PVC plasma bag anomaly detection method for low-temperature environment

Abstract

The invention discloses a PVC plasma bag anomaly detection method for a low-temperature environment, and relates to the field of plasma bag quality detection.The method comprises the steps that S1, in the low-temperature environment of 2-8 DEG C, image acquisition is conducted on a PVC plasma bag in a frozen state; s2, the collected PVC plasma bag image is processed and analyzed, an analysis result is compared with standard data, and processing and analysis at least comprise color feature analysis and texture feature analysis which are executed in parallel; s3, based on the abnormality judgment result, performing fusion judgment through a confidence evaluation rule, and generating a comprehensive judgment result about whether the plasma bag is abnormal or not; according to the method, a special image acquisition device with a low-temperature protection function is deployed, a multi-dimensional image analysis algorithm aiming at plasma color and state characteristics is applied, and finally decision fusion is performed through an interpretable confidence rule, so that automatic and high-precision detection of the appearance quality of the frozen plasma bag is realized.

Description

Technical Field

[0001] This invention relates to plasma bag quality testing technology, specifically to a method for detecting anomalies in PVC plasma bags used in low-temperature environments. Background Technology

[0002] Frozen raw plasma is an important raw material for clinical treatment and biopharmaceutical production, and its quality and safety are of paramount importance. As the direct contact container for plasma, the integrity of PVC plasma bags and the color and state of the plasma inside are key indicators for judging product quality. Normal plasma should be pale yellow, yellow, or pale green, and should be free from defects such as hemolysis (abnormal red), chyle (milky white turbidity), visible foreign matter, and breakage.

[0003] Currently, the industry primarily relies on manual visual inspection in low-temperature environments (typically cold storage or low-temperature workshops at 2-8°C) to test the aforementioned quality indicators of frozen PVC plasma bags. Operators must visually observe and judge each stationary or slowly moving plasma bag in this low-temperature environment. This traditional method has the following inherent drawbacks: First, manual inspection is inefficient and cannot be matched with high-speed automated production lines, becoming a bottleneck in the production process; second, the test results are highly dependent on individual experience and subjective judgment, and the judgment standards of different personnel, or even the same person, may fluctuate under different conditions, making it difficult to guarantee the consistency and accuracy of the tests; third, prolonged low-temperature working environments not only affect personnel health and work efficiency but may also indirectly increase the risk of misjudgment and missed judgment due to decreased human comfort.

[0004] Therefore, there is a lack of existing technologies that can achieve efficient, objective, stable and accurate automated detection in the actual low-temperature environment of plasma storage and circulation, so as to replace the traditional manual visual inspection method. Summary of the Invention

[0005] The purpose of this invention is to provide an anomaly detection method for PVC plasma bags in low-temperature environments, in order to solve the problems of low efficiency, strong subjectivity, and difficulty in guaranteeing accuracy and stability when relying on manual inspection of plasma bags in low-temperature environments in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for detecting anomalies in PVC plasma bags used in low-temperature environments, comprising:

[0008] S1. Image acquisition of PVC plasma bags in a frozen state in a low temperature environment of 2℃ to 8℃;

[0009] S2. Process and analyze the acquired PVC blood plasma bag images, and compare the analysis results with standard data to determine whether there are any abnormalities. The processing and analysis includes at least parallel color feature analysis and texture feature analysis.

[0010] S3. Based on the anomaly judgment result, a fusion judgment is performed using confidence assessment rules to generate a comprehensive judgment result on whether the plasma bag is abnormal;

[0011] The acquisition process utilizes a low-temperature-adaptable image acquisition device to obtain clear images of the PVC plasma bag under the low-temperature environment.

[0012] Furthermore, the color feature analysis includes converting the image to the CIELAB color space and extracting the plasma region. , , Color characteristic value;

[0013] The texture feature analysis includes at least one of gray-level co-occurrence matrix analysis, wavelet transform analysis, and local binary pattern analysis to extract texture features from the image.

[0014] Furthermore, the processing and analysis also include:

[0015] The image undergoes edge detection preprocessing, and morphological processing, including adaptive threshold segmentation and connected component analysis, is used to identify damage or foreign objects.

[0016] Furthermore, the processing and analysis also include:

[0017] The image is input into a trained deep learning model to obtain anomaly prediction results.

[0018] Furthermore, the low-temperature adaptable image acquisition device includes at least an industrial camera with a built-in heating module and an LED light source with an anti-frost coating on its surface.

[0019] Furthermore, the confidence assessment rule includes: assigning weights to different anomaly judgment results and calculating the total confidence, and outputting an anomaly, normal or suspicious judgment result according to the threshold interval to which the total confidence belongs.

[0020] Furthermore, the step of outputting an abnormal, normal, or suspicious judgment result based on the threshold interval to which the total confidence level belongs includes:

[0021] If the total confidence score is higher than the first preset threshold, the plasma bag is determined to be abnormal;

[0022] If the total confidence score is lower than the second preset threshold, the plasma bag is determined to be normal;

[0023] If the total confidence score is between the first preset threshold and the second preset threshold, the comprehensive judgment result is marked as suspicious.

[0024] Furthermore, the LED light source is provided in two sets, which are respectively positioned above and below the PVC blood plasma bag.

[0025] In a second aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0026] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.

[0027] Compared with existing technologies, the present invention provides an abnormal detection method for PVC plasma bags in low-temperature environments. By deploying a dedicated image acquisition device with low-temperature protection and using a multi-dimensional image analysis algorithm targeting the color and state characteristics of plasma, and finally performing decision fusion through interpretable confidence rules, the method achieves automated and high-precision detection of the appearance quality of frozen plasma bags.

[0028] The beneficial effects of this invention are specifically reflected in the following aspects:

[0029] This invention enables fully automated, non-contact online inspection, with a single inspection completed within hundreds of milliseconds. It can seamlessly match the existing production line's conveying speed of 3.8 meters per minute or higher, improving inspection efficiency by several orders of magnitude.

[0030] The objective and quantitative algorithm standards replace subjective and volatile human judgment, ensuring high accuracy and consistency of detection results: by converting sensory indicators such as color and texture into precisely measurable parameters such as CIELAB spatial values ​​and GLCM feature values, and by setting unified judgment thresholds and fusion rules, the subjective differences between different people and the state fluctuations of the same person are completely eliminated.

[0031] The multi-algorithm parallel analysis and confidence fusion mechanism significantly improves the system's ability to distinguish complex and boundary cases, enabling the overall detection accuracy to be maintained at over 99.4%, with extremely low false positive and false negative rates. The detection quality is far superior to that of manual detection and is stable and reliable.

[0032] By employing industrial cameras with built-in heating modules and professional light sources with anti-frost coatings, the invention effectively prevents frost and condensation of core imaging components in low-temperature and high-humidity environments, ensuring long-term stability of image acquisition quality. This allows the system to be directly installed at actual plasma storage or processing stations at 2-8℃ for continuous and stable operation, meeting actual production needs.

[0033] The automation and digitalization of the entire process enable the recording, traceability, and statistical analysis of the test results, characteristic data, and images of each bag of plasma, providing a solid data foundation for quality traceability and process optimization. At the same time, the system reduces reliance on a large number of skilled quality inspectors, alleviates the workload of personnel in harsh environments, and improves the overall modernization level of production management. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0035] Figure 1 A flowchart illustrating the method steps provided in this embodiment of the invention;

[0036] Figure 2 A flowchart illustrating the comprehensive judgment process provided in this embodiment of the invention. Detailed Implementation

[0037] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0038] As attached Figure 1 and 2 As shown:

[0039] Example 1:

[0040] This invention provides an anomaly detection method for PVC blood plasma bags in low-temperature environments. The method is executed at an automated detection station deployed in a low-temperature cold storage facility (ambient temperature 2-8℃), and specifically includes the following steps:

[0041] S1. Low-temperature environment image acquisition:

[0042] The frozen PVC plasma bags (standard size 285mm × 120mm) to be inspected are transported to a closed inspection station via a conveyor line. This station is equipped with a transparent insulation cover to reduce heat exchange between the inside and outside. Subsequently, a low-temperature-adaptive image acquisition device is activated to acquire images of the plasma bags.

[0043] The specific components of an image acquisition device:

[0044] Industrial Camera: A 5-megapixel industrial area scan camera was selected. Its key features include an operating temperature range of -10℃ to 50℃, an IP67 protection rating, and a built-in intelligent heating module. This heating module consists of a temperature sensor and a PTC heating element inside the camera. It automatically activates when the ambient temperature drops below 5℃, maintaining the core temperature of the camera between 5-10℃, effectively preventing frost formation on the lens and sensor window, and ensuring clear and stable imaging.

[0045] Lighting System: Utilizes two sets of high-brightness white LED light sources (color temperature 5500K). All light source surfaces are coated with a nano-anti-frost coating. One set is a ring light source, installed directly above the plasma bag for surface illumination, highlighting the bag's surface texture and color; the other set is a backlight light source, installed below the plasma bag for transmitted illumination, enhancing the visibility of internal foreign objects and edge contours. Light source brightness can be continuously adjusted within the 0-100% range via software.

[0046] Optical lens: It is equipped with an 8mm fixed-focus industrial lens with a working distance of 200mm. The field of view can completely cover the blood plasma bag area, and the lens itself has wide temperature operating characteristics (-20℃~60℃).

[0047] S2. Image Processing and Multi-Feature Analysis:

[0048] The acquired color images (RGB format) are transmitted to an industrial computer for processing and analysis. This step performs color feature analysis and texture feature analysis in parallel.

[0049] Color Feature Analysis (CIELAB Space Transformation and Feature Extraction):

[0050] Preprocessing: First, Gaussian filtering is applied to the image to remove noise, and histogram equalization is used to enhance contrast.

[0051] Color space conversion: Convert the denoised RGB image to the device-independent CIELAB color space. The conversion formula is based on the standard illuminant D65 and the 2° standard observer condition, and the specific calculation is as follows:

[0052] First, convert RGB to XYZ space (according to the sRGB standard), then calculate using the following formula. , , value:

[0053]

[0054]

[0055]

[0056] in, , , It is the tristimulus value of a total internal reflection diffuser under a D65 standard illuminator.

[0057] Feature extraction and comparison: Segment the plasma region (ROI) in the image and calculate the number of pixels within that region. , , The average of the values. The obtained , , The value is compared with the pre-stored standard normal range. The normal plasma color range is defined as:

[0058] Pale yellow:

[0059] yellow:

[0060] Light green:

[0061] If the extracted value exceeds the above range, it is initially judged to be a color anomaly. A high value (>10) indicates that red blood cell rupture leads to the release of hemoglobin. The value is too high (>85) and the color saturation is low, resulting in a cloudy, milky white color. A value higher than 45 indicates excessive bilirubin content; record the color difference between this value and the standard value. ,when A value greater than 3 indicates a color anomaly.

[0062] Texture feature analysis (multi-algorithm fusion):

[0063] Image grayscale conversion: Converting the original RGB image to a grayscale image.

[0064] Gray-Level Co-occurrence Matrix (GLCM) Analysis: On a grayscale image, calculate the GLCM in four directions: 0°, 45°, 90°, and 135°, with a distance d=1. Extract four features from the GLCM in each direction: Contrast, Correlation, Energy, and Homogeneity. Averaging the feature values ​​in the four directions yields a 4-dimensional GLCM feature vector. Chyloery blood typically exhibits increased contrast and decreased energy.

[0065] Wavelet transform analysis: A one-level two-dimensional discrete wavelet decomposition (using the Daubechies4 wavelet basis) is performed on the grayscale image to obtain four sub-bands: LL, LH, HL, and HH. The energy of the three high-frequency sub-bands, LH, HL, and HH, is calculated as texture features. Chyle and foreign matter can cause abnormal high-frequency energy distribution.

[0066] Local Binary Pattern (LBP) analysis: The LBP map of the entire plasma region was calculated using the circular neighborhood LBP operator (P=8, R=1), and its 256-dimensional LBP histogram was statistically analyzed. The LBP histogram distribution of normal plasma is relatively dispersed, while that of chylous blood, due to its uniform texture, tends to cluster on a few patterns.

[0067] The GLCM, wavelet energy, and LBP histogram features mentioned above are compared with a pre-established database of normal blood plasma texture features (using Euclidean distance or cosine similarity) to determine whether the texture is abnormal.

[0068] S3. Confidence fusion and comprehensive judgment:

[0069] The system receives preliminary judgment results from color analysis and texture analysis ("normal color", "suspected hemolysis", "abnormal texture - chyle", etc.), as well as their respective difference scores (color difference ΔE, texture similarity score).

[0070] Confidence calculation: A weighted summation method is used for fusion. Preset base weights: color analysis results have a weight of 0.4, and texture analysis results have a weight of 0.4. Sub-confidence scores are mapped from 0 to 100 based on the difference score (the larger the ΔE, the higher the sub-confidence of color anomalies; the lower the texture similarity, the higher the sub-confidence of texture anomalies). Total confidence = color sub-confidence × 0.4 + texture sub-confidence × 0.4.

[0071] Threshold determination: Two thresholds are set: the first preset threshold (high threshold) is 95 points, and the second preset threshold (low threshold) is 70 points.

[0072] If the total confidence score is higher than 95, the overall judgment result is "abnormal", and the system will trigger the subsequent removal mechanism.

[0073] If the total confidence score is below 70, the overall judgment result is "normal", and the plasma bag is released.

[0074] If the total confidence score is between 70 and 95, the overall judgment result is marked as "suspicious", the system issues an alarm, prompting manual review, and the bag is redirected to the review station.

[0075] Example 2:

[0076] This embodiment is basically the same as the previous embodiment, except that in step S2, while performing color and texture analysis in parallel, an independent morphological analysis thread is added.

[0077] Edge detection: The Canny operator is used to perform edge detection on the acquired raw image (or denoised grayscale image) to obtain a clear edge image. Damage to plasma bags is usually manifested as unnatural breaks or burrs on the edge lines.

[0078] Adaptive thresholding: The image is binarized using a local adaptive threshold (Gaussian weighted adaptive threshold) to separate possible foreground objects (fibers, dust) from the background.

[0079] Connectivity analysis and feature selection: Eight-neighborhood connected regions are labeled in the binarized image. The area, perimeter, circularity, and minimum bounding rectangle of each connected region are calculated. Area thresholds (areas smaller than 10 pixels are considered noise and ignored) and shape regularity (long, thin shapes may be fibers, irregular clumps may be stains) are set to filter out regions suspected of containing foreign objects or being damaged.

[0080] Output results: The location, type, and size information of the identified suspected defects are used as a new "morphological anomaly" judgment result and output to the fusion judgment module of S3.

[0081] Fusion adjustments in step S3:

[0082] At this point, the system receives three analysis results: color, texture, and morphology. The weighting is adjusted to: color 0.3, texture 0.3, and morphology 0.3. The formula for calculating the total confidence score becomes: Total Confidence = Color Sub-Confidence × 0.3 + Texture Sub-Confidence × 0.3 + Morphology Sub-Confidence × 0.3. The decision threshold rule remains unchanged. If the morphological analysis detects obvious damage (such as large areas of missing edges), its sub-confidence can be directly set to 100 points, greatly increasing the total confidence score and ensuring accurate rejection.

[0083] Example 3:

[0084] This embodiment is basically the same as the previous embodiment, except that a deep learning model is introduced as a higher-order judge to improve the ability to identify complex and boundary cases.

[0085] A. Construction and training of deep learning models:

[0086] Model Architecture: A ResNet-50 pre-trained on the ImageNet dataset is used as the backbone network. To adapt to this task, the top fully connected classification layer of the original model was removed, and a Global Attention Module (CBAM) and a new classification head were added at the end. The new classification head is a fully connected layer with an output dimension of 6, corresponding to: normal, hemolysis, chyluria, jaundice, rupture, and foreign body.

[0087] Dataset Preparation: Over 10,000 images of frozen blood plasma bags taken in low-temperature environments were collected and precisely labeled by experts to create a high-quality dataset. This dataset includes approximately 5,000 normal samples and approximately 1,000 samples from each of the various abnormal categories. Data augmentation was performed using methods such as random rotation (±10°), horizontal / vertical flipping, and fine-tuning of brightness and contrast, expanding the dataset by 5 times.

[0088] Model training: The PyTorch framework was used, with cross-entropy loss as the loss function and the Adam optimizer employed (initial learning rate lr = 0.001). The ReduceLROnPlateau strategy was used during training (the learning rate was halved when the validation loss stalled). The batch size was set to 32, and training lasted for 100 epochs. The training, validation, and test sets were partitioned in a 7:2:1 ratio. The final model achieved an accuracy of over 99.5% and a recall of over 99% on the test set, with a model file size of approximately 98MB.

[0089] B. Online detection integration:

[0090] In step S2, a new deep learning model analysis thread is added. The blood plasma bag image, after normalization and preprocessing (scaled to 224×224, normalized), is input into the pre-trained, solidified model. The model outputs a 6-dimensional probability vector, for example [0.01, 0.85, 0.10, 0.02, 0.01, 0.01], indicating the highest confidence level for the model's prediction of "hemolysis".

[0091] Fusion adjustments in step S3:

[0092] The system receives preliminary binary judgment results ("abnormal" or "normal") from four parallel analysis modules:

[0093] Confidence score allocation: If the output of each analysis module (color, texture, morphology, deep learning model) is "abnormal", it is assigned a fixed confidence score of 25 points; if the output is "normal", it is assigned 0 points.

[0094] Total confidence score calculation: Total confidence score = color contribution score + texture contribution score + morphological contribution score + deep learning model contribution score. Therefore, the total confidence score can be 0, 25, 50, 75, or 100.

[0095] Threshold determination:

[0096] If the total confidence score is ≥95, the overall judgment result is "abnormal".

[0097] If the total confidence level is ≤70, the overall judgment result is "normal".

[0098] If 70 points < total confidence level < 95 points (i.e., total confidence level is 75 points), the overall judgment result is marked as "suspicious" and requires manual review.

[0099] Example 4:

[0100] This embodiment provides an automated detection system for PVC plasma bags used in low-temperature environments.

[0101] System components:

[0102] The system hardware is integrated into a stainless steel cabinet, including: a low-temperature conveyor module, a closed testing chamber (containing the image acquisition device of Example 1), an industrial control computer, and a human-machine interface (HMI).

[0103] Workflow:

[0104] Frozen plasma bags are fed into the testing chamber via a conveyor line at a speed of 3.8 m / min. Upon triggering the photoelectric sensor, the system simultaneously triggers the camera to take a picture and the light source to illuminate. The industrial control computer completes image acquisition, preprocessing, and parallel analysis of color, texture, morphology, and deep learning models within 50 milliseconds (deep learning inference takes approximately 15 milliseconds). Confidence fusion calculation is completed within 100 milliseconds, and a judgment is made. All detection data (images, feature values, judgment results, timestamps) are stored in the database in real time, and statistical reports can be generated on the HMI.

[0105] Performance test data:

[0106] The system was subjected to a stable low-temperature environment of 2-5℃ for 720 consecutive hours (30 days). Standard samples were used for testing, with 1000 samples of each type. The results are as follows:

[0107] Detection accuracy:

[0108] Testing items accuracy False positive rate False negative rate Normal color recognition ≥99.8% ≤0.2% ≤0.0% Hemolysis test ≥99.5% ≤0.3% ≤0.2% chyle detection ≥99.0% ≤0.5% ≤0.5% Jaundice test ≥99.2% ≤0.4% ≤0.4% Damage detection ≥99.5% ≤0.3% ≤0.2% Visible foreign object detection ≥99.0% ≤0.6% ≤0.4% Comprehensive testing ≥99.4% ≤0.3% ≤0.2%

[0109] Detection speed:

[0110] The average processing time for a single image is ≤50ms / image.

[0111] Processing time per image (maximum) ≤ 100ms / image;

[0112] The system's maximum throughput is ≥120 frames per second;

[0113] The conveyor line speed must be ≥3.8m / min;

[0114] Delay time (from image acquisition to output) ≤ 100ms;

[0115] Maximum daily testing capacity (24 hours) ≥ 1,000,000 bags.

[0116] Environmental adaptability:

[0117] After working continuously for 24 hours at 2℃, the camera and light source showed no frost formation, and the system operated without faults.

[0118] Mean Time Between Failures (MTBF) test value: >1200 hours.

[0119] The system operates normally within a power supply voltage fluctuation range of ±10%.

[0120] Working principle:

[0121] This invention solves the problem of poor imaging quality in the low-temperature environment of plasma storage by adapting proprietary low-temperature protective hardware (heated camera, anti-frost light source) to general machine vision technology. Furthermore, it designs a multi-dimensional analysis algorithm combination targeting the biological characteristics of plasma (CIELAB color space for accurate diagnosis of hemolysis / jaundice, multi-texture algorithm for identifying chyle / foreign bodies, and morphological detection of damage), and introduces a weighted confidence fusion decision mechanism of "algorithm + deep learning model". Ultimately, it achieves high-speed, high-accuracy, and high-stability automated detection of color and physical anomalies in frozen PVC plasma bags, replacing inefficient, subjective, and environmentally susceptible manual visual inspection.

[0122] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for detecting abnormalities in PVC plasma bags used in low-temperature environments, characterized in that, include: S1. Image acquisition of PVC plasma bags in a frozen state in a low temperature environment of 2℃ to 8℃; S2. Process and analyze the acquired PVC blood plasma bag images, and compare the analysis results with standard data to determine whether there are any abnormalities. The processing and analysis includes at least parallel color feature analysis and texture feature analysis. S3. Based on the anomaly judgment result, a fusion judgment is performed using confidence assessment rules to generate a comprehensive judgment result on whether the plasma bag is abnormal; The acquisition process utilizes a low-temperature-adaptive image acquisition device to obtain clear images of the PVC plasma bag under the low-temperature environment.

2. The method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 1, characterized in that, The color feature analysis includes converting the image to the CIELAB color space and extracting the plasma region. , , Color characteristic value; The texture feature analysis includes at least one of gray-level co-occurrence matrix analysis, wavelet transform analysis, and local binary pattern analysis to extract texture features from the image.

3. The method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 2, characterized in that, The processing and analysis also include: The image is preprocessed with edge detection, and then morphological processing, including adaptive threshold segmentation and connected component analysis, is used to identify damage or foreign objects.

4. A method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 2 or 3, characterized in that, The processing and analysis also include: The image is input into a trained deep learning model to obtain anomaly prediction results.

5. The method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 1, characterized in that, The low-temperature adaptable image acquisition device includes at least an industrial camera with a built-in heating module and an LED light source with an anti-frost coating on its surface.

6. The method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 4, characterized in that, The confidence assessment rules include: assigning weights to different anomaly judgment results and calculating the total confidence score, and outputting the judgment result of abnormal, normal or suspicious according to the threshold interval to which the total confidence score belongs.

7. The method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 6, characterized in that, The determination result of abnormal, normal, or suspicious based on the threshold interval to which the total confidence level belongs includes: If the total confidence score is higher than the first preset threshold, the plasma bag is determined to be abnormal; If the total confidence score is lower than the second preset threshold, the plasma bag is determined to be normal; If the total confidence score is between the first preset threshold and the second preset threshold, the comprehensive judgment result is marked as suspicious.

8. The method for detecting abnormalities in PVC plasma bags used in low-temperature environments according to claim 5, characterized in that, The LED light source is provided in two sets, which are respectively positioned above and below the PVC blood plasma bag.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-8.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1-8.

Images