System and method for detecting intestinal flora

Abstract

The invention discloses an intestinal flora detection system and method, and relates to the technical field of intestinal flora detection.The method specifically comprises the following steps that fluorescence signal stability information of all areas in a fluorescence microscopic image is obtained in real time, analysis is conducted after obtaining, and the fluorescence signal stability information is obtained; respectively generating a fluorescence signal attenuation contrast coefficient and a fluorescence signal non-uniformity coefficient of each region; constructing a fluorescence signal bleaching evaluation model for the generated fluorescence signal attenuation contrast coefficient and the fluorescence signal non-uniformity coefficient of each region, and respectively generating a fluorescence signal bleaching evaluation coefficient of each region; and evaluating the influence degree of fluorescence signal attenuation caused by a fluorescence bleaching effect under a long-time exposure condition on the flora segmentation precision of each region in the fluorescence microscopic image. According to the method, the problem that the flora segmentation precision is reduced due to the fluorescent bleaching effect under long-time exposure is solved, and the accuracy and the reliability of flora detection in the fluorescent microscopic image are remarkably improved by dynamically adjusting the segmentation strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of intestinal flora detection, and in particular to an intestinal flora detection system and method. Background Art

[0002] Intestinal flora refers to various microbial communities existing in the human intestine, including bacteria, fungi, viruses and other microorganisms. These microorganisms live together with the host and maintain the ecological balance of the intestine. Intestinal flora not only plays an important role in the physiological functions of human digestion and absorption, immune system, metabolic regulation, but is also closely related to the occurrence and development of many diseases, such as obesity, diabetes, inflammatory bowel disease, irritable bowel syndrome, etc. In recent years, studies have shown that the imbalance of intestinal flora may be the root cause of many chronic diseases, so the detection of intestinal flora has become crucial. By accurately detecting the types and quantities of intestinal flora, it can help reveal the health status of individuals and provide a scientific basis for early diagnosis, precise treatment and personalized health management of diseases. Modern scientific and technological means, such as detection methods based on image analysis, can efficiently and accurately analyze the distribution and changes of intestinal flora, making intestinal flora detection not only an effective tool for disease diagnosis, but also promoting the further development of microbiome and precision medicine.

[0003] Existing intestinal flora detection technologies mainly rely on molecular biological methods, such as 16S rRNA gene sequencing, metagenomic sequencing, and metabolomics analysis. These methods usually extract DNA from intestinal samples and perform high-throughput sequencing to identify the types and abundance of microorganisms. However, traditional detection methods are usually time-consuming and require a large amount of experimental equipment and professional operations, and the results are complicated to analyze, making it difficult to quickly and accurately provide intuitive diagnostic information. In order to solve these problems, intestinal flora detection methods based on image processing technology have emerged. These methods achieve rapid detection of intestinal flora by efficiently processing and analyzing images of intestinal microorganisms under a microscope. Using computer vision and deep learning algorithms, image processing software can automatically identify and classify the morphological characteristics, distribution, and quantity changes of different microorganisms in intestinal samples, thereby quickly and accurately evaluating the status of intestinal flora. This method not only improves detection efficiency, but also helps researchers and doctors understand the structure and functional status of intestinal flora more intuitively through the visualization of image data, further promoting the development of personalized medicine and precision health management. Through image analysis technology combined with artificial intelligence algorithms, efficient screening can be carried out in large-scale samples to explore potential health risks and pathological changes, greatly improving the convenience and accuracy of intestinal flora testing.

[0004] The prior art has the following deficiencies: When using a fluorescence microscope to detect intestinal flora, if the sample is imaged under long-term exposure, a fluorescence bleaching effect will occur, resulting in the gradual attenuation or even complete loss of the fluorescence signals of some flora. This is because under long-term exposure to the excitation light, the fluorescent dye will undergo photochemical degradation, gradually weakening the fluorescence signals of the flora and finally disappearing. Since the attenuation of the fluorescence signals may be uneven in different regions of the fluorescence microscopic image, the flora in some regions may not be correctly identified due to too weak signals, resulting in a decrease in the contrast, blurred edges or complete absence of the flora in the image, affecting the overall flora distribution information. The existing intestinal flora detection technologies rely on fixed threshold segmentation algorithms or pixel intensity extraction methods to identify the flora. In the signal attenuation regions of the fluorescence microscopic image, it is impossible to evaluate which flora are not correctly identified due to fluorescence bleaching, resulting in an increase in image segmentation errors and a low flora count, affecting subsequent quantitative analysis and data accuracy. If the impact of fluorescence bleaching on the identification of flora in the fluorescence microscopic image cannot be accurately evaluated, it will lead to flora statistical errors, affecting the quantitative analysis of intestinal flora and further reducing the reliability of the detection results.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide a detection system and method for intestinal flora to solve the problems in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solution: A detection method for intestinal flora, specifically including the following steps: When using a fluorescence microscope to perform long-term exposure imaging on intestinal flora, determine the fluorescence microscopic image generated by the fluorescence bleaching effect and evenly divide it into several regions; Obtain the fluorescence signal stability information of each region in the fluorescence microscopic image in real time, and analyze it after obtaining, respectively generating the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient of each region; Construct a fluorescence signal bleaching evaluation model for the generated fluorescence signal attenuation comparison coefficient and fluorescence signal non-uniformity coefficient of each region, and respectively generate the fluorescence signal bleaching evaluation coefficient of each region; Based on the generated fluorescence signal bleaching evaluation coefficient of each region, conduct an evaluation analysis to evaluate the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the flora segmentation accuracy of each region in the fluorescence microscopic image, and divide each region in the fluorescence microscopic image into a signal stable region, a signal transition region and a signal bleaching region according to the evaluation results; According to the division results of each region in the fluorescence microscopic image, different image segmentation strategies are adopted for the signal stable region, the signal transition region, and the signal bleaching region respectively; Continuously monitor the change of the fluorescence signal in each region of the fluorescence microscopic image, and dynamically adjust the corresponding image segmentation strategy according to the real-time monitoring results.

[0008] Preferably, the fluorescence signal stability information of each region in the fluorescence microscopic image is obtained in real time, and after obtaining, it is analyzed to generate the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient of each region respectively, which specifically includes the following steps: Obtain the fluorescence signal stability information of each region in the fluorescence microscopic image in real time, and perform preprocessing after obtaining; Extract the distribution feature information and the dispersion information from the fluorescence signal stability information of each preprocessed region; Analyze the extracted distribution feature information and dispersion information to generate the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient of each region respectively.

[0009] Preferably, the acquisition logic of the fluorescence signal attenuation comparison coefficient of each region is as follows: Extract the distribution feature information from the fluorescence signal stability information of each preprocessed region, which specifically includes the maximum pixel value, the minimum pixel value, the average value and the standard deviation of all pixel values in each region of the fluorescence microscopic image, and are respectively calibrated as 、 、 and , represents the maximum pixel value in the th region of the fluorescence microscopic image, represents the minimum pixel value in the th region of the fluorescence microscopic image, represents the average value of all pixel values in the th region of the fluorescence microscopic image, represents the standard deviation of all pixel values in the th region of the fluorescence microscopic image, , is a positive integer; Calculate the fluorescence signal attenuation comparison coefficient of each region, and the specific calculation formula is as follows:

[0010] In the formula, is the fluorescence signal attenuation comparison coefficient of the th region.

[0011] Preferably, the acquisition logic of the fluorescence signal non-uniformity coefficient of each region is as follows: Extract the dispersion information from the fluorescence signal stability information of each region after preprocessing, specifically including the variance of all pixel values in each region of the fluorescence microscopic image and the ratio of the maximum pixel value to the minimum pixel value, and calibrate them respectively as and , represents the variance of all pixel values in the th region of the fluorescence microscopic image, represents the ratio of the maximum pixel value to the minimum pixel value in the th region of the fluorescence microscopic image, , is a positive integer; Calculate the fluorescence signal non-uniformity coefficient of each region. The specific calculation formula is as follows:

[0012] In the formula, is the fluorescence signal non-uniformity coefficient of the th region, represents the average value of all pixel values in the th region of the fluorescence microscopic image.

[0013] Preferably, for the generated fluorescence signal attenuation contrast coefficient and the fluorescence signal non-uniformity coefficient of each region, construct a fluorescence signal bleaching evaluation model, and generate the fluorescence signal bleaching evaluation coefficient of each region by weighted summation. The specific calculation formula is as follows:

[0014] In the formula, is the fluorescence signal bleaching evaluation coefficient of the th region, and are the non-zero weight coefficients of the fluorescence signal attenuation contrast coefficient and the fluorescence signal non-uniformity coefficient of each region respectively, and .

[0015] Preferably, determine the preset fluorescence signal bleaching evaluation coefficient threshold interval , and after determination, compare it with the generated fluorescence signal bleaching evaluation coefficient Perform comparison, evaluate the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in each region of the fluorescence microscopic image according to the comparison results, and divide each region in the fluorescence microscopic image into a signal stable region, a signal transition region, and a signal bleached region according to the evaluation results. The specific comparison analysis and division are as follows: If , the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in this region of the fluorescence microscopic image is a low influence degree, and this region is divided into a signal stable region; If , the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in this region of the fluorescence microscopic image is a medium influence degree, and this region is divided into a signal transition region; If , the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in this region of the fluorescence microscopic image is a high influence degree, and this region is divided into a signal bleached region.

[0016] Preferably, according to the division results of each region in the fluorescence microscopic image, different image segmentation strategies are adopted for the signal stable region, the signal transition region, and the signal bleached region respectively. Specifically: For the region divided into the signal stable region, the adopted image segmentation strategy is specifically: maintaining the current standard segmentation method; through contrast enhancement and edge sharpening, maintaining a high-resolution segmentation accuracy; For the region divided into the signal transition region, the adopted image segmentation strategy is specifically: using an adaptive threshold adjustment algorithm to dynamically adjust the segmentation threshold according to the signal changes within the region; introducing a multi-scale segmentation method to perform fine-grained adjustment for the signal attenuation region; enhancing the signal contrast through image enhancement technology to optimize the segmentation effect; For the region divided into the signal bleached region, the adopted image segmentation strategy is specifically: adopting a deep learning-assisted semantic segmentation algorithm to perform signal recovery and compensation for this region; combining image enhancement and filling technologies to restore the details of this region and reduce the segmentation error caused by signal loss; through multi-channel image fusion, combining the signal information of different channels to optimize the segmentation accuracy of this region.

[0017] Preferably, a detection system for intestinal flora includes an image division and analysis module, a signal attenuation analysis module, a bleaching evaluation model module, a signal evaluation and region division module, a dynamic segmentation strategy module, and a real-time monitoring and optimization module; An image division and analysis module, when using a fluorescence microscope to perform long-exposure imaging on intestinal flora, determines the fluorescence microscopic images generated by the fluorescence bleaching effect and evenly divides them into several regions; A signal attenuation analysis module, which obtains the fluorescence signal stability information of each region in the fluorescence microscopic image in real time and analyzes it after obtaining, and respectively generates the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient of each region; A bleaching evaluation model module, which constructs a fluorescence signal bleaching evaluation model for the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient of each generated region, and respectively generates the fluorescence signal bleaching evaluation coefficient of each region; A signal evaluation and region division module, which performs evaluation and analysis based on the fluorescence signal bleaching evaluation coefficients of each generated region, evaluates the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-exposure conditions on the segmentation accuracy of the flora in each region of the fluorescence microscopic image, and divides each region in the fluorescence microscopic image into a signal stable region, a signal transition region, and a signal bleaching region according to the evaluation results; A dynamic segmentation strategy module, according to the division results of each region in the fluorescence microscopic image, adopts different image segmentation strategies for the signal stable region, the signal transition region, and the signal bleaching region respectively; A real-time monitoring and optimization module, continuously monitors the fluorescence signal change conditions of each region in the fluorescence microscopic image, and dynamically adjusts the corresponding image segmentation strategy according to the real-time monitoring results.

[0018] In the above technical solution, the technical effects and advantages provided by the present invention are as follows: 1. The present invention effectively evaluates the signal attenuation and non-uniformity of each region in the fluorescence microscopic image by calculating the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient in real time. Through these key parameters, the system can dynamically adjust the image segmentation strategy without relying on the traditional fixed threshold segmentation method, which is particularly effective in regions with severe signal attenuation or bleaching effect. For the signal stable region, a standard segmentation method is adopted to ensure high-resolution segmentation accuracy; for the signal transition region, an adaptive threshold adjustment algorithm is used to adapt to the signal change; and for the signal bleaching region, a deep learning-assisted segmentation method is introduced to recover and compensate the signal in the bleaching region. This refined segmentation strategy can significantly improve the segmentation accuracy under long exposure, avoid edge blurring and mis-segmentation in traditional methods, and ensure the accurate identification of the flora.

[0019] 2. The present invention continuously monitors the fluorescence signal changes in each region of the fluorescence microscopy image and dynamically adjusts the image segmentation strategy according to real-time feedback. As the fluorescence signal decays, especially during long-term exposure, the signal changes may affect the segmentation accuracy. The real-time monitoring system can identify the changes in signal intensity and automatically adjust the segmentation algorithm to ensure that the segmentation strategy matches the current signal state. This dynamic optimization mechanism can not only adapt to the changes in experimental conditions in real time, avoiding the limitations of the fixed threshold method under different conditions, but also enhance the system's adaptability to problems such as signal bleaching and background changes, ensuring that it can still maintain high-efficiency and stable intestinal flora detection capabilities during long-term exposure or the experimental process.

[0020] 3. The present invention can accurately quantify the signal attenuation degree and non-uniformity of each region, and perform detailed regional division and segmentation optimization on the fluorescence microscopy image according to the evaluation results. Different image segmentation strategies are adopted for different types of regions (signal stable region, signal transition region, signal bleaching region), thereby effectively reducing the segmentation errors caused by signal attenuation and bleaching. This segmentation optimization based on quantitative evaluation not only significantly improves the accuracy of image segmentation, but also can accurately reflect the impact of the fluorescence bleaching effect of each region on the intestinal flora segmentation accuracy, providing a reliable basis for subsequent quantitative analysis, greatly reducing the statistical bias caused by segmentation errors, and ensuring the data accuracy and reliability in intestinal flora analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0022] Figure 1 It is a schematic flowchart of a detection system and method for intestinal flora of the present invention.

[0023] Figure 2 It is a schematic block diagram of a detection system and method for intestinal flora of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] Now, the exemplary embodiments will be described more comprehensively with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more complete and comprehensive, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0025] The present invention provides as Figure 1A detection method for gut microbiota specifically includes the following steps: When using a fluorescence microscope to perform long-exposure imaging on gut microbiota, determine the fluorescence microscopic images generated by the fluorescence bleaching effect and evenly divide them into several regions; To determine the fluorescence microscopic images generated by the fluorescence bleaching effect under long-exposure conditions, a fluorescence signal temporal analysis method can be used. Based on the fluorescence image frame sequence captured at different time points by the fluorescence microscope, extract and calculate the characteristics of the fluorescence signal changing with time. The specific implementation method includes the following steps: First, during the fluorescence microscope imaging process, collect the fluorescence images at different exposure times in the same field of view in real time, and extract the pixel gray value matrix and fluorescence signal intensity distribution data of each image. Then, apply pixel-level time series analysis to these temporal images, calculate the fluorescence signal decay curve of each pixel point or local region, and an exponential fitting or polynomial regression method can be used to calculate the decay rate of the fluorescence signal. Next, set a fluorescence signal decay threshold. When it is detected that the fluorescence signal intensity of a certain frame of image is lower than the set threshold in multiple regions and has decreased by more than a preset ratio compared to the fluorescence signal at the initial exposure, it can be determined that this image has been significantly affected by the fluorescence bleaching effect, and thus determine that this fluorescence microscopic image belongs to the bleaching-affected image. In addition, a deep learning model (such as LSTM or Transformer) can also be used for time series prediction to automatically determine which frame of image begins to show an obvious bleaching effect, so as to improve the detection accuracy and automation level.

[0026] After determining the fluorescence microscopic images generated by the fluorescence bleaching effect, it is necessary to evenly divide this image into several regions for subsequent fluorescence signal decay evaluation and dynamic segmentation optimization. It can be achieved by using a region segmentation method based on grid division or an adaptive region growing algorithm. The specific method includes the following steps: First, according to the resolution of the image, set a fixed grid size (such as 20×20 or 50×50 pixels), and divide the entire fluorescence microscopic image in a regular grid manner to ensure that all regions are of the same size, thereby ensuring the uniformity of the calculation results. Another method is an adaptive region division method based on local fluorescence signal characteristics, that is, first calculate the gradient map of the fluorescence signal distribution, and through the K-Means or DBSCAN clustering algorithm, adaptively divide the image into multiple regions according to the change trend of the fluorescence intensity, ensuring that regions with more severe fluorescence signal decay have higher spatial resolution, while regions with relatively stable fluorescence signals maintain lower spatial segmentation accuracy to improve the calculation efficiency and accuracy. No matter which method is used, ultimately it is necessary to ensure that each region can independently perform fluorescence signal decay analysis and adopt different optimization strategies for the bleaching influence degrees of different regions in the subsequent steps.

[0027] The main purpose of automatically identifying and regionally dividing fluorescence bleaching images using the above method is to accurately evaluate the impact of fluorescence bleaching effects on fluorescence microscopy images and improve the accuracy of bacterial colony segmentation through regional optimization. Since long-term exposure can cause the gradual attenuation of fluorescence signals, making the fluorescence signals in some regions become too weak or even completely lost, the traditional fixed-threshold segmentation method cannot correctly identify the bacterial colonies in these regions, ultimately affecting the counting and distribution analysis of bacterial colonies. Therefore, firstly, through the fluorescence signal time-series analysis method, it is possible to effectively distinguish which images are significantly affected by fluorescence bleaching, ensuring that subsequent calculations are only performed on the affected images, thereby improving the calculation efficiency. Secondly, by using the method based on grid division or adaptive region segmentation, the entire fluorescence microscopy image can be divided into multiple independent analysis regions, enabling different regions to adopt different optimization strategies according to the degree of bleaching influence. For example, in regions with relatively light signal attenuation, standard threshold segmentation can be directly applied, while in regions with severe signal attenuation, deep learning enhancement or adaptive threshold adjustment methods can be applied to ensure that the bacterial colonies in all regions can be accurately identified. This can not only improve the segmentability of bleached images but also reduce the impact of signal bleaching on the accuracy of bacterial colony detection, thereby enhancing the reliability and stability of intestinal bacterial colony detection.

[0028] Obtain the fluorescence signal stability information of each region in the fluorescence microscopy image in real time, and perform analysis after acquisition to generate the fluorescence signal attenuation comparison coefficient and fluorescence signal non-uniformity coefficient of each region respectively; In this embodiment, obtaining the fluorescence signal stability information of each region in the fluorescence microscopy image in real time and performing analysis after acquisition to generate the fluorescence signal attenuation comparison coefficient and fluorescence signal non-uniformity coefficient of each region respectively specifically includes the following steps: Obtain the fluorescence signal stability information of each region in the fluorescence microscopy image in real time and perform preprocessing after acquisition; Obtaining the fluorescence signal stability information of each region in the fluorescence microscopy image in real time can be achieved through image segmentation and pixel intensity analysis. Firstly, the image data captured by the fluorescence microscope can be transmitted and read in real time through image processing software, converting each frame of the image into a numerical matrix to obtain the fluorescence intensity value of each pixel. Then, for each region divided in the image (for example, regions based on grid division or adaptive partitioning), using a sliding window or region growing algorithm, the pixel intensity data of each region is extracted, and information such as the mean, standard deviation, and contrast change of the pixel intensity within the region is calculated. Specifically, by using filters in image processing libraries (such as OpenCV, ImageJ, etc.) for real-time image capture and intensity analysis, the stability information can be obtained by comparing the brightness values of each region in the image. During this process, each pixel value of the image can directly reflect the fluorescence intensity at that position, thereby generating a stability index.

[0029] The purpose of preprocessing is to remove noise and unnecessary background information in fluorescence microscopy images to improve the accuracy and reliability of subsequent analysis. The main steps of preprocessing include denoising, background removal, image enhancement, etc. First, random noise and interference generated during imaging are removed through denoising filters (such as Gaussian filtering, mean filtering, etc.) to ensure the clarity of the fluorescence signal. Then, background removal algorithms, such as Background Subtraction, are used to eliminate the influence of background light from the fluorescence signal in each region, avoiding the impact of background light on the final signal intensity calculation. Finally, to enhance the contrast of the signal, histogram equalization or adaptive contrast enhancement methods can be used to improve the distinguishability of weak fluorescence signals in the image, ensuring that even after long exposure, the weakened parts of the signal due to fluorescence bleaching can still be recognized. These preprocessing steps are implemented through image processing software, which can ensure that the quality of the processed image is high enough to provide a reliable basis for subsequent fluorescence signal stability analysis.

[0030] Extract distribution feature information and dispersion information from the fluorescence signal stability information of each preprocessed region; To extract distribution feature information and dispersion information from the fluorescence signal stability information of each preprocessed region, it can be achieved through regional statistical analysis first. Specifically, the distribution feature information can be extracted by calculating the maximum, minimum, mean, and median of the pixel intensities within each region. These values reflect the intensity range and central tendency of the fluorescence signal within the region, revealing the concentration or dispersion of the signal within the region. For the dispersion information, the standard deviation and variance of the pixel intensities within the region can be calculated to measure the volatility and non-uniformity of the signal, reflecting the stability of the signal in this region. In addition, pixel contrast can be used as a supplementary indicator of dispersion to further quantify the brightness differences between different regions in the image. When implementing this process, the regional statistical functions in image processing software (such as OpenCV or ImageJ) can be used to directly calculate these statistics for each region, and image segmentation techniques can be used to divide the image into different regions for independent analysis. In this way, not only the distribution characteristics of the signal within the region can be obtained, but also the uniformity and volatility of the signal can be quantitatively analyzed, providing a basis for subsequent segmentation optimization.

[0031] Analyze the extracted distribution feature information and dispersion information to generate the fluorescence signal attenuation contrast coefficient and fluorescence signal non-uniformity coefficient for each region respectively.

[0032] In this embodiment, the acquisition logic of the fluorescence signal attenuation contrast coefficient for each region is as follows: Extract the distribution feature information from the fluorescence signal stability information of each preprocessed region, specifically including the maximum pixel value, minimum pixel value, average value, and standard deviation of all pixel values within each region in the fluorescence microscopy image, and label them respectively as , , and , represents the maximum pixel value within the th region in the fluorescence microscopy image, represents the minimum pixel value within the th region in the fluorescence microscopy image, represents the average value of all pixel values within the th region in the fluorescence microscopy image, represents the standard deviation of all pixel values within the th region in the fluorescence microscopy image, , is a positive integer; To obtain the maximum pixel value, minimum pixel value, average value, and standard deviation of each region from the fluorescence microscopy image, pixel-level analysis of the image can be achieved through image processing software. First, load the fluorescence microscopy image through an image processing tool (such as OpenCV or ImageJ) and convert it into a numerical matrix, where the value of each pixel represents the fluorescence intensity at that point. Then, divide the image into multiple regions through an image segmentation method (such as grid division or region growing). For each region, the software can automatically calculate the maximum pixel value within the region (i.e., the intensity of the brightest point in the region), the minimum pixel value (i.e., the intensity of the darkest point), the average pixel value (i.e., the average of the intensities of all pixels within the region), and the standard deviation (i.e., the degree of dispersion of pixel intensities, reflecting the volatility of the signal). These calculations can be achieved through the statistical analysis tool in the image analysis function, which will automatically traverse the pixels of each region and extract these quantitative data, thus providing a basis for subsequent signal analysis and attenuation evaluation. These data can accurately quantify the distribution characteristics and volatility of the fluorescence signal and are the basis for subsequent analysis and calculation of key parameters such as the fluorescence decay contrast coefficient.

[0033] Calculate the fluorescence signal decay contrast coefficient for each region. The specific calculation formula is as follows:

[0034] In the formula, is the fluorescence signal decay contrast coefficient for the th region.

[0035] Calculate the fluorescence signal decay contrast coefficient for each region The purpose is to quantify the attenuation degree and its non-uniformity of the fluorescence signal in each region, so as to evaluate the influence of the fluorescence bleaching effect on the accuracy of microbial colony segmentation. First, What is calculated is the ratio of the difference between the maximum fluorescence intensity and the minimum fluorescence intensity in this region to the maximum intensity, which reflects the variation range of the signal in this region. The change in contrast (the difference between the maximum and minimum values) is a key factor in judging the signal attenuation degree, so this ratio is used to measure the severity of signal attenuation. Next, By calculating the standard deviation and the ratio with the average intensity and using logarithmic operation to enhance the sensitivity of signal change. The standard deviation reflects the volatility of the signal in this region, and the logarithmic operation further strengthens the influence of the region with larger attenuation, so as to more accurately capture the influence of the fluorescence bleaching effect on the signal. Finally, through normalization is carried out, so that the calculation result will not deviate from the standard due to the extreme values of the regional signal (such as very strong or very weak fluorescence signals), thus ensuring the stability and universality of the result. Combining these steps, it can effectively reflect the influence degree of the fluorescence bleaching effect in different regions, thus providing a reliable basis for the subsequent optimization of image segmentation.

[0036] The fluorescence signal attenuation contrast coefficient of the th region directly reflects the attenuation degree of the fluorescence signal and the non-uniformity of the signal in this region. Therefore, it has a close relationship with the influence degree of the fluorescence bleaching effect on the accuracy of microbial colony segmentation under long-time exposure conditions. Specifically, when the value is larger, it indicates that the fluorescence signal in this region attenuates more severely and non-uniformly, and the volatility and contrast difference of the signal are larger. This usually means that the signal attenuation in this region may cause the edges of the microbial colonies in the image to be blurred or completely lost, thus affecting the accuracy of the segmentation algorithm. On the contrary, when the value is smaller, it shows that the fluorescence signal in this region is relatively stable and uniform, and the attenuation effect is smaller, which helps to ensure the clear segmentation of the microbial colonies and higher segmentation accuracy. Therefore, the size of the value provides a quantitative index to measure the influence of the fluorescence bleaching effect on the image quality and subsequent segmentation tasks. The larger the value, the more significant the influence on the segmentation accuracy, and vice versa.

[0037] In this embodiment, the acquisition logic of the fluorescence signal non-uniformity coefficient of each region is as follows: Extract the dispersion information from the fluorescence signal stability information of each region after preprocessing, specifically including the variance of all pixel values in each region of the fluorescence microscopic image and the ratio of the maximum pixel value to the minimum pixel value, and calibrate them as and , represents the variance of all pixel values in the th region in the fluorescence microscopy image, represents the ratio of the maximum pixel value to the minimum pixel value in the th region in the fluorescence microscopy image, , where is a positive integer; To obtain the variance of all pixel values and the ratio of the maximum pixel value to the minimum pixel value in each region of the fluorescence microscopy image, pixel-level data extraction can be achieved through image processing software. First, the image is divided into multiple independent regions by image segmentation methods (such as threshold-based segmentation or region growing algorithms), which ensures that the pixel data in each region is analyzed independently. Then, for each region, statistical analysis tools in image processing software (such as OpenCV or ImageJ) are used to calculate the maximum and minimum pixel intensities in the region. These two values can be directly extracted from the intensity values of all pixels in the region. The maximum value reflects the strongest signal in the region, and the minimum value reflects the weakest point of the signal. Next, the variance of the pixel intensities in the region is calculated, which is the average of the squares of the differences between all pixel intensities in the region and the average value of the region. The larger the variance, the stronger the signal inhomogeneity in the region, which may be related to the attenuation or bleaching effect of the fluorescence signal. Finally, the ratio of the maximum value to the minimum value can reflect the extreme difference of the signal in the region, which is used to further quantify the signal inhomogeneity. All these statistical data can be achieved through the region statistics function in image processing software, ensuring that the obtained data is based on the quantitative information of the image itself, providing a basis for subsequent image analysis and optimization.

[0038] Calculate the fluorescence signal inhomogeneity coefficient for each region. The specific calculation formula is as follows:

[0039] In the formula, is the fluorescence signal inhomogeneity coefficient of the th region, represents the average value of all pixel values in the th region in the fluorescence microscopy image.

[0040] The purpose of calculating the fluorescence signal inhomogeneity coefficient for each region is to quantify the inhomogeneity of the fluorescence signal in the region, especially the change in signal distribution caused by the fluorescence bleaching effect during long-term exposure. First, The ratio of the maximum pixel value to the minimum pixel value reflects the extreme difference in the signals of this area and can measure whether there are large fluctuations or uneven distributions in the signals. If this value is large, it means that the signal intensity difference within this area is large, and the bleaching effect may cause some bacterial colonies to not be accurately identified. Secondly, The variance of all pixel values within the area measures the volatility of the signal intensity within this area. The larger the variance, the more unstable the signal, further emphasizing the degree of signal attenuation and unevenness. To better reflect the impact of these fluctuations on image segmentation, the variance is combined with the average pixel value of the area, and through logarithmic operations, the impact of areas with relatively large variances is amplified to ensure that the changing trend of the signal is fully considered during the calculation process. Finally, dividing by is to normalize the signal intensity of the area so that the calculation result is not affected by the absolute value of the signal intensity, thereby obtaining a standardized fluorescence signal non-uniformity coefficient that does not depend on the light intensity . Through this calculation method, the signal non-uniformity within the area can be accurately quantified, providing an effective basis for subsequent bleaching effect evaluation and image segmentation optimization.

[0041] The fluorescence signal non-uniformity coefficient of the nth area has a direct relationship with the degree of influence on the segmentation accuracy of bacterial colonies in each area of the fluorescence microscopic image caused by the fluorescence signal attenuation due to the fluorescence bleaching effect under long-term exposure conditions. Specifically, when the value is large, it indicates that the fluorescence signal distribution in this area is more uneven, the signal intensity difference is large, and there may be an obvious bleaching effect, that is, the fluorescence signal attenuation in some areas is more serious, resulting in a decrease in the image contrast, and the edges of some bacterial colonies are blurred or lost, thus affecting the accuracy of the segmentation algorithm. The larger the value, the more obvious the signal attenuation and non-uniformity in this area, and the greater the error in the segmentation accuracy. On the contrary, when the value is small, it indicates that the fluorescence signal in this area is relatively uniform, the signal attenuation is small, the image contrast is maintained well, and the segmentation result is more accurate. Therefore, the size of the value can effectively quantify the signal non-uniformity, and further predict the degree of influence of the fluorescence bleaching effect on the segmentation accuracy under long-term exposure, providing a key basis for subsequent image optimization and analysis.

[0042] Construct a fluorescence signal bleaching evaluation model for the fluorescence signal attenuation contrast coefficients and fluorescence signal non-uniformity coefficients of each generated area, and generate fluorescence signal bleaching evaluation coefficients for each area respectively; In this embodiment, for the fluorescence signal attenuation contrast coefficients generated for each area and the fluorescence signal non-uniformity coefficients Construct a fluorescence signal bleaching evaluation model, and generate fluorescence signal bleaching evaluation coefficients for each region through weighted summation. The specific calculation formula is as follows:

[0043] In the formula, is the fluorescence signal bleaching evaluation coefficient of the th region, and are respectively the fluorescence signal attenuation comparison coefficients of each region and the fluorescence signal non-uniformity coefficient non-zero weight coefficients, and .

[0044] To calculate the fluorescence signal bleaching evaluation coefficient , it is necessary to perform weighted summation on the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient . First, for each region, calculate and respectively. These two parameters have been obtained through the aforementioned image analysis method, and respectively reflect the amplitude of signal attenuation and the non-uniformity of the signal within the region. Then, these two indicators are combined through weighted summation to obtain the final fluorescence signal bleaching evaluation coefficient of the region. When calculating, and are weight coefficients, which represent the and relative contributions to , and satisfy the constraint of . and can be adjusted according to specific experimental designs or data analyses. For example, if you want to pay more attention to the impact of signal attenuation when the severe impact of fluorescence bleaching is greater, you can increase the value of ; if the impact of non-uniformity is greater, you can increase the value of . This weighting mechanism ensures that the model can flexibly adjust the influencing factors according to different experimental conditions or the characteristics of specific regions, so as to more accurately evaluate the impact of the fluorescence bleaching effect on the image segmentation accuracy under long-term exposure.

[0045] Based on the generated fluorescence signal bleaching evaluation coefficients of each region, conduct an evaluation and analysis to evaluate the degree of influence of the fluorescence signal attenuation caused by the fluorescence bleaching effect on the bacterial colony segmentation accuracy of each region in the fluorescence microscopy image under long-term exposure conditions, and divide each region in the fluorescence microscopy image into a signal stable region, a signal transition region, and a signal bleaching region according to the evaluation results; In this embodiment, a threshold interval of the pre-set fluorescence signal bleaching evaluation coefficient is determined , and after the determination, it is compared with the fluorescence signal bleaching evaluation coefficients of each generated region . According to the comparison results, the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the bacteria colony segmentation accuracy of each region in the fluorescence microscopic image is evaluated, and each region in the fluorescence microscopic image is divided into a signal stable region, a signal transition region, and a signal bleaching region according to the evaluation results. The specific comparison analysis and division are as follows: If , the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the bacteria colony segmentation accuracy of this region in the fluorescence microscopic image is a low influence degree, and this region is divided into a signal stable region; This situation means that the fluorescence signal attenuation in this region is relatively light, that is, the signal change in this region is small and basically remains stable. This indicates that the fluorescence bleaching effect in this region under long-term exposure conditions has a low influence, and the intensity and distribution of its fluorescence signal are relatively uniform, and it will not seriously affect the bacteria colony segmentation accuracy. During the image segmentation process, since the signal is relatively stable, the segmentation algorithm can more accurately identify and segment the bacteria colonies, and the segmentation accuracy is high. Therefore, such regions are divided into signal stable regions, no additional optimization measures are required, and the segmentation results are relatively reliable.

[0046] If , the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the bacteria colony segmentation accuracy of this region in the fluorescence microscopic image is a medium influence degree, and this region is divided into a signal transition region; This situation indicates that the fluorescence signal in this region is affected by medium-degree bleaching. This situation usually means that there is a certain degree of non-uniformity or fluctuation in the fluorescence signal attenuation in this region, the intensity of the signal is weak, but a certain contrast can still be maintained. Since the signal change has not reached the degree of complete failure, the influence degree on the segmentation accuracy is medium. During the image segmentation process, some mis-segmentations may occur, especially in the regions with weak signals. The segmentation algorithm may be interfered by the bleaching effect, resulting in blurred edges or incorrect recognition of some bacteria colonies. Therefore, these regions are divided into signal transition regions, and the image segmentation algorithm needs to be optimized or compensated to improve the segmentation accuracy.

[0047] If , the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the bacteria colony segmentation accuracy of this region in the fluorescence microscopic image is a high influence degree, and this region is divided into a signal bleaching region.

[0048] This situation indicates that the fluorescence signal attenuation in this area is severe, that is, the signal in this area is almost completely lost. Due to the severe bleaching of the fluorescence signal, the image contrast in this area decreases significantly, the signal becomes very weak, and it may even lead to the inability to identify the boundaries of the microbial population. In this case, the accuracy of the segmentation algorithm will be severely affected, which may result in the complete loss or missegmentation of the microbial population, thus greatly reducing the segmentation accuracy. This area is classified as a signal bleaching area, and special segmentation optimization strategies need to be adopted, such as deep learning enhanced segmentation or signal recovery algorithms, to cope with the segmentation errors caused by severe bleaching and improve the reliability of the results.

[0049] Determine the pre-set threshold interval of the fluorescence signal bleaching evaluation coefficient, which can be achieved by combining statistical analysis and experimental data analysis. First, through the analysis of the experimental data of a large number of fluorescence microscopic images, calculate the fluorescence signal bleaching evaluation coefficient of each area . In these data, calibrate the signal attenuation degree of different areas, and analyze the distribution characteristics under different exposure times and experimental conditions. Then, use statistical analysis methods, such as calculating the mean, standard deviation and distribution range of the data, to determine a suitable threshold interval. Specifically, the percentile can be selected as the threshold boundary to ensure that the of most signal stable areas is lower than the lower limit of this threshold interval, while the of severely bleached areas is higher than the upper limit. For example, set the minimum value of the threshold interval as the 10th percentile and the maximum value as the 90th percentile, so as to ensure that the threshold interval covers the vast majority of common situations and avoid the influence of outliers on the interval. After determining the threshold interval, real-time calculation and adjustment can be carried out through software tools (such as data analysis libraries in MATLAB or Python) to adapt to different experimental scenarios and image data sets. This method can automatically adjust the threshold interval flexibly according to different data sets and experimental conditions, providing an accurate basis for subsequent image segmentation optimization.

[0050] According to the division results of each area in the fluorescence microscopic image, different image segmentation strategies are adopted for the signal stable area, signal transition area and signal bleaching area respectively; In this embodiment, according to the division results of each area in the fluorescence microscopic image, different image segmentation strategies are adopted for the signal stable area, signal transition area and signal bleaching area respectively, specifically as follows: For the area classified as the signal stable area, the specific image segmentation strategy adopted is: maintain the current standard segmentation method, such as fixed threshold segmentation, to ensure accurate segmentation of the microbial population with stable signal intensity in this area and avoid additional computational overhead; through contrast enhancement and edge sharpening, maintain high-resolution segmentation accuracy; In the signal stable region, since the fluorescence signal in this region is relatively stable, the image segmentation strategy can be achieved by the fixed threshold segmentation method. This is because the signal variation in this region is small, and the signal intensity remains consistent under long-time exposure. Therefore, the fixed threshold segmentation can quickly and accurately segment the bacterial population without complex dynamic adjustment. In implementation, the pixel intensity of the image can be directly used through image processing software (such as OpenCV), and a fixed threshold is set. All parts with pixel values greater than this threshold are regarded as the foreground (bacterial population region), and the rest are the background. This method is simple and efficient, reducing the computational overhead and ensuring accurate segmentation in the high signal-to-noise ratio region. This method is chosen because the contrast in the signal stable region is high, and no additional enhancement or optimization is required. The fixed threshold segmentation can not only ensure the accuracy but also save computational resources.

[0051] For the regions classified as signal transition regions, the specific image segmentation strategy is as follows: Use the adaptive threshold adjustment algorithm to dynamically adjust the segmentation threshold according to the signal variation in the region; introduce the multi-scale segmentation method to perform fine-grained adjustment for the signal attenuation region to adapt to different signal intensity variations; enhance the signal contrast through image enhancement techniques (such as histogram equalization) to optimize the segmentation effect. In the signal transition region, due to the certain non-uniformity of signal attenuation, the adaptive threshold adjustment algorithm is an ideal segmentation strategy. Specifically, the threshold of each pixel can be dynamically adjusted in real time through image processing software, so that segmentation can be performed according to the illumination change and signal intensity fluctuation in the local region. During the implementation process, the local contrast or local mean can be used to determine the threshold of each region, and the entire region is scanned through the sliding window technique. The segmentation threshold is dynamically adjusted according to the mean or variance of the pixel intensity within the window. This method can effectively process the regions with uneven signal intensity changes, especially in the signal attenuation part, reducing the mis-segmentation caused by the traditional fixed threshold method. The reason for choosing this strategy is that although the signal in this region attenuates, it still has a certain local contrast, and dynamically adjusting the threshold can improve the segmentation accuracy and adapt to the non-uniformity of the signal.

[0052] For the regions classified as signal bleaching regions, the specific image segmentation strategy is as follows: Adopt the deep learning-assisted semantic segmentation algorithm, such as U-Net, to perform signal restoration and compensation for this region; combine image enhancement and filling techniques to restore the details of this region and reduce the segmentation error caused by signal loss; optimize the segmentation accuracy of this region through multi-channel image fusion, combining the signal information of different channels to ensure that the bacterial population in these regions is not missed or mis-segmented.

[0053] In the signal bleaching area, due to severe signal attenuation, traditional threshold segmentation methods may not be able to effectively distinguish bacteria colonies from the background. Therefore, using a deep learning-assisted semantic segmentation algorithm such as U-Net is the most appropriate solution. By using this advanced convolutional neural network (CNN) model, the software can automatically learn and identify complex patterns in the image, especially for those areas with weak signals and blurred edges. In implementation, the U-Net network is used for image training, and the model is trained based on the labeled dataset to enable it to identify and segment bacteria colonies in the bleached area. For image compensation and enhancement, image inpainting techniques can be combined, such as the image restoration module in a deep convolutional neural network, to compensate for the missing or blurred signal parts, thereby enhancing image details and improving segmentation accuracy. The reason for choosing this method is that the image quality in the signal bleaching area has severely deteriorated, and deep learning techniques are needed for complex signal restoration and segmentation to ensure that bacteria colonies in the bleached area are not missed or mis-segmented.

[0054] Continuously monitor the fluorescence signal changes in each area of the fluorescence microscopic image, and dynamically adjust the corresponding image segmentation strategy according to the real-time monitoring results to ensure the accuracy of bacteria colony recognition and improve the detection reliability of fluorescence microscopic images under long-term exposure conditions.

[0055] Continuously monitoring the fluorescence signal changes in each area of the fluorescence microscopic image can be achieved through real-time image processing and dynamic update algorithms. First, using continuous image capture and analysis techniques, the software can receive image data from the fluorescence microscope in real time and automatically calculate statistical quantities such as the fluorescence signal intensity, standard deviation, and mean of each area. During the image segmentation process, the software dynamically monitors and analyzes each area through a sliding window or region growing algorithm, and updates its fluorescence signal attenuation contrast coefficient and inhomogeneity coefficient in real time. By comparing the signal changes between the current image and the previous frame, the software can immediately detect which areas have significant signal attenuation or changes, and based on this, adjust the segmentation threshold or switch the segmentation strategy. For example, when it is found that the signal intensity in some areas has dropped significantly, the system can automatically select a deep learning segmentation algorithm that is more adaptable to the bleaching effect to cope with the problem of severe signal attenuation.

[0056] Real-time monitoring and dynamically adjusting the image segmentation strategy aims to improve the accuracy of microbiota recognition. Especially under long-term exposure conditions, the fluorescence bleaching effect often leads to changes in image quality. As the exposure time increases, the signal attenuation in some areas is relatively significant. Traditional static segmentation methods often cannot cope with the uneven signal changes, which may result in incorrect segmentation or missed detection. By dynamically adjusting the segmentation strategy and selecting the most suitable segmentation method according to real-time signal changes, signal attenuation can be automatically compensated to ensure segmentation accuracy. Adopting this real-time monitoring and adjustment strategy can handle signal changes in different regions, ensuring the reliability of detection under different image quality conditions, avoiding errors caused by long-term exposure, and improving the accuracy and stability of the entire fluorescence microscopy image analysis.

[0057] As Figure 2 shown, a detection system for intestinal microbiota includes an image division and analysis module, a signal attenuation analysis module, a bleaching evaluation model module, a signal evaluation and region division module, a dynamic segmentation strategy module, and a real-time monitoring and optimization module; The image division and analysis module, when using a fluorescence microscope to perform long-term exposure imaging on intestinal microbiota, determines the fluorescence microscopy image generated by the fluorescence bleaching effect and evenly divides it into several regions; The signal attenuation analysis module obtains the fluorescence signal stability information of each region in the fluorescence microscopy image in real time and analyzes it after acquisition, respectively generating the fluorescence signal attenuation comparison coefficient and the fluorescence signal non-uniformity coefficient of each region; The bleaching evaluation model module constructs a fluorescence signal bleaching evaluation model for the generated fluorescence signal attenuation comparison coefficient and fluorescence signal non-uniformity coefficient of each region, respectively generating the fluorescence signal bleaching evaluation coefficient of each region; The signal evaluation and region division module conducts evaluation and analysis based on the generated fluorescence signal bleaching evaluation coefficient of each region, evaluates the influence degree of the fluorescence signal attenuation caused by the fluorescence bleaching effect on the microbiota segmentation accuracy of each region in the fluorescence microscopy image under long-term exposure conditions, and divides each region in the fluorescence microscopy image into a signal stable region, a signal transition region, and a signal bleaching region according to the evaluation results; The dynamic segmentation strategy module adopts different image segmentation strategies for the signal stable region, the signal transition region, and the signal bleaching region respectively according to the division results of each region in the fluorescence microscopy image; The real-time monitoring and optimization module continuously monitors the fluorescence signal change situation of each region in the fluorescence microscopy image and dynamically adjusts the corresponding image segmentation strategy according to the real-time monitoring results.

[0058] The above formulas are all dimensionless and only take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0059] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0060] It should be understood that in various embodiments of the present application, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0061] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0062] In several embodiments provided by the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the above-described embodiments are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be in electrical, mechanical or other forms.

[0063] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0064] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0065] As mentioned above, the above are only the specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for detecting intestinal flora, characterized in that: The specific steps include: When using fluorescence microscopy for long-exposure imaging of intestinal flora, determine the fluorescence microscopic image produced by the fluorescence bleaching effect and divide it evenly into several areas; Acquire the fluorescence signal stability information of each area in the fluorescence microscopy image in real time, and analyze it after acquisition to generate the fluorescence signal attenuation contrast coefficient and the fluorescence signal inhomogeneity coefficient of each area respectively; A fluorescence signal bleaching evaluation model is constructed based on the generated fluorescence signal attenuation contrast coefficient and fluorescence signal inhomogeneity coefficient of each region, and a fluorescence signal bleaching evaluation coefficient of each region is generated respectively; Based on the generated fluorescence signal bleaching evaluation coefficients of each region, an evaluation analysis is performed to evaluate the influence of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in each region of the fluorescence microscopy image, and each region in the fluorescence microscopy image is divided into a signal stable region, a signal transition region, and a signal bleaching region according to the evaluation results; According to the division results of each area in the fluorescence microscopy image, different image segmentation strategies are adopted for the signal stable area, signal transition area and signal bleaching area. Continuously monitor the changes in fluorescence signals in each area of ​​the fluorescence microscopy image, and dynamically adjust the corresponding image segmentation strategy based on the real-time monitoring results.

2. The method for detecting intestinal flora according to claim 1, characterized in that: The fluorescence signal stability information of each area in the fluorescence microscopic image is acquired in real time, and analyzed after acquisition to generate the fluorescence signal attenuation contrast coefficient and the fluorescence signal non-uniformity coefficient of each area, which specifically includes the following steps: Acquire the fluorescence signal stability information of each area in the fluorescence microscopy image in real time and perform preprocessing after acquisition; Extracting distribution characteristic information and dispersion information from the pre-processed fluorescence signal stability information of each region; The extracted distribution characteristic information and dispersion information are analyzed to generate the fluorescence signal attenuation contrast coefficient and the fluorescence signal inhomogeneity coefficient of each area respectively.

3. A method for detecting intestinal flora according to claim 2, characterized in that: The logic for obtaining the fluorescence signal attenuation contrast coefficient of each region is as follows: The distribution characteristic information is extracted from the fluorescence signal stability information of each region after preprocessing, including the maximum pixel value, the minimum pixel value, and the mean and standard deviation of all pixel values ​​in each region of the fluorescence microscopy image, and calibrated as , , and , Indicates the fluorescence microscopy image The maximum pixel value in the region, Indicates the fluorescence microscopy image The minimum pixel value in the region, Indicates the fluorescence microscopy image The average value of all pixel values ​​in the region Indicates the fluorescence microscopy image The standard deviation of all pixel values ​​in the region is , is a positive integer; Calculate the fluorescence signal attenuation contrast coefficient of each area. The specific calculation formula is as follows: ； In the formula, For the The fluorescence signal attenuation contrast coefficient of each area.

4. A method for detecting intestinal flora according to claim 3, characterized in that: The logic for obtaining the fluorescence signal non-uniformity coefficient of each region is as follows: The dispersion information is extracted from the preprocessed fluorescence signal stability information of each region, including the variance of all pixel values ​​in each region of the fluorescence microscopy image and the ratio of the maximum pixel value to the minimum pixel value, and is calibrated as and , Indicates the fluorescence microscopy image The variance of all pixel values ​​in the region is Indicates the fluorescence microscopy image The ratio of the maximum pixel value to the minimum pixel value in the region. , is a positive integer; Calculate the fluorescence signal inhomogeneity coefficient of each area. The specific calculation formula is as follows: ； In the formula, For the The fluorescence signal inhomogeneity coefficient of the region is Indicates the fluorescence microscopy image The average value of all pixels in the region.

5. A method for detecting intestinal flora according to claim 4, characterized in that: The fluorescence signal attenuation contrast coefficient of each generated area and fluorescence signal inhomogeneity coefficient A fluorescence signal bleaching assessment model was constructed, and the fluorescence signal bleaching assessment coefficients of each region were generated by weighted summation. The specific calculation formula is as follows: ； In the formula, For the The fluorescence signal bleaching evaluation coefficient of the region is and are the fluorescence signal attenuation contrast coefficients of each region. and fluorescence signal inhomogeneity coefficient The non-zero weight coefficient of .

6. A method for detecting intestinal flora according to claim 5, characterized in that: Determine the pre-set fluorescence signal bleaching evaluation coefficient threshold interval , and after determining the fluorescence signal photobleaching evaluation coefficient for each region generated The comparison was performed, and the influence of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in each area of ​​the fluorescence microscopy image was evaluated according to the comparison results. According to the evaluation results, each area in the fluorescence microscopy image was divided into a signal stable area, a signal transition area, and a signal bleaching area. The specific comparison analysis and division are as follows: like , the influence of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in this area of ​​the fluorescence microscopy image is low, and this area is divided into a signal stable area; like , the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions has a moderate impact on the accuracy of bacterial colony segmentation in this area of ​​the fluorescence microscopy image, and this area is divided into a signal transition zone; like ,The attenuation of the fluorescence signal caused by the fluorescence bleaching effect under long exposure conditions has a high impact on the accuracy of bacterial colony segmentation in this area in the fluorescence microscopy image, and this area is classified as the signal bleaching area.

7. A method for detecting intestinal flora according to claim 6, characterized in that: According to the division results of each area in the fluorescence microscopy image, different image segmentation strategies are adopted for the signal stable area, signal transition area and signal bleaching area, specifically: For the area classified as the signal stable area, the image segmentation strategy adopted is as follows: maintain the current standard segmentation method; maintain high-resolution segmentation accuracy through contrast enhancement and edge sharpening; For the area classified as the signal transition zone, the image segmentation strategy adopted is as follows: using an adaptive threshold adjustment algorithm to dynamically adjust the segmentation threshold according to the change of the signal in the area; introducing a multi-scale segmentation method to make fine-grained adjustments for the signal attenuation area; using image enhancement technology to improve signal contrast and optimize segmentation effects; For the area classified as the signal bleaching area, the image segmentation strategy adopted is as follows: using a deep learning-assisted semantic segmentation algorithm to restore and compensate the signal in the area; combining image enhancement and filling techniques to restore the details of the area and reduce the segmentation error caused by signal loss; through multi-channel image fusion, combining the signal information of different channels, the segmentation accuracy of the area is optimized.

8. A system for detecting intestinal flora, used to implement the method for detecting intestinal flora as described in any one of claims 1 to 7, characterized in that: It includes image segmentation and analysis module, signal attenuation analysis module, bleaching assessment model module, signal assessment and area segmentation module, dynamic segmentation strategy module and real-time monitoring and optimization module; The image segmentation and analysis module determines the fluorescence microscopic image produced by the fluorescence bleaching effect when the intestinal flora is imaged for a long time using a fluorescence microscope, and evenly divides it into several areas; The signal attenuation analysis module acquires the fluorescence signal stability information of each area in the fluorescence microscopy image in real time, and analyzes it after acquisition to generate the fluorescence signal attenuation contrast coefficient and the fluorescence signal non-uniformity coefficient of each area respectively; The bleaching assessment model module constructs a fluorescence signal bleaching assessment model based on the generated fluorescence signal attenuation contrast coefficient and fluorescence signal inhomogeneity coefficient of each region, and generates fluorescence signal bleaching assessment coefficients of each region respectively; The signal evaluation and region division module performs evaluation and analysis based on the generated fluorescence signal bleaching evaluation coefficients of each region, evaluates the influence of the fluorescence signal attenuation caused by the fluorescence bleaching effect under long-term exposure conditions on the accuracy of bacterial colony segmentation in each region of the fluorescence microscopy image, and divides each region in the fluorescence microscopy image into a signal stable region, a signal transition region, and a signal bleaching region according to the evaluation results; The dynamic segmentation strategy module adopts different image segmentation strategies for the signal stable area, signal transition area and signal bleaching area according to the division results of each area in the fluorescence microscopy image; The real-time monitoring and optimization module continuously monitors the changes in fluorescence signals in various areas of the fluorescence microscopy image, and dynamically adjusts the corresponding image segmentation strategy based on the real-time monitoring results.

Images