device, to execute all or part of the steps of the methods in the embodiments of this application. The storage medium includes any medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory, random access memory, magnetic disk, or optical disk.

[0055] The above are merely specific embodiments of this application, and the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should fall within the scope of protection of this application.

Claims

1. A method for screening thermoresistant lactic acid bacteria, characterized in that, include: S1: Acquire multiple time-series images of candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture conditions; S2: For each pixel in multiple time-series images, extract the color value of the pixel in each time-series image to form the color change curve of the pixel; S3: For each pixel, compare the starting time and direction of the color change curve of the pixel with those of each pixel in its spatial neighborhood. If the starting time of the pixel's change lags behind that of a certain pixel in its neighborhood and the two change directions are consistent, then mark it as a passively diffused pixel; otherwise, mark it as a self-generating acid pixel. The consistent direction of change refers to the following: within the acquisition time range jointly covered by the pixel and its neighboring pixels, comparing the red component change sign, green component change sign, and blue component change sign of the pixel and its neighboring pixels from each time point to the next time point to see if all three are the same; if the proportion of the number of time points in consecutive time points with the same red component change sign, green component change sign, and blue component change sign to the total number of time points in common coverage reaches a preset proportion threshold, then the direction of change is determined to be consistent. S4: Cluster the autonomous acid-producing pixels according to their spatial location to form multiple candidate colony regions; S5: For each passively diffused pixel, trace it pixel by pixel along the direction of the pixel with the earliest color change start time in its neighborhood to form a propagation streamline until it reaches a candidate colony area, and assign the passively diffused pixel to the candidate colony area. S6: Merge the autonomous acid-producing pixels and passive diffusion pixels corresponding to each candidate colony region to obtain independent acid-producing color regions, and screen thermotolerant lactic acid bacteria based on the area change rate of the independent acid-producing color regions in multiple time-series images. The high-temperature culture conditions are 45℃~55℃, and the thermotolerant lactic acid bacteria are lactic acid bacteria that can grow and produce acid normally under culture conditions of 45℃~55℃.

2. The method for screening thermoresistant lactic acid bacteria according to claim 1, characterized in that, S1 includes: Place the culture medium containing the indicator, inoculated with multiple candidate lactic acid bacteria colonies, in a high-temperature incubation environment; Set a preset time interval and use an industrial camera to take pictures of culture medium plates at multiple acquisition time points to obtain single-frame plate images corresponding to each acquisition time point; The single-frame plate images corresponding to each acquisition time point are arranged in the order of acquisition time to form a multi-time sequence image.

3. The method for screening thermoresistant lactic acid bacteria according to claim 1, characterized in that, S2 includes: Determine the color value corresponding to the same pixel location in each time series image from multiple time series images; Arrange the color values ​​corresponding to the same pixel location in each time series image according to the order of acquisition time; The arranged color value sequence is used as the color change curve of the corresponding pixel.

4. The method for screening thermoresistant lactic acid bacteria according to claim 1, characterized in that, S3: Includes: For each pixel, identify the moment when the color value first breaks through the preset stable fluctuation range from the color change curve of the corresponding pixel, and take it as the starting moment of change; For each pixel, determine the neighboring pixel with the earliest start change time in the neighborhood. If the start change time of the pixel lags behind the start change time of the earliest neighboring pixel, and the lag time is within the preset neighborhood propagation lag range, then the pixel is pre-judged as a candidate passive diffusion pixel. For each candidate passive diffusion pixel, compare the color change curve of the pixel after the initial change time with the color change curve of the earliest neighboring pixel after the initial change time at multiple consecutive time points to see if the change direction is consistent. Candidate passive diffusion pixels with consistent change direction are marked as passive diffusion pixels, and candidate passive diffusion pixels with inconsistent change direction are marked as autonomous acid-producing pixels. Pixels that were not pre-judged as candidate passive diffusion pixels are marked as autonomous acid-producing pixels.

5. The method for screening thermoresistant lactic acid bacteria according to claim 4, characterized in that, Candidate passively diffused pixels with consistent change directions are marked as passively diffused pixels. This includes: comparing the color change curve of the candidate passively diffused pixel after the initial change time with the color change curve of the earliest neighboring pixel after the initial change time, and comparing the number of consecutive time points in which the color value moves along the same trajectory in the color space point by point. If the proportion of the number of consecutive time points to the total number of time points compared is greater than or equal to a preset proportion threshold, the change direction is determined to be consistent. If the proportion of the number of consecutive time points to the total number of time points compared is less than the preset proportion threshold, the change direction is determined to be inconsistent.

6. The method for screening thermoresistant lactic acid bacteria according to claim 1, characterized in that, S4 includes: Identify autonomous acid-producing pixels that exhibit local density maxima in space from all autonomous acid-producing pixels and use them as cluster center pixels; Starting from each cluster center pixel, autonomous acid-producing pixels that are spatially located within the preset neighborhood of the cluster center pixel and have the same direction of change on the color change curve as the cluster center pixel are assigned to the cluster corresponding to the cluster center pixel. The spatial region covered by autonomous acid-producing pixels belonging to the same cluster is considered as a candidate colony region.

7. The method for screening thermoresistant lactic acid bacteria according to claim 6, characterized in that, Autonomous acid-producing pixels that are spatially located within a preset neighborhood of the cluster center pixel and have the same direction of change on the color change curve as the cluster center pixel are assigned to the cluster corresponding to the cluster center pixel. This includes: comparing the direction of change of each autonomous acid-producing pixel within the preset neighborhood of the cluster center pixel with the direction of change of the cluster center pixel; if the direction of change of the autonomous acid-producing pixel is the same as the direction of change of the cluster center pixel, then the autonomous acid-producing pixel is assigned to the cluster corresponding to the cluster center pixel; otherwise, it is not assigned to the cluster corresponding to the cluster center pixel.

8. The method for screening thermoresistant lactic acid bacteria according to claim 1, characterized in that, S5 include: For each passively diffused pixel, the neighboring pixel whose color change started earliest in its spatial neighborhood is determined as the tracking direction pixel. Starting from the passive diffusion pixel, the tracking direction pixel is taken as the current tracking pixel. The earliest neighboring pixel in the spatial neighborhood of the current tracking pixel is repeatedly determined as the next tracking pixel, and the propagation streamline is gradually formed. When the propagation stream advances to the point where the currently tracked pixel is an autonomous acid-producing pixel, the advancement of the propagation stream is terminated; the passive diffusion pixel at the beginning of the propagation stream is assigned to the candidate colony region where the autonomous acid-producing pixel at the end is located.

9. The method for screening thermoresistant lactic acid bacteria according to claim 1, characterized in that, S6 include: For each candidate colony region, all autonomous acid-producing pixels and all passive diffusion pixels belonging to the candidate colony region are merged to form an independent acid-producing color development region. In each frame of the multi-time series images, the area of ​​the independent acid-producing colorimetric region is calculated, and the area change rate over time is obtained. When the area change rate of the independent acid-producing colorimetric region of the candidate colony area is greater than or equal to the preset area change rate threshold, the candidate lactic acid bacteria colony corresponding to the candidate colony area is determined to be thermostable lactic acid bacteria.

10. A screening system for thermoresistant lactic acid bacteria, used to implement the screening method for thermoresistant lactic acid bacteria according to any one of claims 1-9, characterized in that, include: The image acquisition module is used to acquire multi-time-series images of multiple candidate lactic acid bacteria colonies on a culture medium containing an indicator under high-temperature culture conditions; The curve forming module is used to extract the color value of each pixel in multiple time-series images and form the color change curve of the pixel. The pixel marking module is used to compare the starting time and direction of the color change curve of each pixel with those of each pixel in its spatial neighborhood. If the starting time of the pixel's change lags behind that of a neighboring pixel and the two change directions are consistent, it is marked as a passively diffusing pixel; otherwise, it is marked as a self-generating acid pixel. The region formation module is used to cluster autonomous acid-producing pixels according to their spatial location to form multiple candidate colony regions; The pixel allocation module is used to track each passively diffused pixel pixel by pixel along the direction of the pixel with the earliest color change start time in its neighborhood, forming a propagation streamline until it reaches a candidate colony area, and then allocate the passively diffused pixel to the candidate colony area. The lactic acid screening module is used to merge the autonomous acid-producing pixels and passive diffusion pixels corresponding to each candidate colony region to obtain an independent acid-producing colorimetric region. The heat-resistant lactic acid bacteria are screened based on the area change rate of the independent acid-producing colorimetric region in multiple time-series images.

Images