Intelligent marine environmental pollution early warning method based on biomarkers

Abstract

The invention discloses an intelligent marine environmental pollution early warning method based on a biomarker, and belongs to the technical field of marine ecotoxicology, and the method comprises the steps: building an image quality control standard, obtaining a marine mussel histopathology image, building a data set according to the standard, carrying out the data marking and quality control, and carrying out the preprocessing. A model is constructed based on a YOLOv11 architecture, a multi-task segmentation function is realized by taking an image segmentation module as a core, and model training, verification and evaluation are carried out; different feature pixel points in the histopathology image are extracted through the trained model, then the shortest distance between the different feature pixel points is calculated through the Euclidean distance, and a result image containing distance marks is generated after averaging; and the marine environmental pollution early warning grade is divided by calculating the response index H. According to the method, the tissue pathology image is accurately analyzed by means of deep learning, identification and segmentation of different tissue structures of the marine mussels are realized, and an efficient technical means is provided for marine environment monitoring.

Description

Technical Field

[0001] The present invention belongs to the technical field of marine ecotoxicology, and specifically relates to an intelligent marine environmental pollution early warning method based on biomarkers. Background Art

[0002] Marine mussels ( Mytilus galloprovincialis ) As a typical filter-feeding economic shellfish, it has an important ecological function in the nearshore ecosystem. Through its characteristics of continuous water filtration and feeding, it can efficiently enrich toxic substances such as heavy metals and persistent organic pollutants in seawater, making it a key indicator species in marine ecotoxicology research. This species has a rich biomarker system, covering gene expression regulation and enzyme activity response at the molecular level, DNA damage at the cellular level, and pathological changes at the tissue and organ level. Among them, tissue pathological damage, as the key hub connecting molecular responses and organism phenotypes, provides irreplaceable morphological evidence for evaluating the biological effects of environmental pollution by intuitively displaying abnormalities in the structure of biological tissues.

[0003] Histopathology studies pathological changes in biological tissues and cells, assessing their adaptation and response to the environment. Histopathological damage, as a sensitive indicator, can serve as an early warning. When organisms exhibit pathological damage to their tissues and cells, while they may appear to survive at the individual level, they are less able to cope with environmental changes. Assessing their stress response to pollution by studying pathological changes can reveal the health of both the environment and the organisms themselves. Common histopathological analysis methods include qualitative, semi-quantitative, and quantitative analyses. Within the histopathological assessment technology landscape, quantitative indicators are a research priority due to their objectivity and reproducibility. Digestive gland luminal cell thickness and gill filament epithelial thickness, as stable quantitative parameters, have been shown to show significant correlations with pollutant exposure dose and serve as an important basis for constructing dose-effect models. Traditional methods rely on manual microscopic observation combined with image processing software to measure morphological parameters. These methods require manual tracing of tissue boundaries and geometric transformations to calculate average thickness, or direct measurement. However, to achieve statistical significance, a single test requires processing hundreds of pathological slide images, significantly increasing the workload.

[0004] More importantly, manual interpretation has significant technical limitations: on the one hand, differences in operator experience may lead to boundary identification errors; on the other hand, repetitive measurements of complex tissue structures can easily cause visual fatigue and result in data deviation.

[0005] Based on the deep learning intelligent analysis architecture combined with the image processing ability of CNN, the system realizes the end-to-end automatic analysis of the mussel tissue structure, including feature extraction, target segmentation and parameter calculation. Not only solves the long-existing efficiency bottleneck in the field of marine mussel histopathology quantitative analysis, but also lays a technical foundation for building an intelligent marine ecological toxicology detection platform, which has important application value for realizing the rapid evaluation and early warning of marine pollution biological effects. SUMMARY

[0006] In view of the deficiencies in the background art, the purpose of the present application is to provide an intelligent marine environmental pollution early warning method based on biomarkers. After image acquisition, quality control, labeling and preprocessing, a data set is established, and then a multi-task segmentation network is constructed using an improved YOLOv11 architecture. The average thickness is calculated by Euclidean distance, and the pollution is monitored by H index classification. The organization analysis is automated, the measurement accuracy and efficiency are improved, and technical support is provided for marine ecological toxicology research to help discover pollution in a timely manner and protect the marine environment.

[0007] The technical scheme adopted by the present application is as follows: An intelligent marine environmental pollution early warning method based on biomarkers, comprising the following steps: S1) Establish image quality control standards, acquire marine mussel histopathology images, construct a data set according to the established image quality control standards, then data labeling and quality control, and then format conversion and division to preprocess the data for model training; S2) Build a model based on YOLOv11 architecture, realize multi-task segmentation function with image segmentation module as the core, and perform model training, verification and evaluation to realize recognition and segmentation of different structures in marine mussel histopathology images; S3) Extract different feature pixel points in the histopathology image through the trained model, calculate the shortest distance between different feature pixel points using Euclidean distance, and generate a result image containing distance labels after averaging; S4) Divide the marine environmental pollution early warning level according to the value of response index H, and monitor the marine environmental pollution situation.

[0008] Preferably, in step S1), the image quality control standards are established, the marine mussel histopathology images are acquired, the data set is constructed according to the established image quality control standards, then the data is labeled and quality controlled, and then the data is preprocessed through format conversion and division. The specific process is as follows: (1) Data preparation Sa1) Establish image quality control standards, the standards are as follows: (1) Tissue morphological integrity: the tissue section is complete, without folding or damage; (2) Staining quality specification: uniform staining, clear contrast between cell nucleus and cytoplasm; (3) Optical imaging accuracy: accurate image focusing, no blurred area; Sa2) Image acquisition and data set construction: collect Mytilus galloprovincialis gill and digestive gland tissue samples and perform quality screening, use staining method to make histopathological slides, use high-resolution digital pathology scanning system to obtain histopathological images that meet the standards, and construct a data set that covers morphological variations of different individuals and tissues; Sa3) Data annotation: use professional annotation tool Labelme to annotate target structures in histopathological images, including: for gill filament tissue, select the outer wall and inner wall of gill filament cells as features and annotate; for digestive gland tissue, select the inner wall and outer wall of lumen cells as features and annotate in layers; all annotations are saved in JSON format files; Sa4) Data quality control: randomly select annotated histopathological images, calculate , the formula is where A and B are the sets of annotation results of two researchers, represents the intersection of the two annotation results, and represent the absolute values of the total number of two annotation results respectively; assuming the calculation result reaches 0.92±0.03, indicating that it meets the quality requirements; (2) Data preprocessing Sb1) Data format conversion: convert JSON format annotated information into YOLO format TXT file, each TXT file corresponds to one histopathological image, and histopathological images and corresponding TXT files are used as data set together; Sb2) Data set division: divide the data set into training set and test set to ensure the balance of data distribution.

[0009] Preferably, in step S2), the model is constructed based on YOLOv11 architecture, the image segmentation module is used as the core to realize multi-task segmentation function, and model training, verification and evaluation are performed, the specific steps are as follows: Model construction: YOLOv11 architecture recognition and segmentation engine is adopted; the model extracts image features through its efficient Backbone network, and fuses multi-scale features through the Neck part, and finally outputs binary segmentation masks of gill epithelial cells and digestive gland lumen tissues through its segmentation head, wherein the segmentation head part is designed as a double branch structure: branch A gill cell segmentation and branch B digestive gland lumen cell hierarchical segmentation; the role of YOLOv11 in this stage is to provide high-precision, pixel-level target contour for subsequent geometric parameter calculation; Model training: first, input the training set and validation set after data preparation and preprocessing into the model, set the parameters to train for 300 rounds, the number of samples per batch is 16, use the stochastic gradient descent (SGD) optimizer, update the model by iterative calculation of parameter gradient, the formula is as follows:

[0010] In the formula, is the model parameter vector after the first iteration, is the learning rate, is the loss function , the gradient of the parameter vector is the model parameter vector after the first iteration, is the learning rate, is the loss function , the gradient of the parameter vector is the model parameter vector after the first iteration; Then, use the Cosine annealing learning rate scheduling, set the maximum value of the learning rate to 0.01, and the minimum value of the learning rate to ; Finally, based on the index, when the model does not improve significantly after more than 100 rounds of training in the training process, stop training; Wherein, the is used as the key indicator of model performance, and the average value in the interval is calculated to obtain , the calculation formula is as follows:

[0011] In the formula, is the average precision when the confidence threshold is t%; The value range of t is 50-95, which represents the average precision at each confidence level within the confidence threshold range from 50% to 95%; Model verification and model evaluation: after training is completed, the trained model is verified on the validation set that did not participate in training to evaluate its performance on the data; Wherein, the calculation formula of the model accuracy is as follows:

[0012] In the formula, the number of samples correctly predicted refers to the number of samples successfully predicted by the model as the true label on the test data set; the total number of samples refers to the total number of samples in the test data set; The calculation formula of recall rate is as follows:

[0013] In the formula, the number of true cases is the number of actual positive samples successfully predicted by the model as positive cases; the number of false negative cases is the number of samples that are actually positive cases but are incorrectly predicted by the model as negative cases.

[0014] Preferably, in the step S3), different feature pixel points in the histopathology image are extracted by the trained model, that is, the outer wall pixel points and the inner wall pixel points of the gill filament cells are extracted, the outer wall pixel points and the inner wall pixel points of the duct lumen cells of the digestive gland tissue are extracted, and then the shortest distance between the outer wall pixel points and the inner wall pixel points, that is, the thickness, is calculated by using the Euclidean distance, and the average thickness is calculated to generate a result image containing thickness labels. The specific steps are as follows: Sc1) Digestive gland lumen cell and gill filament cell segmentation: using the above trained model to segment the digestive gland lumen cell and gill filament cell in the histopathology image, and extracting the pixel points constituting the outer wall and the inner wall of the digestive gland lumen cell or the gill filament cell, and using the pixel points to define the internal structure and the external boundary of the cell, that is, the outer wall pixel points and the inner wall pixel points; Sc2) Euclidean distance calculation: Euclidean distance d is used for calculation. For the inner wall pixel point and the outer wall pixel point, the inner wall pixel point is assumed to be and the outer wall pixel point is The distance between each inner wall and outer wall is calculated, and the shortest distance between the inner wall pixel point and the outer wall pixel point is taken, and the calculation formula is:

[0015] Wherein, d is the straight line distance between the inner wall pixel point and the outer wall pixel point, is the coordinate difference value of the inner wall pixel point and the outer wall pixel point in the horizontal direction, is the coordinate difference value of the inner wall pixel point and the outer wall pixel point in the vertical direction; The minimum distance of the multiple Euclidean distances from the same inner wall pixel point to different outer wall pixel points is selected, that is, the thickness value, and the calculation formula is: Sc3) Thickness average value calculation: the thickness values of all positions are averaged, and the calculation formula is: Wherein, N represents the number of thickness values of all positions, represents the thickness value of the th position, and the result is labeled on the segmented cell, and the number of recognized cells is recorded at the same time; ​Sc4) Statistical result generation: generate the result image of batch labeling, and show the quantitative results of cell thickness analysis.

[0016] Preferably, in the step S4), the response index H of cell thickness is calculated, the marine environmental pollution warning level is divided according to the value of the response index H, and the marine environmental pollution is monitored, and the specific steps are as follows: By comparing the cell thickness of the marine mussel tissue in the test sea area and the control sea area, the response index H of cell thickness is calculated, and the calculation formula is as follows:

[0017] Wherein, T1 represents the thickness of the tissue in the test sea area, and T0 represents the thickness of the tissue in the control sea area; According to the value of H, four warning levels are divided: H≤10% represents warning A level, 10%<H≤20% represents warning B level, 20%<H≤30% represents warning C level, and 30%<H represents warning D level.

[0018] Compared with the prior art, the present application proposes an intelligent marine environmental pollution warning method based on biomarkers, and the advantages of the method are: 1. The present application establishes an intelligent recognition and quantitative analysis method for marine mussel tissue pathology image based on deep learning through systematic experimental verification, breaks through the limitation of traditional dependence on artificial experience, and provides a scientific and standardized technical path for tissue pathology analysis; 2. The present application constructs a deep learning model suitable for marine mussel tissue pathology image analysis, realizes the automatic and accurate measurement of the thickness of digestive gland tube cavity and gill filament epithelium relying on the improved YOLOv11 double-branch segmentation network, and avoids the measurement error of artificial measurement; 3. The combination of deep learning model and traditional artificial detection method verifies the labeling consistency through Dice coefficient (0.92±0.03), and the mAP 50-95 Index guarantees the performance of the model, and significantly improves the stability and reliability of the quantitative analysis of marine mussel tissue pathology; The intelligent recognition method of tissue pathology image based on deep learning can greatly improve the analysis throughput and detection efficiency compared with artificial detection, can quickly generate thickness labeled image and pollution warning result, and provides efficient technical support for marine environment monitoring. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The flow framework diagram of the present application; Figure 2The flow chart of the preparation and image acquisition process of the marine mussel histopathology section of the application; A, marine mussel tissue extraction and fixation; B, dehydration and transparency; C, paraffin embedding; D, tissue section; E, spreading and baking; F, hematoxylin and eosylin staining; G, microscopic observation and image acquisition; Figure 3 The flow chart of batch cell thickness identification; H, histopathology image of digestive gland; I, use YOLOv11 to segment the cells in the original image, P0 represents the outer wall, and P1 represents the inner wall; J, calculate the Euclidean distance between each outer wall and inner wall point, take the inner wall point closest to each outer wall point as the cell thickness at the current outer wall position, and take the average value as the average cell thickness; K, output the result image, the blue area represents the cell area, and the label is the cell category and the current cell thickness; Figure 4 The control group image of the histopathological changes induced by sulfamethoxazole. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the application will be further described in detail below with reference to the drawings in the embodiments of the application. It should be explained that the described embodiments are only a part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0021] In order to make the application purposes, technical solutions and advantages of the application more clear, the embodiments of the application will be further described in detail below with reference to the drawings of the specification: In order to more clearly understand the above-mentioned purposes, features and advantages of the application, the advantages of the application will be further explained by comparing the embodiments with reference to the drawings and specific embodiments.

[0022] The application proposes an intelligent marine environmental pollution early warning method based on biomarkers, as shown in the accompanying Figure 1 The steps of the method are described in detail as follows: S1) Establish image quality control standards, obtain marine mussel histopathology images, construct a data set according to the established image quality control standards, then data annotation and quality control, and then format conversion and division to preprocess the data, and prepare for model training; Specifically, in the step S1), the image quality control standards are established, the marine mussel histopathology images are obtained, the data set is constructed according to the established image quality control standards, then the data is annotated and quality controlled, and then the data is preprocessed through format conversion and division, and the specific process is as follows: (1) Data preparation Sa1) Establish image quality control standards, the standards are as follows: (1) tissue morphology integrity: (1) Tissue slice integrity: no folding or damage; (2) Staining quality: uniform staining, clear contrast between nucleus and cytoplasm, ensuring accurate identification of cell morphological features; (3) Optical imaging accuracy: accurate image focus, no blurred areas; to ensure that the image quality meets the requirements of deep learning model training; Sa2) Image acquisition and dataset construction: Collect Mytilus galloprovincialis gill and digestive gland tissue samples and perform quality screening. Use H&E staining to prepare histopathological slides. Use a high-resolution digital pathology scanning system to obtain histopathological images (including tissue microstructure, cell distribution, and staining characteristics). As shown in FIG. 1, through standardized staining and digital scanning, the transformation of Mytilus galloprovincialis pathological samples from “physical slices” to “digital assets” is achieved, providing a standardized and reproducible basic data carrier for subsequent use; Figure 2 Sa2) Image acquisition and dataset construction: Collect Mytilus galloprovincialis gill and digestive gland tissue samples and perform quality screening. Use H&E staining to prepare histopathological slides. Use a high-resolution digital pathology scanning system to obtain histopathological images (including tissue microstructure, cell distribution, and staining characteristics). As shown in FIG. 1, through standardized staining and digital scanning, the transformation of Mytilus galloprovincialis pathological samples from “physical slices” to “digital assets” is achieved, providing a standardized and reproducible basic data carrier for subsequent use; Based on the above standards, 1000 histopathological images that meet the standards are obtained by quality screening of 200 samples, and a dataset is constructed, which covers morphological variations of different individuals and tissues; Sa3) Data annotation: completed by two marine biological pathologists with more than 5 years of experience; use professional annotation tool Labelme to accurately annotate target structures in histopathological images, including: for gill filament tissue, select the outer wall and inner wall of gill filament cells as features and label; for digestive gland tissue, select the inner wall and outer wall of lumen cells as features and label in layers; all annotations are saved in JSON format files, including polygon vertex coordinates and class labels; Sa4) Data quality control: randomly select 100 labeled histopathological images, calculate , the formula is where A and B are the sets of annotation results of the two researchers, represents the intersection of the two annotation results, and represent the absolute values of the total number of the two annotation results; assuming the calculation result reaches 0.92±0.03 (mean±standard deviation), indicating that it meets the quality requirements of histopathological images; (2) Data preprocessing Sb1) Data format conversion: convert the polygon vertex information annotated in JSON format to TXT files in YOLO format, with each TXT file corresponding to a histopathological image. The file contains the number of target categories and normalized bounding box coordinates. Histopathological images and corresponding TXT files are used as a dataset; Sb2) Dataset division: divide the dataset into training set (80%) and test set (20%) to ensure balanced data distribution.

[0023] S2) building a model based on YOLOv11 architecture, realizing multi-task segmentation function with image segmentation module as the core, and performing model training, verification and evaluation to realize recognition and segmentation of different structures in marine mussel histopathology images; Specifically, in the step S2), a model is built based on YOLOv11 architecture, multi-task segmentation function is realized with an image segmentation module as the core, and model training, verification and evaluation are performed, and the specific steps are as follows: Model building: YOLOv11 architecture recognition and segmentation engine is adopted; the model uses its efficient Backbone network as the feature extractor of the model, gradually extracts features from the input image, and finally outputs the binary segmentation mask of the gill filament epithelial cells and the digestive gland lumen tissue through the multi-scale feature fusion of the Neck part and the segmentation head, wherein the segmentation head part is designed as a double-branch structure: branch A gill filament cell segmentation and branch B digestive gland lumen cell hierarchical segmentation, this double-branch design enables the model to simultaneously process two different but related segmentation tasks, breaking the limitations of single-task segmentation and improving the analysis capability of complex tissue images (such as marine mussel histopathology images); the role of YOLOv11 in this stage is to provide high-precision, pixel-level target contours for subsequent geometric parameter calculation; Model training: first, input the training set and validation set after data preparation and preprocessing into the model; the parameter setting is to train for 300 rounds, and the number of samples per batch is 16; the number of rounds determines the number of learning times of the model on the data set, and the batch size affects the memory utilization and parameter update frequency; the stochastic gradient descent (SGD) optimizer is adopted, and the model is updated by iterative calculation of parameter gradient, and the formula is as follows:

[0024] In the formula, is the model parameter vector after the th iteration, is the learning rate, is the loss function , the gradient of the parameter vector , is the model parameter vector after the th iteration; Then, the cosine annealing learning rate scheduling is adopted, the cosine annealing is to make the learning rate change periodically according to the cosine function, the learning rate converges faster in the early stage, and the learning rate is small in the later stage to fine-tune the parameters, avoiding overfitting or slow convergence, wherein the maximum value of the learning rate is set to 0.01, and the minimum value of the learning rate is set to ; Finally, based on the The indicator is used to stop the training when the model does not show significant improvement after more than 100 rounds of training during the training process, which can prevent the model from overfitting, save computing resources, and further realize the recognition and segmentation of the characteristics of digestive gland lumen cells and gill filament cells in the marine mussel histopathology image. wherein, (mean Average Precision between 50% to 95%) is a key indicator of model performance commonly used in target detection tasks, which is used to measure the average precision of the model at different confidence thresholds; the average value in the interval is obtained by calculating , the calculation formula is as follows:

[0025] In the formula, is the average precision when the confidence threshold is t%; The value range of mAP Model verification and model evaluation: after the training is completed, the trained model is verified on the validation set that does not participate in the training to evaluate its performance on the data; by analyzing the prediction results of the model on the validation set, the accuracy of the recognition and segmentation of the digestive gland lumen cells and gill filament cells is evaluated; this step is crucial for confirming the generalization ability and accuracy of the model, and provides a basis for further optimization and application; The accuracy of the trained model reaches 88.8%, the recall rate reaches 93.1%, and the mAP 50 (B) is 0.869, and the mAP 50-95 (B) is 0.63, which further confirms the excellent performance of the model in the pixel-level segmentation task; wherein, the calculation formula of the model accuracy is as follows:

[0026] In the formula, the number of correctly predicted samples refers to the number of samples successfully predicted by the model on the test data set; the total number of samples is the total number of samples in the test data set; The calculation formula of the recall rate is as follows:

[0027] In the formula, the number of true positives is the number of actual positive samples successfully predicted by the model as positive; the number of false negatives is the number of samples that are actually positive but are incorrectly predicted by the model as negative.

[0028] ​S3) Extract different feature pixel points in the histopathology image by the trained model, and then calculate the shortest distance between different feature pixel points by Euclidean distance to generate a result image containing distance labels after averaging; Specifically, in the step S3), different feature pixel points in the histopathology image are extracted by the trained model, that is, the outer wall pixel points and the inner wall pixel points of the gill filament cells are extracted, the outer wall pixel points and the inner wall pixel points of the gland duct lumen cells are extracted, and then the shortest distance between the outer wall pixel points and the inner wall pixel points, that is, the thickness, is calculated by Euclidean distance, and a result image containing thickness labels is generated after averaging the thickness, as shown in the accompanying drawings. The specific steps are as follows: Figure 3 Sc1) Digestive gland lumen cell and gill filament cell segmentation: the above trained model is used to accurately segment the digestive gland lumen and gill filament cells in the histopathology image, and the pixel points constituting the outer wall and the inner wall of the digestive gland lumen or the gill filament cells are extracted, and the internal structure and the external boundary of the cells, that is, the outer wall pixel points and the inner wall pixel points, are defined by the pixel points; Sc2) Euclidean distance measurement and calculation: based on the obtained inner wall pixel points and outer wall pixel points, the distance between each inner wall pixel point and outer wall pixel point is calculated, and the shortest distance between the inner wall pixel point and the outer wall pixel point is taken, that is, the cell thickness at the position of the inner wall pixel point and the outer wall pixel point; The Euclidean distance d is used for calculation. For the inner wall pixel point and the outer wall pixel point, the inner wall pixel point is assumed to be and the outer wall pixel point is The distance between each inner wall and outer wall is calculated, and the shortest distance between the outer wall pixel point and the inner wall pixel point is taken, and the calculation formula is:

[0029] wherein d is the straight line distance between the outer wall pixel point and the inner wall pixel point, is the coordinate difference value of the outer wall pixel point and the inner wall pixel point in the horizontal direction, is the coordinate difference value of the outer wall pixel point and the inner wall pixel point in the vertical direction; The multiple Euclidean distances of the same inner wall pixel point to different outer wall pixel points are selected, and the minimum distance is taken as the thickness value, and the calculation formula is: Sc3) Thickness average value calculation: the thickness values of all positions are averaged, and the calculation formula is: wherein N represents the number of thickness values of all positions, represents the thickness value of the th position, the result is labeled on the segmented cells, and the number of recognized cells is recorded at the same time; ​​S4) Statistical result generation: generate the result image of batch annotation, and show the quantitative result of cell thickness analysis.

[0030] S4) By calculating the response index H, the marine environmental pollution early warning level is divided according to the value, and the marine environmental pollution is monitored.

[0031] Specifically, in step S4), by calculating the response index H of the cell thickness, the marine environmental pollution early warning level is divided according to the value, and the marine environmental pollution is monitored, and the specific steps are as follows: By comparing the tissue thickness of the test sea area (may be contaminated) and the control sea area (supposed to be uncontaminated and stable environment), the tissue thickness response index H is calculated, and the calculation formula is as follows:

[0032] Wherein, T1 represents the tissue thickness of the test sea area, and T0 represents the tissue thickness of the control sea area; According to the value of H, it is divided into four early warning levels: H≤10% represents early warning A level, 10%<H≤20% represents early warning B level, 20%<H≤30% represents early warning C level, and 30%<H represents early warning D level; The pollution degree of the related sea area is characterized by calculating the tissue thickness response index. Under normal circumstances, the tissue thickness of the marine mussel remains in a relatively stable range, and thickening (such as tissue swelling, inflammatory reaction, etc.) and thinning (such as tissue necrosis, abnormal nutrition, etc.) all suggest that the marine mussel may be under stress of pollutants; if the tissue thickness response index of the related sea area is greater than 10%, it proves that there is a stress source in the sea area, and the environmental protection should be strengthened.

[0033] The specific experiments of the present application are described in detail in combination with the technical solutions and the drawings as follows: (1) Comparative Example 1 (recognition accuracy) ① Use the histopathological images of marine mussels accumulated by the research team, randomly select 20 histopathological images of digestive gland and gill tissues from them, and use them to form a test data set; ② The digestive gland lumen and gill epithelial cells in the test data set are labeled by using the method, artificial recognition (unfamiliar) and artificial recognition (familiar) respectively, and the accuracy of the three recognition methods is judged by experienced researchers; Results analysis: As shown in Table 1, the recognition effect of the method and artificial recognition (skilled) is better, which is significantly higher than artificial recognition (unskilled); the gill filament structure is relatively simple, and the gill filament recognition effect is as follows: artificial recognition (skilled) > the method > artificial recognition (unskilled). The digestive gland lumen structure is relatively complex, and the digestive gland lumen recognition effect is as follows: the method > artificial recognition (skilled) > artificial recognition (unskilled). As can be seen, the method reduces the weight of experience in the recognition of key tissues of marine mussels, and is more conducive to the promotion of the quantitative detection method of marine mussel histopathology.

[0034] (2) Comparative Example 2 (measurement time) ① The test data set construction method is the same as step ① in Comparative Example 1, only the digestive gland data set is constructed; ② The thickness of the digestive gland lumen and gill filament epithelial cells in the test data set is measured by the method, artificial recognition (geometric conversion method) and artificial recognition (direct measurement method), respectively, and the measurement time is compared; ③ Main steps of artificial recognition (geometric conversion method): first, outline the digestive gland lumen profile, measure the relevant parameters using the image processing software ImageJ, and then convert the image, and then calculate the thickness of the digestive gland lumen using the formula; ④ Main steps of artificial recognition (direct measurement method): randomly select five positions on each digestive gland lumen, measure them using the image processing software ImageJ, and take the average value as the average thickness of the digestive gland lumen; ⑤ Results analysis: As shown in Table 2, the interpretation time of the three methods is as follows: the method (5 min) < direct measurement method (45 min) < geometric conversion method (55 min). The method greatly saves the interpretation time, improves the detection throughput and efficiency.

[0035] (3) Comparative Example 3 (effect of sulfamethoxazole stress on the thickness of the digestive gland lumen and gill filament of marine mussels) ① Laboratory acclimation of marine mussels: collect marine mussels and acclimate them in the laboratory, and the specific culture conditions are as follows: salinity 32.10±0.21‰, dissolved oxygen 82.71±3.36%, pH 8.12±0.22, temperature 18.30±0.27℃, light cycle 12 h darkness: 12 h light. Replace the artificial seawater every day, and feed the feed algae (isochrysis galbana), and after the marine mussels are stable, start the stress experiment; ② Sulfamethoxazole stress experiment setting: The stress experiment was set up with two experimental conditions, namely, the control group (only adding artificial seawater) and the sulfamethoxazole stress group (50 μg / L); the stress experiment lasted for 6 days, and the artificial seawater, food and sulfamethoxazole working solution were added every day to ensure the stability of the stress system; the sulfamethoxazole stress concentration was selected based on the environmental concentration reported in the literature to simulate the pollution in the actual sea area; ③ Tissue sampling: After the stress experiment, 6 mussels were randomly collected from each treatment group as biological repeats, and the standard H&E staining method was used to prepare the histopathological slides, and the high-resolution digital pathology scanning system was used to obtain the histopathological images; the slide samples were observed under an ordinary optical microscope (400x magnification), and the random sampling strategy was used to collect the histopathological images, ensuring that not less than 50 high-quality microscopic images were obtained for each sample; ④ The deep learning image analysis method and the direct measurement method (same as step ④ of Comparative Experiment 2) were used to quantitatively analyze the marine mussel histopathology images, and the tissue thickness response index was calculated; ⑤ The results show that after sulfamethoxazole stress, the digestive gland and gill of marine mussels show obvious histopathological response, the digestive gland lumen shows thickness increase, and the gill epithelium shows thickness decrease, i.e. the quantitative index of marine mussel histopathology can be used to represent the biological effect of marine pollution; as shown in the accompanying Figure 4 The results of the digestive gland lumen thickness and gill epithelium thickness obtained by the method and the traditional method are consistent, and there is no significant difference, indicating that the analysis results of the method are accurate, stable and greatly improve the detection efficiency; the gill tissue thickness response index H = 22.04% (warning C level), and the digestive gland tissue thickness response index H = 14.78% (warning B level), indicating that the environmental pollutants in the test sea area have caused obvious histopathological damage to marine mussels, and protective measures should be taken.

[0036] In summary, the present application provides a deep learning-based intelligent quantitative analysis method for marine mussel histopathology, which significantly improves the analysis throughput and detection efficiency through automatic image recognition and accurate measurement technology; the present application provides an efficient, stable and reliable implementation scheme for the biological effect evaluation of marine environmental pollution, and provides important technical support for marine ecotoxicology research and environmental monitoring field.

[0037] Table 1 Comparison of recognition effects of different methods

[0038] Table 2 Comparison of recognition time of different methods

[0039] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, it is intended that such additions and modifications be included within the scope of the application. It is the following claims, including any amendments thereto, which define the scope of the application.

[0040] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A biomarker-based intelligent marine environment pollution early warning method, characterized in that, The application relates to a method for monitoring marine environmental pollution based on a marine mussel histopathology image recognition model. The method comprises the following steps: S1) establishing an image quality control standard, obtaining marine mussel histopathology images, constructing a data set according to the established image quality control standard, then data labeling and quality control, and then pre-processing the data by format conversion and division to prepare for model training; S2) constructing a model based on a YOLOv11 architecture, taking an image segmentation module as a core to realize a multi-task segmentation function, and performing model training, verification and evaluation to realize recognition and segmentation of different structures in marine mussel histopathology images; S3) extracting different feature pixel points in the histopathology images through the trained model, calculating the shortest distance between the different feature pixel points by using a Euclidean distance, and generating a result image containing distance labels after averaging; 2. The biomarker-based intelligent early warning method of marine environmental pollution according to claim 1, characterized in that, S4) dividing marine environmental pollution early warning levels according to a response index H, and monitoring marine environmental pollution. In the step S1), the image quality control standard is established, the marine mussel histopathology images are obtained, the data set is constructed according to the established image quality control standard, then the data is labeled and quality controlled, and the data is preprocessed by format conversion and division, and the specific process is as follows: (1) data preparation Sa1) establish an image quality control standard, and the standard is as follows: (1) tissue morphology integrity: the tissue section integrity is good, and there is no folding or damage; (2) dyeing quality standardization: the dyeing is uniform, and the contrast between the cell nucleus and the cytoplasm is clear; (3) optical imaging accuracy: the image focusing is accurate, and there is no blurred area; Sa2) image acquisition and data set construction: collect the gill and digestive gland tissue samples of marine mussels and perform quality screening, prepare histopathology slides by using a dyeing method, obtain the histopathology images meeting the standard by using a high-resolution digital pathology scanning system, and construct a data set, which covers the morphological variations of different individuals and tissues; Sa4) Data quality control: randomly select the labeled histopathology images, calculate , the formula is , wherein A and B are the sets of labeling results of two researchers, represents the intersection of the two labeling results, and respectively represent the absolute values of the total number of two labeling results; it is assumed that The calculation result reaches , indicating that the quality requirements are met; Sa3) data labeling: the target structures in the histopathology images are labeled by using a professional labeling tool Labelme, including: for the gill filament tissue, the outer wall and the inner wall of the gill filament cell are selected as features and are labeled; for the digestive gland tissue, the inner wall and the outer wall of the lumen cell are selected as features and are labeled in layers; all the labels are saved in a JSON format file; (2) data preprocessing Sb1) data format conversion: the JSON format labeled information is converted into a YOLO format TXT file, each TXT file corresponds to one histopathology image, and the histopathology image and the corresponding TXT file are taken as a data set together; 3.The biomarker-based intelligent marine environment pollution early warning method according to claim 1, characterized in that, Sb2) data set division: the data set is divided into a training set and a test set to ensure the balance of data distribution. In the step S2), the model is constructed based on the YOLOv11 architecture, the multi-task segmentation function is realized by taking the image segmentation module as a core, and the specific steps are as follows: Model construction: YOLOv11 architecture recognition and segmentation engine is adopted; the model extracts image features through its efficient Backbone network, and fuses multi-scale features through the Neck part, and finally outputs the binary segmentation mask of the gill epithelial cells and the digestive gland lumen tissue through the segmentation head, wherein the segmentation head part is designed as a double-branch structure: branch A gill cell segmentation and branch B digestive gland lumen cell hierarchical segmentation; the role of YOLOv11 in this stage is to provide high-precision, pixel-level target contour for subsequent geometric parameter calculation; Model training: first, input the training set and validation set after data preparation and preprocessing into the model, set the parameters to train for 300 rounds, the number of samples per batch is 16, use the stochastic gradient descent (SGD) optimizer, update the model by iteratively calculating the parameter gradient, the formula is as follows: ； In the formula It is The model parameter vector after iterations, is the learning rate, is the loss function For parameter vector The gradient, It is The model parameter vector after iterations; Then, the Cosine annealing learning rate schedule is adopted, wherein the maximum value of the learning rate is set to 0.01, and the minimum value of the learning rate is set to ; Finally, based on the metric, training is stopped when the model does not improve significantly after 100 rounds of training during the training process. wherein the use of As a key indicator of model performance, the mean of the absolute difference between the predicted and actual values is calculated The average value within the interval is obtained The formula is as follows: ; In the formula is the average precision when the confidence threshold is t%; The value range of is 50-95, which represents the average precision at each confidence value calculated in the range of confidence threshold from 50% to 95%. Model validation and model evaluation: after training is completed, the trained model is verified on the validation set that did not participate in the training, and its performance on the data is evaluated; The calculation formula of the model accuracy is as follows: ; In the formula, the number of samples correctly predicted is the number of samples that the model successfully predicts the true label on the test data set; the total number of samples is the total number of samples in the test data set; The calculation formula of the recall rate is as follows: ; In the formula, the number of true positives is the number of actual positive samples that the model successfully predicts as positive; the number of false negatives is the number of samples that are actually positive but are incorrectly predicted as negative by the model. 4.The biomarker-based intelligent marine environment pollution early warning method according to claim 1, characterized in that, In step S3), different feature pixels in the histopathology image are extracted by the trained model, i.e., the outer wall pixel points and inner wall pixel points of the gill cell are extracted, the inner wall pixel points and outer wall pixel points of the digestive gland tissue lumen cell are extracted, and the shortest distance between the outer wall pixel points and the inner wall pixel points is calculated using the Euclidean distance, i.e., the thickness, and the average thickness is calculated to generate a result image containing thickness annotations, the specific steps are as follows: Sc1) Digestive gland lumen cell and gill cell segmentation: use the above trained model to segment the digestive gland lumen cell and gill cell in the histopathology image, and extract the pixel points constituting the outer wall and inner wall of the digestive gland lumen cell or gill cell, and define the internal structure and external boundary of the cell using the pixel points, i.e., the outer wall pixel points and the inner wall pixel points; Sc2) Euclidean distance calculation: Euclidean distance d is used for calculation, for inner wall pixel points and outer wall pixel points, assuming that the inner wall pixel point is and the outer wall pixel point is , the distance between each inner wall and outer wall is calculated, and the shortest distance between the inner wall pixel point and the outer wall pixel point is taken, and the calculation formula is: ; wherein d is the straight-line distance between the inner-wall pixel point and the outer-wall pixel point, is the coordinate difference in the horizontal direction between the inner-wall pixel point and the outer-wall pixel point, is the coordinate difference in the vertical direction between the inner-wall pixel point and the outer-wall pixel point. The minimum distance of the multiple Euclidean distances from the same inner wall pixel point to different outer wall pixel points is the thickness value, and the calculation formula is: ; Sc3) Thickness average calculation: average the thickness values of all positions, the calculation formula is: where N represents the number of thickness values of all positions, the thickness value of the i-th position, mark the result on the segmented cell, and record the number of recognized cells at the same time; ​ Sc4) Statistical result generation: generate a batch of annotated result images to show the quantitative results of cell thickness analysis. 5.The biomarker-based intelligent early warning method of marine environmental pollution according to claim 1, characterized in that, In step S4), the response index H of cell thickness is calculated, and the marine environmental pollution warning level is divided according to the value of H to monitor the marine environmental pollution, the specific steps are as follows: By comparing the cell thickness of the test sea area and the control sea area, the response index H of cell thickness is calculated, and the formula is as follows: ； Wherein, T1 represents the thickness of the test sea area, and T0 represents the thickness of the control sea area; According to the value of H, it is divided into 4 warning levels: H≤10% represents warning level A, 10%<H≤20% represents warning level B, 20%<H≤30% represents warning level C, and 30%<H represents warning level D.

Images