Medical fabric washing control method and system based on visual inspection

Abstract

The invention discloses a medical fabric washing control method and system based on visual inspection, and relates to the technical field of visual inspection, and the method comprises the steps: determining a plurality of actual stain parts of a medical fabric according to a plurality of sub-visual inspection items, front and back images of the medical fabric and a preliminary stain region; determining a plurality of washing parameters of the medical fabric in the washing process on the basis of the plurality of actual stain parts, the type of the medical fabric and the washing mode of the washing equipment, so that a key washing area of the medical fabric is determined on the basis of the washing effect grade of the medical fabric at each position and the next use scene of the medical fabric; a plurality of cleanliness is determined based on detection of a key washing area of the medical fabric, the difference quantity of two adjacent cleanliness is determined, and the gradient type washing mode of the medical fabric is determined according to the positions of the cleanliness, the area form of the key washing area and the difference quantity of the two adjacent cleanliness. The washing effect of the medical fabric is improved.

Description

Technical Field

[0001] This invention relates to the technical field of visual inspection, and more particularly to a washing control method and system for medical fabrics based on visual inspection. Background Technology

[0002] With the development of technology, medical textiles refer to functional textiles specifically used in the fields of medical care, hygiene, and nursing. They not only possess the physical properties of ordinary textiles, but also have higher requirements in terms of antibacterial properties, impermeability, comfort, and biocompatibility due to the special nature of their use. Medical textiles will have corresponding stains after repeated use by users. In the existing technology, the medical textile is photographed and the current image of the medical textile is output. The stains are identified based on the current image of the medical textile and the stains are washed according to a single logic. However, the type of medical textile and the washing mode of the washing equipment are not taken into account, which affects the accuracy of multiple washing parameters in the washing process of medical textiles, thus affecting the washing effect of medical textiles. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a washing control method and system for medical fabrics based on visual detection.

[0004] This invention provides a vision-based method for controlling the washing of medical fabrics, comprising: determining the usage record of the medical fabric based on its model; determining a preliminary stained area based on the usage record; determining a vision detection mode for the medical fabric based on the preliminary stained area, the type of medical fabric, and the area of ​​use of the medical fabric; determining multiple sub-vision detection items based on the recognition of the vision detection mode; determining multiple actual stained portions of the medical fabric based on the multiple sub-vision detection items, the front and back images of the medical fabric, and the preliminary stained area; determining multiple washing parameters of the medical fabric during the washing process based on the multiple actual stained portions, the type of medical fabric, and the washing mode of the washing equipment; determining the washing effect level of the medical fabric at each location based on the multiple washing parameters; determining a critical washing area of ​​the medical fabric based on the washing effect level of the medical fabric at each location and the next use scenario of the medical fabric; determining multiple cleanliness levels based on the detection of the critical washing area of ​​the medical fabric, and determining the difference between two adjacent cleanliness levels; and determining a gradient washing method for the medical fabric based on the location of the multiple cleanliness levels, the regional morphology of the critical washing area, and the difference between two adjacent cleanliness levels.

[0005] This invention provides a vision-based medical fabric washing control system, which is applied to the aforementioned vision-based medical fabric washing control method. The vision-based medical fabric washing control system includes: The preliminary stain area module is used to determine the usage record of medical fabrics based on the model of the medical fabrics, and to determine the preliminary stain area based on the usage record of the medical fabrics. The multimodal data module is used to determine the visual inspection pattern of the medical fabric based on the initial stain area, the type of medical fabric, and the area of ​​the medical fabric used; and to determine multiple sub-visual inspection items based on the recognition of the visual inspection pattern. The washing parameter module is used to determine multiple actual stained parts of the medical fabric based on multiple sub-visual detection items, front and back images of the medical fabric, and preliminary stained areas; and to determine multiple washing parameters of the medical fabric during the washing process based on multiple actual stained parts, the type of medical fabric, and the washing mode of the washing equipment. The critical washing area module is used to determine the washing effect level of medical fabrics at various locations based on multiple washing parameters, and to determine the critical washing areas of medical fabrics based on the washing effect level of medical fabrics at various locations and the next use scenario of medical fabrics. The gradient washing mode module is used to determine multiple cleanliness levels based on the detection of key washing areas of medical fabrics, and to determine the difference between two adjacent cleanliness levels. The gradient washing mode of the medical fabric is determined based on the location of multiple cleanliness levels, the regional morphology of the key washing areas, and the difference between two adjacent cleanliness levels.

[0006] Compared with the prior art, the beneficial effects of the present invention are: In this embodiment of the invention, multiple sub-visual detection items are determined based on the recognition of the visual detection mode using the method described in this embodiment; multiple actual stained portions of the medical fabric are determined based on the multiple sub-visual detection items, the front and back images of the medical fabric, and the preliminary stained area; multiple washing parameters of the medical fabric during the washing process are determined based on the multiple actual stained portions, the type of medical fabric, and the washing mode of the washing equipment. This introduces a visual detection mode for the medical fabric, which is compatible with the overall consideration of multiple actual stained portions, the type of medical fabric, and the washing mode of the washing equipment, thereby improving the accuracy of multiple washing parameters of the medical fabric during the washing process.

[0007] Therefore, the washing effect level of medical fabrics at various locations is determined based on multiple washing parameters. Based on the washing effect level of the medical fabrics at various locations and the next usage scenario, key washing areas of the medical fabrics are identified. Multiple cleanliness levels are determined based on the detection of these key washing areas, and the difference between two adjacent cleanliness levels is determined. A gradient washing method for the medical fabrics is determined based on the location of multiple cleanliness levels, the regional morphology of the key washing areas, and the difference between two adjacent cleanliness levels. The introduction of key washing areas further controls these areas, achieving a holistic consideration of the location of multiple cleanliness levels, the regional morphology of the key washing areas, and the difference between two adjacent cleanliness levels. This improves the washing effect of the medical fabrics and enhances the accuracy of the gradient washing method. Attached Figure Description

[0008] Figure 1 This is a schematic flowchart of a vision-based medical fabric washing control method according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating step S11 of the vision-based medical fabric washing control method in an embodiment of the present invention. Figure 3 This is a flowchart illustrating step S12 in the vision-based medical fabric washing control method according to an embodiment of the present invention. Figure 4 This is a flowchart illustrating step S13 in the vision-based medical fabric washing control method according to an embodiment of the present invention. Figure 5 This is a flowchart illustrating step S14 of the vision-based medical fabric washing control method in an embodiment of the present invention. Figure 6 This is a flowchart illustrating step S15 of the vision-based medical fabric washing control method in an embodiment of the present invention. Figure 7 This is a schematic diagram of the structural composition of a vision-based medical fabric washing control system according to an embodiment of the present invention. Detailed Implementation

[0009] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0010] Please see Figures 1 to 7 A vision-based method for controlling the washing of medical fabrics, applied to vision-based inspection scenarios; the vision-based method for controlling the washing of medical fabrics includes: Step S11: Determine the usage record of the medical fabric based on its model, and determine the initial stained area based on the usage record of the medical fabric. Step S12: Determine the visual inspection pattern of the medical fabric based on the initial stain area, the type of medical fabric, and the area of ​​the medical fabric used; determine multiple sub-visual inspection items based on the recognition of this visual inspection pattern. Step S13: Determine multiple actual stained portions of the medical fabric based on multiple sub-visual detection items, front and back images of the medical fabric, and preliminary stained areas; determine multiple washing parameters of the medical fabric during the washing process based on multiple actual stained portions, the type of medical fabric, and the washing mode of the washing equipment. Step S14: Determine the washing effect level of the medical fabric at each location based on multiple washing parameters, and determine the key washing areas of the medical fabric based on the washing effect level of the medical fabric at each location and the next use scenario of the medical fabric. Step S15: Determine multiple cleanliness levels based on the detection of key washing areas of medical fabrics, and determine the difference between two adjacent cleanliness levels. Determine the gradient washing method of medical fabrics based on the location of multiple cleanliness levels, the regional morphology of key washing areas, and the difference between two adjacent cleanliness levels.

[0011] refer to Figure 2 In step S11, the specific steps are as follows: S111: Collect the model of the medical fabric, determine the past use events of the medical fabric based on the matching of the model of the medical fabric and the medical fabric database, determine multiple sub-use events based on the detection of the past use events of the medical fabric, and determine the use record of the medical fabric based on the identification of multiple sub-use events. S112: Based on the identification of medical fabric usage records, identify the stained areas of medical fabric during each use, mark multiple previous stained locations of medical fabric, and determine the initial stained areas based on the multiple previous stained locations of medical fabric.

[0012] In the embodiments of this application, unique identification information of medical fabrics is obtained through barcodes, RFID tags or image recognition, etc. For example, a medical fabric processing device scans a surgical gown with a built-in RFID reader and obtains its model number "SM-2023-A surgical gown". This model number includes information such as fabric material (70% polyester + 30% cotton), size (L), and production batch.

[0013] The system will match the obtained model information with the medical fabric database; the database stores the complete life cycle information of each fabric; for example, after querying "SM-2023-A surgical gown", the system will return all usage records of the surgical gown since it was put into use on August 15, 2023.

[0014] The system breaks down each usage event into finer-grained sub-use events for analysis. For example, if the surgical gown was used for 4 hours in the operating room on October 5, 2023, this usage event can be broken down into the following sub-use events: Sub-event 1: Surgical preparation stage (30 minutes), contact with disinfectants such as iodine; Sub-event 2: Surgical procedure stage (3 hours), contact with blood and body fluids; Sub-event 3: Surgical end stage (30 minutes), contact with cleaning solutions such as saline solution.

[0015] The system breaks down each usage event into finer-grained sub-use events for analysis. For example, if the surgical gown was used for 4 hours in the operating room on October 5, 2023, this usage event can be broken down into the following sub-use events: Sub-event 1: Surgical preparation stage (30 minutes), contact with disinfectants such as iodine; Sub-event 2: Surgical procedure stage (3 hours), contact with blood and body fluids; Sub-event 3: Surgical end stage (30 minutes), contact with cleaning solutions such as saline solution.

[0016] Furthermore, the system first analyzes the usage records of medical textiles to extract stain information generated during each use. This information includes key parameters such as stain type, location, area, and concentration. Based on the sub-use event data in the usage records and combined with the structural characteristics of the medical textiles, the system predicts and determines the stain area that will appear during each use.

[0017] For example, for medical fabrics such as surgical gowns, the system analyzes the types and locations of contaminants that come into contact with them at different stages, such as surgical preparation, surgery, and surgery completion. The front area of ​​the surgical gown is most likely to come into contact with blood and body fluids during the surgery, while the cuff area comes into contact with disinfectant during the surgical preparation stage. Based on the analysis results, the system generates a corresponding stain distribution map for each use.

[0018] The system marks all historical stain locations on the digital model or image of the medical fabric. These marks typically use a coordinate system, establishing a two-dimensional or three-dimensional coordinate system with a fixed point on the fabric as the origin, to accurately record the location of each stain. The marking process considers multiple attributes of the stain, including: stain type (blood, body fluids, drugs, disinfectants, etc.); stain location coordinates (x, y coordinates); stain area size; stain concentration or severity; and stain formation time. The system stores this marking information in a database, forming a "stain map" of the medical fabric, providing data support for subsequent analysis.

[0019] The system analyzes all marked previous stain locations and uses statistical analysis and pattern recognition algorithms to determine the initial stain areas of medical fabrics; identifies areas with high stain frequency; clusters adjacent or nearby stain locations into areas; analyzes the changing trends of stain distribution; assesses the probability and severity of stains in different areas; based on these analyses, the system determines one or more initial stain areas.

[0020] Specifically, the system first analyzed its usage records and found that the surgical gown had been used 15 times in the past 3 months, mainly in the operating room. The system marked 42 historical stain locations on the digital model of the surgical gown, including: front area: 28 stain marks, of which 20 were blood stains and 8 were bodily fluid stains; right cuff area: 8 stain marks, of which 5 were disinfectant stains and 3 were medication stains; left cuff area: 4 stain marks, all of which were disinfectant stains; collar area: 2 stain marks, which were sweat stains.

[0021] In the initial stain area identification stage, the system analysis revealed the following: 18 blood stains appeared within a 50mm radius around the center point of the front placket area (coordinates x=150mm, y=200mm), forming a clear stain hotspot; 6 disinfectant stains appeared within a 30mm radius around the right cuff area (coordinates x=50mm, y=300mm); and 3 disinfectant stains appeared within a 20mm radius around the left cuff area (coordinates x=250mm, y=300mm). Based on these analyses, the system identified three initial stain areas: Primary stain area: the center area of ​​the front placket (coordinates x=100-200mm, y=150-250mm), with a risk assessment of "high"; Secondary stain area: the right cuff area (coordinates x=35-65mm, y=285-315mm), with a risk assessment of "medium"; and Secondary stain area: the left cuff area (coordinates x=235-265mm, y=285-315mm), with a risk assessment of "medium". (y=290-310mm), the risk assessment is "low to medium".

[0022] refer to Figure 3 In step S12, the specific steps are as follows: S121: Acquire the current image of the medical fabric based on the photograph of the medical fabric, determine the surface texture of the medical fabric based on the texture recognition of the current image of the medical fabric, determine the type of the medical fabric based on the surface texture and model of the medical fabric, and determine the usage area of ​​the medical fabric based on the area recognition of the current image of the medical fabric. S122: Determine the first detection coefficient based on the initial stain area and the type of medical fabric, determine the second detection coefficient based on the initial stain area and the area of ​​use of the medical fabric, and determine the visual detection mode of the medical fabric based on the mapping relationship between the first detection coefficient, the second detection coefficient and the detection mode. S123: Collect the visual inspection pattern of medical fabrics, determine the corresponding visual inspection procedure table based on the recognition of the visual inspection pattern of medical fabrics, determine multiple sub-visual inspection items based on the recognition of the visual inspection procedure table, and determine the inspection content corresponding to each visual inspection item with different priorities. Multiple sub-visual inspection items perform orderly visual inspection of medical fabrics.

[0023] In the embodiments of this application, an industrial-grade high-resolution camera (such as 20 megapixels) is used in conjunction with a standard light source system for shooting; a multi-angle shooting scheme is adopted, typically including four angles: 45° and 90° from the front and 45° and 90° from the back; the light source configuration uses a D65 standard light source with a color temperature of 6500K to ensure accurate color reproduction.

[0024] The Gray-Level Co-occurrence Matrix (GLCM) algorithm was used to extract texture feature parameters; four key texture indices, namely contrast, energy, uniformity, and correlation, were calculated; multi-scale texture analysis was performed using wavelet transform; the periodicity and directionality of the texture were analyzed using Fourier transform; a texture feature database containing standard texture parameters for various medical fabrics was established; simultaneously, a fabric type classification decision tree was built, including texture features, model information, and physical properties; and multi-class classification was performed using the Support Vector Machine (SVM) algorithm.

[0025] The fabric outline is identified using edge detection algorithms (such as the Canny operator); image pixels are converted into actual physical dimensions using pixel calibration techniques; effective and non-use areas are distinguished using image segmentation techniques; and the ratio of the total fabric area to the effective use area is calculated.

[0026] Furthermore, based on the initial stain area and the type of medical fabric, the first detection coefficient is determined. The system parameterizes the location, area, shape, and other features of the initial stain area. Different weighting coefficients are set according to the material characteristics and usage features of the fabric type (such as surgical gowns, bed sheets, surgical towels, etc.). The weighted scoring method is used to calculate the matching degree between the stain area and the fabric type. The formula for calculating the first detection coefficient is: K1 = Σ(stain feature parameter i × weighting coefficient i).

[0027] Specifically, taking a "SM-2023-A type surgical gown" as an example, the initial stained areas obtained by the system from step S112 include: the central area of ​​the front placket: area 150cm², irregular shape, risk assessment "high"; the right cuff area: area 45cm², strip-shaped, risk assessment "medium"; the left cuff area: area 38cm², strip-shaped, risk assessment "low to medium". Based on the material characteristics of the surgical gown (70% polyester, 30% cotton) and its usage, the system sets the following weighting coefficients: front placket area weight: 0.6 (the front placket is the main contaminated area); cuff area weight: 0.3 (the cuff area is the second most important); other areas weight: 0.1; the first detection coefficient is calculated as: K1 = (150×0.6 + 45×0.3 + 38×0.3) / (150+45+38) = (90 + 13.5 +11.4) / 233 = 114.9 / 233 = 0.493; The system normalizes the K1 value to 0.49, indicating that the stain distribution characteristics of the surgical gown match the typical staining pattern of the surgical gown by 49%.

[0028] The second detection coefficient is determined based on the initial stained area and the area of ​​medical fabric used. The ratio of the total area of ​​the initial stained area to the area of ​​fabric used is calculated. The distribution density and concentration of the stained area are taken into account. A stain severity correction factor is introduced to adjust the coefficient. The formula for calculating the second detection coefficient is: K2 = (total stained area / area used) × distribution density factor × severity factor.

[0029] Specifically, the initial total area of ​​the stained area is 150 + 45 + 38 = 233 cm²; the area used, obtained from step S121, is 6500 cm²; the stain distribution density is mainly concentrated in three areas, with a distribution density factor of 1.2; the stain severity is assessed as "high" for the front area, with a severity factor of 1.3; the cuff area is assessed as "medium" and "low-medium", with a severity factor of 1.1; the second detection coefficient is calculated as follows: K² = (233 / 6500) × 1.2 × [(150×1.3 +45×1.1 + 38×1.1) / (150+45+38)] = 0.0358 × 1.2 × (195 + 49.5 + 41.8) / 233 = 0.0358 × 1.2 × 286.3 / 233 = 0.0358 × 1.2 × 1.229 = 0.0528; The system standardizes the K2 value to 0.053, indicating that the stain coverage of the surgical gown is 5.3%.

[0030] The visual detection mode of medical fabrics is determined based on the first detection coefficient, the second detection coefficient, and the mapping relationship of detection modes. By combining the two detection coefficients, the most suitable visual detection mode is determined, a detection mode mapping relationship table is established, and visual detection modes corresponding to different coefficient combinations are defined. The visual detection modes include: basic mode, standard mode, fine mode, and professional mode. Each detection mode corresponds to different detection parameters (such as resolution, light source conditions, algorithm complexity, etc.).

[0031] Specifically, based on the calculated K1=0.49 and K2=0.053, the system queries the preset detection mode mapping table, as shown in Table 1: Table 1: Mapping Relationship of Detection Modes

[0032] Based on K1=0.49 (within the range of 0.3-0.6) and K2=0.053 (within the range of 0.02-0.06), the system determines the visual detection mode of the surgical gown as "standard mode". The specific detection parameters of the standard mode include: image acquisition resolution: 12 million pixels; light source configuration: main light source (D65 standard light source) + auxiliary light source (side 45° LED light source); detection algorithm: standard stain recognition algorithm, including color analysis, texture analysis and shape analysis; detection accuracy: the smallest identifiable stain area is 0.5mm².

[0033] Therefore, the system retrieves the corresponding detection parameter configuration from the detection mode database; the parameters include image acquisition parameters, light source control parameters, algorithm processing parameters, etc.; the system automatically adjusts the working status of the hardware equipment according to the detection mode; the detection modes are stored in parameterized form for easy retrieval and switching; at the same time, the detection procedure table contains information such as detection sequence, detection content, and detection method; different detection modes correspond to different procedure tables to ensure detection efficiency and quality.

[0034] Specifically, the visual inspection mode obtained by the system from S122 is "standard mode". The system immediately retrieves the complete parameter configuration of the standard mode from the database: Image acquisition parameters: resolution 12 megapixels, exposure time 1 / 100 second, ISO value 200; Light source control parameters: main light source brightness 80%, auxiliary light source brightness 60%, color temperature 6500K; Algorithm processing parameters: color analysis accuracy 95%, texture analysis accuracy 90%, shape analysis accuracy 85%. For surgical gown inspection in standard mode, the system determines the visual inspection procedure table, which is shown in Table 2: Table 2 Visual Inspection Process Table

[0035] Based on the identification of the visual inspection process list, multiple sub-visual inspection items are identified. The system decomposes the process list into specific sub-inspection items and sets priorities: each sub-inspection item corresponds to specific inspection content and methods; priority settings are based on medical safety requirements and inspection efficiency considerations; high-priority items are executed first to ensure that key issues are detected in a timely manner; there are logical relationships between sub-items to form a complete inspection network.

[0036] A dynamic scheduling algorithm is used to ensure that high-priority projects are executed first; real-time feedback is provided during the detection process, and serious problems can be detected and terminated in advance with an alarm; detection results are recorded in real time to form a complete detection report; the system can dynamically adjust subsequent detection parameters based on the detection results.

[0037] Specifically, based on the above process schedule, the sub-visual inspection items determined by the system are as follows: Priority 1 (highest priority): P1-01: Blood residue detection (for the front placket area); P1-02: Body fluid contamination detection (for the front placket and cuff areas); P1-03: Obvious damage detection (overall inspection) Priority 2: P2-01: Disinfectant residue detection (overall inspection); P2-02: Fine stain detection (for the front placket area); P2-03: Seam integrity inspection (all seam areas); Priority 3: P3-01: General cleanliness assessment (overall inspection); P3-02: Fabric aging assessment (overall inspection); P3-03: Color consistency inspection (overall inspection).

[0038] The "SM-2023-A surgical gown" underwent sequential testing: Phase 1 (Priority 1 test, 12 seconds): The system first performed P1-01 blood residue detection, finding two suspected blood residue points (0.8mm and 1.2mm in area) in the front area; then P1-02 body fluid contamination detection, finding one body fluid contamination mark (2.5mm in area) in the right cuff area; finally, P1-03 obvious damage detection, finding no obvious damage. Phase 2 (Priority 2 test, 10 seconds): P2-01 disinfectant residue detection, finding normal disinfectant concentration in the front area; P2-02 fine stain detection, confirming the blood residue points found in Phase 1; P2-03 suture integrity detection, finding slight loosening of the sutures at the left cuff.

[0039] Phase 3 (Priority 3 inspection, 8 seconds): Perform P3-01 general cleanliness assessment, overall cleanliness score is 85 points (out of 100); Perform P3-02 fabric aging assessment, aging level is "mild"; Perform P3-03 color consistency check, color consistency is good; After the inspection is completed, the system generates a complete inspection report, including: Main problems found: blood residue in the front area, body fluid contamination on the right cuff, loose seam on the left cuff; Comprehensive assessment result: needs to be washed again, focusing on the front and right cuff areas; Recommended washing parameters: increase the washing intensity of the front area by 20%, extend the washing time by 15%.

[0040] refer to Figure 4 In step S13, the specific steps are as follows: S131: Acquire the current image of the medical fabric, determine the front and back images of the medical fabric based on the division of the current image of the medical fabric, determine multiple stain surfaces based on the visual inspection of the front and back images of the medical fabric, and determine the first stain combination based on multiple sub-visual inspection items and multiple stain surfaces. S132: Determine a second stain combination based on multiple stained surfaces and initial stained areas, determine multiple actual stained portions of the medical fabric based on the matching of the first stain combination and the second stain combination, and mark the location and stain area of ​​the multiple actual stained portions; S133: Collect the types of medical fabrics, determine the washing mode of the washing equipment based on multiple operating parameters of the washing equipment, determine the washing content corresponding to multiple actual stains based on multiple actual stains, the types of medical fabrics and the washing mode of the washing equipment, determine the corresponding washing parameters based on the identification of the washing content corresponding to multiple actual stains, and output multiple washing parameters of medical fabrics during the washing process.

[0041] In the embodiments of this application, a current image of the medical fabric is acquired, and the front and back images of the medical fabric are determined based on the division of the current image of the medical fabric. Multiple stain surfaces are determined based on the visual inspection of the front and back images of the medical fabric. A first stain combination is determined based on multiple sub-visual inspection items and multiple stain surfaces, which takes into account the overall consideration of multiple sub-visual inspection items and multiple stain surfaces, and ensures the accuracy of the first stain combination.

[0042] At this point, the current image of the medical fabric is acquired, and the fabric outline is identified using an edge detection algorithm (Canny operator); the front and back of the fabric are distinguished using a texture analysis algorithm (LBP texture feature); the front and back of the fabric are classified and identified using a deep learning model (CNN network); the accuracy of front and back division is required to reach more than 98%; the division result includes the region coordinates and confidence scores of the front and back of the fabric.

[0043] Simultaneously, multispectral image analysis technology is employed, combining visible light, near-infrared, and ultraviolet spectral information; color space conversion (RGB, HSV, Lab) is used to enhance stain contrast; image segmentation algorithms (watershed algorithm, region growing algorithm) are applied to separate stain areas; stain surface identification includes feature parameters such as stain type, area, concentration, and shape; identification accuracy requirements: minimum identifiable stain area 0.1 mm², type identification accuracy 95%.

[0044] Specifically, taking a "SM-2023-A surgical gown" as an example, the system acquires images under standard lighting conditions: using a 50-megapixel industrial camera equipped with a D65 standard light source; acquisition distance: 50cm from the fabric surface to the camera lens; lighting conditions: illuminance 2000 lux, color temperature 6500K; acquisition angles: front 0°, back 180°, left 45°, right 135°; image resolution: 8192×5464 pixels; acquisition time: exposure time of 1 / 60 second for each angle, with a total acquisition time not exceeding 5 seconds; generating 4 RAW format original images and 4 JPEG format preview images.

[0045] Specifically, the acquired "SM-2023-A type surgical gown" images were divided into front and back sides: Edge detection: The Canny operator was used to detect the outline of the surgical gown and identify the complete outer boundary; Texture analysis: Front side texture features: smooth surface, texture uniformity 0.92, texture complexity 0.35; Back side texture features: rough surface, texture uniformity 0.76, texture complexity 0.68; Deep learning classification: Input: image texture features and color features; Model output: front side confidence 0.99, back side confidence 0.97; Segmentation results: Front side image region: coordinates (x100-700, y50-550), confidence 0.99; Back side image region: coordinates (x100-700, y650-1150), confidence 0.97; Segmentation accuracy: 99.2%.

[0046] Specifically, stain surface identification was performed on the front and back images of the "SM-2023-A surgical gown": Front image stain identification: Stain surface S1: Location (x350-450, y200-300), area 150mm², type blood residue, concentration 0.8 (high concentration), irregular shape; Stain surface S2: Location (x500-550, y350-400), area 75mm², type bodily fluid contamination, concentration 0.6 (medium concentration), oval shape; Stain surface S3: Location (x250-300, y400-450), area 30mm², type grease stain, concentration 0.4 (low concentration), round shape.

[0047] Reverse image stain recognition: Stain surface S4: Location (x400-500, y800-900), area 200mm², type disinfectant residue, concentration 0.5 (medium concentration), shape diffused; Stain surface S5: Location (x300-350, y750-800), area 45mm², type general stain, concentration 0.3 (low concentration), shape striped; Stain surface feature parameters: A total of 5 stain surfaces were identified, with a total area of ​​500mm²; Stain type distribution: 1 blood residue, 1 body fluid contamination, 1 grease stain, 1 disinfectant residue, and 1 general stain; Stain concentration distribution: 1 high concentration, 2 medium concentrations, and 2 low concentrations; Recognition accuracy: 96.5%.

[0048] The first stain combination is determined based on multiple sub-visual detection items and multiple stained surfaces, and a mapping relationship matrix between sub-visual detection items and stain types is established. A priority matching algorithm is used to pair the stained surfaces with the most suitable detection items. The first stain combination includes stain surface information, corresponding detection items, and detection priorities. The combination result supports dynamic adjustment, and the matching relationship is updated in real time according to changes in stain characteristics.

[0049] Specifically, based on the sub-visual detection items and identified stain surfaces in S123: Sub-visual detection item list: P1-01: Blood residue detection (priority 1); P1-02: Body fluid contamination detection (priority 1); P2-01: Grease stain detection (priority 2); P2-02: Fine stain detection (priority 2); P3-01: Disinfectant residue detection (priority 3); P3-02: General cleanliness assessment (priority 3); Stain surface matching with detection items: S1 surface (blood residue) matches item P1-01, matching degree 0.95; S2 surface (body fluid contamination) matches item P1-02, matching degree 0.92; S3 surface (grease stain) matches item P2-01, matching degree 0.88; S4 surface (disinfectant residue) matches item P3-01, matching degree 0.90; S5 surface (general stain) matches item P3-02, matching degree 0.85.

[0050] First stain combination determined: Combination item 1: S1 surface + P1-01 blood residue detection, priority 1; Combination item 2: S2 surface + P1-02 bodily fluid contamination detection, priority 1; Combination item 3: S3 surface + P2-01 grease stain detection, priority 2; Combination item 4: S4 surface + P3-01 disinfectant residue detection, priority 3; Combination item 5: S5 surface + P3-02 general cleanliness assessment, priority 3; Testing sequence arrangement: First stage: Combination items 1 and 2 (priority 1), estimated testing time 15 seconds; Second stage: Combination item 3 (priority 2), estimated testing time 10 seconds; Third stage: Combination items 4 and 5 (priority 3), estimated testing time 20 seconds; Total testing time: 45 seconds.

[0051] This systematic method for identifying stain combinations allows the system to accurately match the identified stain surfaces with the most suitable detection items and perform detection according to priority. This ensures timely identification of critical stains and improves overall detection efficiency. This initial stain combination provides important basic data for subsequent determination of actual stain components and optimization of washing parameters.

[0052] Furthermore, a second stain combination is determined based on multiple stained surfaces and preliminary stained areas. Multiple actual stained portions of the medical fabric are determined based on the matching of the first and second stain combinations, and the positions and stain areas of the multiple actual stained portions are marked. This approach takes into account the overall consideration of matching the first and second stain combinations, ensuring the accuracy of the multiple actual stained portions of the medical fabric.

[0053] At this point, a second stain combination is determined based on multiple stained surfaces and the initial stained area. The system spatially matches the multiple stained surfaces identified in S131 with the initial stained area determined in S112; it uses a region overlap algorithm (such as IoU, Intersection over Union) to calculate the overlap between the stained surfaces and the initial stained area; it assigns a weighted score to each stained surface, based on factors including: the degree of overlap with the initial stained area; the frequency of the stained surface in historical usage events; and the impact of the stain type on medical safety; stained surfaces with scores exceeding a threshold are included in the second stain combination.

[0054] Specifically, taking a "SM-2023-A type surgical gown" as an example: Initial stain areas (from S112): Front placket area (coordinates: x100–300mm, y150–400mm); Left cuff area (coordinates: x235–265mm, y290–310mm); Stain surface (from S131): S1: Blood residue (coordinates: x120–280mm, y160–390mm); S2: Body fluid contamination (coordinates: x240–260mm, y295–305mm); S3: Grease stains (coordinates: x50–90mm, y200–250mm); S4: Disinfectant residue (coordinates: x310–350mm, y180–220mm).

[0055] Matching process: S1 overlaps with the front placket area by 95%, score 0.95; S2 overlaps with the left cuff area by 90%, score 0.90; S3 does not overlap with the front placket area, score 0.10; S4 does not overlap with the initial stain area, score 0.05; Second stain combination: S1 (blood residue) + front placket area; S2 (body fluid contamination) + left cuff area.

[0056] Based on the matching of the first and second stain combinations, multiple actual stained portions of the medical fabric are identified. The first and second stain combinations are compared to filter out stained surfaces that exist in both combinations. A confidence assessment algorithm, combined with image features and historical data, is used to perform a secondary verification of the stain authenticity. Bounding boxes are marked on the confirmed stained surfaces to generate a list of actual stained portions. Simultaneously, an image annotation tool is used to draw bounding boxes around the actual stained portions. The pixel area of ​​each bounding box is calculated and converted into the actual physical area (mm²) according to the image resolution. The center coordinates, boundary coordinates, and area information of each stain are recorded. The marked information is stored as structured data for subsequent washing parameter optimization.

[0057] Specifically, the first stain combination is: S1 (blood residue) + P1-01 item; S2 (body fluid contamination) + P1-02 item; S3 (grease stain) + P2-01 item; S4 (disinfectant residue) + P3-01 item; S5 (general stain) + P3-02 item; the second stain combination (this step) is: S1 (blood residue) + front area; S2 (body fluid contamination) + left cuff area; matching results: S1 exists in both combinations and is confirmed as an actual stain; S2 exists in both combinations and is confirmed as an actual stain; S3, S4, and S5 only exist in the first stain combination and are not included in the actual stains; list of actual stains: actual stain 1: blood residue, located in the front area; actual stain 2: body fluid contamination, located in the left cuff area.

[0058] Actual stain part 1: Blood residue; Center coordinates: x200mm, y275mm; Boundary coordinates: x120–280mm, y160–390mm; Pixel area: 15,000 pixels; Actual area: 160mm × 230mm = 36,800mm² (approximately 368cm²); Actual stain part 2: Bodily fluid contamination; Center coordinates: x250mm, y300mm; Boundary coordinates: x240–260mm, y295–305mm; Pixel area: 500 pixels; Actual area: 20mm × 10mm = 200mm² (approximately 2cm²).

[0059] Therefore, this method collects information on the types of medical fabrics and determines the washing mode of the washing equipment based on multiple operating parameters. It then determines the washing content corresponding to each actual stain based on multiple actual stain areas, the types of medical fabrics, and the washing mode of the washing equipment. Based on the identification of the washing content corresponding to these stain areas, corresponding washing parameters are determined to output multiple washing parameters for the medical fabrics during the washing process. This method incorporates a holistic approach to identifying the washing content corresponding to multiple actual stain areas, ensuring the accuracy of the corresponding washing parameters. Furthermore, it introduces a visual detection mode for the medical fabrics, further enhancing the accuracy of the multiple washing parameters during the washing process by considering the overall consideration of multiple actual stain areas, the types of medical fabrics, and the washing mode of the washing equipment.

[0060] At this time, the system automatically obtains the type information of medical fabrics through RFID tags, barcodes or image recognition technology; the types of fabrics include but are not limited to: surgical gowns, surgical towels, hospital gowns, bed sheets, duvet covers, etc.; different types of fabrics have different material properties (such as cotton, polyester-cotton blends, non-woven fabrics, etc.) and washing requirements; the system has a built-in fabric database that stores key information such as material parameters, temperature resistance, and color fastness of various fabrics.

[0061] The system determines the washing mode of the washing equipment based on multiple operating parameters. These parameters are collected in real time, including: water temperature range (30-90℃); washing time range (5-60 minutes); mechanical strength adjustment (low, medium, and high); detergent type (alkaline, neutral, enzyme, disinfectant, etc.); and spin speed (400-1200 rpm). Based on the equipment parameters and fabric type, the system selects the most suitable washing mode from a preset washing mode library. The washing modes include lightly soiled, moderately soiled, and heavily soiled modes.

[0062] Specifically, taking an "SM-2023-A surgical gown" as an example: The system obtains fabric type information through RFID scanning: Fabric type: surgical gown; Material: 65% polyester, 35% cotton; Temperature resistance: maximum withstands 85℃; Color fastness: Grade 4 (1-5, Grade 5 is best); Special requirements: chlorine bleach cannot be used; For the SM-2023-A surgical gown, the system collects the washing equipment parameters: Water temperature adjustment range: 30-85℃; Washing time range: 5-45 minutes; Mechanical strength: low, medium, and high; Applicable detergents: alkaline detergent, neutral detergent, enzyme detergent, disinfectant; Spin-drying speed: 400-1000 rpm; Based on the fabric characteristics and equipment parameters, the system determines the washing mode to be "moderately soiled".

[0063] Based on multiple actual stains, the type of medical fabric, and the washing mode of the washing equipment, the system determines the washing content corresponding to multiple actual stains and performs matching analysis between the actual stains determined in S132 and the washing mode.

[0064] Based on information such as stain type, area, and location, specific washing procedures are determined for each stain. These procedures include: pre-treatment methods (such as spot soaking, manual treatment, etc.); main wash parameters (water temperature, time, detergent type and dosage); number of rinses and parameters; disinfection methods (heat disinfection, chemical disinfection, etc.); and consideration of the limitations imposed by fabric type on the washing procedures, such as temperature resistance and chemical sensitivity.

[0065] The system integrates washing parameters for all stained areas, generating comprehensive washing parameters. It adopts a "strictest principle," meaning that when different stained areas have different requirements for the same parameter, the most stringent requirement is selected. Washing parameters include: pre-treatment parameters (time, temperature, detergent type and dosage); main wash parameters (water temperature, time, mechanical intensity, detergent type and dosage); rinsing parameters (number of rinses, time, water temperature); disinfection parameters (method, temperature, time, disinfectant dosage); and spin-drying parameters (spindle speed, time). Therefore, multiple washing parameters are included, encompassing pre-treatment, main wash, rinsing, disinfection, and spin-drying parameters. The system also considers energy efficiency and environmental requirements, optimizing parameters while ensuring washing effectiveness.

[0066] The system outputs the determined washing parameters in a structured format, which can be directly transmitted to the washing equipment control system. The output format is usually XML or JSON, containing detailed parameters for all washing stages. The system also generates washing parameter description documents, explaining the basis for parameter settings and expected effects. The output parameters include real-time monitoring indicators, such as water temperature sensor data and turbidity sensor data, for quality control of the washing process.

[0067] Specifically, for the SM-2023-A surgical gown, the system identified two actual stained areas: Stain 1 (S1): Type: Blood residue; Area: 36,800 pixels (approximately 184 cm²); Location: Front area; Stain severity: Severe; Stain 2 (S2): Type: Body fluid contamination; Area: 200 pixels (approximately 1 cm²); Location: Left cuff; Stain severity: Moderate; The system determined the following washing steps: Stain 1 (S1): Pretreatment: Soak in a localized enzyme detergent for 10 minutes; Main wash... Section 1: Water temperature 75℃, time 25 minutes, use 120ml alkaline detergent; Rinse: 3 times, 8 minutes each time, water temperature gradually decreases (60℃→45℃→30℃); Disinfection: Heat disinfection, water temperature 70℃, time 10 minutes; Stain 2 (S2): Pre-treatment: No special pre-treatment required; Main wash stage: Water temperature 65℃, time 20 minutes, use 100ml neutral detergent; Rinse: 2 times, 8 minutes each time, water temperature gradually decreases (50℃→30℃); Disinfection: Chemical disinfection, use 80ml disinfectant.

[0068] Specifically, for the SM-2023-A surgical gown, the system integrates the washing procedures for both stained areas, determining the final washing parameters: Pre-treatment stage: Time: 10 minutes; Temperature: Room temperature (25℃); Detergent: Enzyme-based detergent, 50ml (for local application); Main wash stage: Water temperature: 75℃ (strict principle, meeting S1 requirements); Time: 25 minutes (strict principle, meeting S1 requirements); Mechanical strength: Medium; Detergent: Alkaline detergent, 120ml; Additive: Blood stain remover enhancer, 30ml; Rinsing stage: Number of rinses: 3 (strict principle, meeting S1 requirements); Time per rinse: 8 minutes; Water temperature: 60℃→45℃→30℃ (strict principle, meeting S1 requirements); Disinfection stage: Method: Heat disinfection (strict principle, meeting S1 requirements); Temperature: 70℃; Time: 10 minutes; Disinfectant: None (heat disinfection does not require chemical disinfectants); Dehydration stage: Spin speed: 800rpm; Time: 5 minutes.

[0069] refer to Figure 5 In step S14, the specific steps are as follows: S141: Collect multiple washing parameters and the morphology of the medical fabric during the washing process. Based on the multiple washing parameters and the morphology of the medical fabric during the washing process, determine multiple washing areas of the medical fabric. In each washing area, determine the corresponding washing parameter combination based on the identification of each washing area. Based on the identification of the washing parameter combination, determine the washing effect level of the medical fabric in that washing area, so as to output the washing effect level of the medical fabric at each position. S142: Collect future usage records of medical textiles, determine the next usage event of medical textiles based on the identification of future usage records of medical textiles, determine multiple usage scenario features based on the detection of the next usage event of medical textiles, and determine the next usage scenario of medical textiles based on multiple usage scenario features. S143: Based on the washing effect level of the medical fabric at each location and the next use scenario of the medical fabric, determine multiple parts of the medical fabric that are of concern, and construct the key washing area of ​​the medical fabric based on the multiple parts of the medical fabric that are of concern. In the embodiments of this application, the washing parameters are collected in real time using embedded sensors: water temperature (range: 20℃–90℃, accuracy ±0.5℃); drum speed (range: 0–1200rpm, accuracy ±10rpm); washing liquid pH value (range: 3–12, accuracy ±0.1); washing liquid turbidity (NTU value, range: 0–1000, accuracy ±1); detergent concentration (ml / L, accuracy ±0.1); washing time (second-level accuracy); sampling frequency: 1Hz (once per second).

[0070] Morphological acquisition: A high-speed industrial camera (100fps) was used to capture the movement of the fabric in the drum; image processing algorithms were used to analyze the fabric's: degree of unfolding (0–100%); folding state (number of layers, folding angle); motion trajectory (speed, direction); surface tension distribution; acquisition environment: standard light source (D65 light source), neutral gray background.

[0071] Based on the collected washing parameters and morphological data, the system divides the fabric into multiple washing regions: using a grid-based method, the fabric surface is divided into basic units of 10mm×10mm; the natural fold lines and unfolding boundaries of the fabric are identified based on the morphological data; combined with the distribution of washing parameters, regions with different washing intensities are identified; and a clustering algorithm is used to merge adjacent basic units with similar washing characteristics into washing regions. Each washing region includes: region boundary coordinates; region area; average washing intensity; and morphological change characteristics.

[0072] For each washing zone, the system analyzes the combination of washing parameters it experiences: extracts key parameters for each washing stage of the zone; calculates the time-weighted average of the parameters; analyzes the trend of parameter changes (increase, decrease, fluctuation); and generates a feature vector of parameter combinations. The parameter combinations include: average water temperature; average rotation speed; average pH value; average turbidity change rate; detergent contact time; and mechanical action intensity.

[0073] Specifically, the washing parameters were collected as follows: Pre-wash stage: water temperature 30℃, time 5 minutes, spin speed 300 rpm, pH value 7.2, turbidity 150 NTU; Main wash stage: water temperature 60℃, time 15 minutes, spin speed 500 rpm, pH value 10.5, turbidity 450 NTU; Rinse stage: water temperature successively 60℃, 45℃, and 30℃, 8 minutes each, spin speed 400 rpm, pH value gradually decreasing to 7.0; Disinfection stage: water temperature 70℃, time 10 minutes, spin speed 300 rpm, pH value 8.5; Dehydration stage: spin speed 800 rpm, time 5 minutes.

[0074] Morphological sampling: Pre-wash stage: The fabric is tightly rolled up with an unfolding degree of 30% and an average number of folds of 5; Main wash stage: The fabric gradually unfolds, with the unfolding degree increasing to 70% and the number of folds decreasing to 2; Rinse stage: The fabric is fully unfolded with an unfolding degree of 95% and almost no folds; Disinfection stage: The fabric remains unfolded with an unfolding degree of 90%; Dehydration stage: The fabric adheres tightly to the drum wall, and the unfolding degree decreases to 40%.

[0075] Based on the washing data of the surgical gowns mentioned above, the system identified the following washing areas: Area 1 (High-intensity washing area): Location: Central chest area; Coordinate range: x150-200mm, y120-170mm; Area: 2500mm²; Average washing intensity: 85%; Morphological characteristics: Fully unfolded during the main wash stage, with sufficient contact with the detergent; Area 2 (Medium-intensity washing area): Location: Cuff area; Coordinate range: x50-100mm, y200-250mm (left sleeve), x250-300mm, y200-250mm (right sleeve); Area: 1000mm² each; Average washing intensity: 60%; Morphological characteristics: Partially unfolded during the main wash stage, with slight folds; Area 3 (Low-intensity washing area): Location: Neckline area; Coordinate range: x100-250mm, y50-100mm; Area: 3750mm²; Average washing intensity: 40%; Morphological characteristics: Remains folded during most of the washing stages.

[0076] Washing parameters for Zone 1: Average water temperature: 58.5℃; Average spin speed: 480 rpm; Average pH: 9.8; Average turbidity change rate: -30 NTU / min; Detergent contact time: 25 minutes; Mechanical action intensity: High (continuous high-intensity friction); Washing parameters for Zone 2: Average water temperature: 55.2℃; Average spin speed: 420 rpm; Average pH: 9.2; Average turbidity change rate: -25 NTU / min; Detergent contact time: 20 minutes; Mechanical action intensity: Medium (intermittent friction); Washing parameters for Zone 3: Average water temperature: 48.6℃; Average spin speed: 350 rpm; Average pH: 8.5; Average turbidity change rate: -15 NTU / min; Detergent contact time: 15 minutes; Mechanical action intensity: Low (slight friction).

[0077] Based on the combination of washing parameters for each region, the system evaluates the washing effect using a multi-parameter evaluation model, comprehensively considering: stain removal rate (calculated by comparing images before and after washing); fabric damage level (through fiber structure analysis); color fastness retention rate (through color measurement); and bacterial removal rate (through microbial detection). A 5-level rating system is adopted: Level 5: Excellent (removal rate >95%, damage <5%); Level 4: Good (removal rate 85-95%, damage 5-10%); Level 3: Moderate (removal rate 70-85%, damage 10-15%); Level 2: Poor (removal rate 50-70%, damage 15-20%); Level 1: Poor (removal rate <50%, damage >20%). A washing effect level distribution map is generated.

[0078] Specifically, based on the above parameter combinations, the system evaluates the washing effect of each area: Washing effect of Area 1: Stain removal rate: 96%; Fabric damage: 4%; Color fastness retention: 98%; Bacterial removal rate: 99.9%; Overall score: 4.8 points; Washing effect level: 5 (Excellent); Washing effect of Area 2: Stain removal rate: 88%; Fabric damage: 8%; Color fastness retention: 95%; Bacterial removal rate: 99.5%; Overall score: 4.2 points; Washing effect level: 4 (Good); Washing effect of Area 3: Stain removal rate: 75%; Fabric damage: 12%; Color fastness retention: 92%; Bacterial removal rate: 98%; Overall score: 3.5 points; Washing effect level: 3 (Medium).

[0079] The system generates a detailed washing performance rating report, including: a washing performance rating distribution map (using a heatmap to display the washing performance rating at each location); color coding: Level 5 (dark green), Level 4 (light green), Level 3 (yellow), Level 2 (orange), Level 1 (red); structured data: boundary coordinates of each washing area; corresponding washing parameter combinations; washing performance rating and scoring details; improvement suggestions (e.g., increasing washing time or water temperature for Level 3 areas); statistical summary: overall average washing performance rating: 4.1; percentage of excellent areas: 33%; percentage of good areas: 33%; percentage of average areas: 34%; area requiring improvement: collar area (suggestion to optimize folding).

[0080] Furthermore, collect future usage records for medical textiles, including: scheduled usage time (accurate to the minute); department using the textiles (e.g., surgery, internal medicine, ICU); type of use (surgery, nursing, examination, etc.); personnel using the textiles (surgeon, nurse, etc.); location of use (operating room number, ward number, etc.); and special requirements (e.g., sterility level, anti-static requirements, etc.).

[0081] The system uses a rule engine to match usage records; employs a machine learning model to predict event types; considers time weights (events that are closer together have higher weights); and classifies events as follows: surgical events (routine surgery, minimally invasive surgery, open surgery, etc.); nursing events (routine nursing, special nursing, etc.); examination events (CT scans, MRI scans, etc.); and emergency events (emergency surgery, resuscitation, etc.). Event attribute extraction includes: duration prediction; risk level assessment; pollution risk prediction; and comfort requirement level.

[0082] Specifically, taking a surgical gown with the serial number "SM-2023-A001" as an example: Future usage record collection: Appointment time: January 15, 2024, 09:30; Department using: Surgical operating room; Type of use: Cholecystectomy; Surgeon: Dr. A; Operating room: Operating room 3; Special requirements: Requires sterile packaging and anti-static treatment; Next usage event determination: Event type: Minimally invasive surgical event; Predicted duration: 2.5 hours; Risk level: Medium (Cholecystectomy); Contamination risk: Medium risk of contact with blood and body fluids; Comfort requirement: High (required for long-term surgery).

[0083] Multiple usage scenario characteristics are determined based on the detection of the next usage event. The scenario feature extraction methods are: feature mapping based on knowledge base; statistical analysis of historical data; rule reasoning of expert system. The main feature categories are: environmental features: temperature and humidity requirements; cleanliness level; electromagnetic environment requirements; operational features: range of motion requirements; bending and stretching frequency; contact object type; safety features: sterility requirement level; protection requirement level; antistatic requirements; functional features: breathability requirements; moisture absorption requirements; strength requirements.

[0084] The next use scenario for medical textiles is determined based on multiple usage scenario characteristics using multidimensional feature space mapping; a clustering algorithm is used to determine the scenario type; scenario similarity is calculated; scenario classification includes: high-risk surgical scenarios; routine surgical scenarios; special examination scenarios; and emergency rescue scenarios; scenario description elements include: scenario name; scenario type; key feature combination; risk level; and a list of special requirements. Specifically, the scene characteristics extracted based on minimally invasive surgical events are as follows: Environmental characteristics: Temperature requirement: 22-24℃; Humidity requirement: 45-55%; Cleanliness: Class 10,000 clean operating room; Electromagnetic environment: Electrosurgical equipment is present, requiring anti-static measures; Operational characteristics: Range of motion: Moderate (mainly upper limb movements); Bending frequency: Low (2-3 times per minute); Contact objects: Surgical instruments, patient's bodily fluids; Safety characteristics: Sterility requirement: High level (requires sterile packaging); Protection requirement: Moderate (prevents bodily fluid penetration); Antistatic requirement: High (prevents electrosurgical interference); Functional characteristics: Breathability: High (for prolonged wear); Moisture absorption: Moderate (for moderate perspiration); Strength requirement: High (requires tensile strength).

[0085] Based on the extracted features, the following usage scenarios are constructed: Scenario Name: Standard scenario for minimally invasive gallbladder surgery; Scenario Type: Routine surgical scenario; Risk Level: Medium risk; Key Feature Combinations: Environmental Control: Precise temperature and humidity control; Operational Requirements: Primarily upper limb movement; Safety Requirements: Sterility + Antistatic; Functional Requirements: High breathability + High strength; Special Requirements List: Requires sterile treatment; Must have antistatic function; Collar and cuffs need enhanced sealing; Material needs to have good breathability; Requires a certain degree of moisture absorption.

[0086] Therefore, the washing effect level data includes indicators such as cleanliness, residue, and fiber damage in various areas of the fabric, with each area represented by a level (e.g., 1-5, with 5 being the best). The next usage scenario characteristics include information such as the usage environment, contact objects, activity range, functional requirements, and safety level. The system matches key parts in the usage scenario (e.g., the front, cuffs, and collar of surgical gowns) with the washing effect level to determine which areas are "areas of concern." Weighting mechanism: If an area has a low washing effect level (e.g., level 2) but high importance in the usage scenario (e.g., the front of the gown comes into contact with blood), its attention weight is increased; if an area has a high washing effect level (e.g., level 5) but low importance in the usage scenario (e.g., the non-contact area on the back), its weight is decreased. Threshold determination: The system sets an attention threshold; areas exceeding this value are marked as "areas of concern."

[0087] The system constructs key washing areas for medical fabrics based on multiple areas of interest. Based on the location coordinates of these areas, the system expands outwards by a certain range (e.g., 5cm forward and backward) to form a complete washing area. The key washing areas are prioritized according to the importance weight of each area. Each key washing area is labeled with the following information: area number; boundary coordinates; priority; washing effect requirements; usage scenario requirements; and monitoring indicators. Based on the characteristics of the key washing areas, the system generates targeted washing optimization suggestions: for high-priority areas, it suggests increasing washing intensity or extending washing time; for areas with residual stains, it suggests using a special detergent; and for areas with wear, it suggests reducing mechanical strength.

[0088] Specifically, taking a surgical gown with the serial number "SM-2023-A001" as an example: Washing effect level data: Front area: Level 2 (slight stains remain); Cuff area: Level 3 (basically clean, but with slight wear); Collar area: Level 4 (good cleanliness); Back area: Level 5 (completely clean); Next use scenario characteristics: Use type: cholecystectomy; Key contact areas: Front (contact with blood), Cuffs (frequent operation); Functional requirements: High protection, high comfort; Safety requirements: Sterile, anti-static.

[0089] Process for determining the areas of concern: Front placket area: Washing grade 2 + High importance of use → Attention weight 90%; Cuff area: Washing grade 3 + High importance of use → Attention weight 75%; Neckline area: Washing grade 4 + Medium importance of use → Attention weight 40%; Back area: Washing grade 5 + Low importance of use → Attention weight 10%; Output result: List of areas of concern: Front placket area (priority 1); Cuff area (priority 2).

[0090] Taking the surgical gown "SM-2023-A001" as an example: Areas of concern: Front placket area (Priority 1); Cuff area (Priority 2); Key washing area construction: Front placket key area: Area number: KR-001; Boundary coordinates: 5cm below the neckline to 10cm above the hem, 5cm on each side; Priority: 1; Washing effect requirement: Level 5 (Completely clean); Usage scenario requirement: High protection; Monitoring indicators: Stains remaining, fiber integrity; Cuff key area: Area number: KR-002; Boundary coordinates: 15cm inward from the cuff; Priority: 2; Washing effect requirement: Level 4 (Good clean); Usage scenario requirement: High flexibility; Monitoring indicators: Wear level, elasticity retention rate; Washing optimization suggestions: Front placket area: It is recommended to increase the pre-wash time by 2 minutes; Use a special blood stain remover; Increase the washing temperature to 60℃; Cuff area: It is recommended to reduce the machine speed by 20%; Use fabric softener; Maintain the washing temperature at 40℃.

[0091] refer to Figure 6 In step S15, the specific steps are as follows: S151: Collect the key washing area of ​​the medical fabric, determine multiple sub-key washing areas of the medical fabric based on the key washing area and various functional parts of the medical fabric, determine the corresponding cleanliness based on the detection of multiple sub-key washing areas, and output multiple cleanliness levels of the key washing area, with each cleanliness level marked at a different position on the medical fabric. S152: Multiple cleanliness levels are sequentially presented in the key washing areas of medical fabrics, and the difference between two adjacent cleanliness levels is determined based on the comparison between two adjacent cleanliness levels; the first washing sequence is determined based on the location of the multiple cleanliness levels and the regional morphology of the key washing areas. S153: Determine the second washing sequence based on the location of multiple cleanliness levels and the difference between two adjacent cleanliness levels, and determine the gradient washing method of medical fabrics according to the first washing sequence, the second washing sequence, and multiple sub-key washing areas.

[0092] In the embodiments of this application, key washing areas of medical fabrics are collected, and the different functional parts contained in each key washing area are systematically analyzed. The functional parts are identified based on: structural features (such as sutures, edges, folds, etc.); usage features (such as high-frequency contact areas, friction areas, load-bearing areas, etc.); and material features (such as the joints of different materials). A grid subdivision method is used to divide the key washing areas into basic units. The unit size is determined according to the fabric precision requirements (usually 1cm×1cm or 2cm×2cm). The principle for merging adjacent functional units is: the same or similar functions; the same contamination risk level; and the same washing process requirements. Sub-region attribute labeling: a unique identifier is assigned to each sub-key washing area. The following are recorded for each sub-region: precise boundary coordinates; functional attributes; material attributes; contamination risk level; and washing sensitivity.

[0093] Cleanliness testing of key washing areas utilizes a multispectral imaging system: visible light imaging (400-700nm): detecting surface stains; near-infrared imaging (700-1000nm): detecting organic residues; ultraviolet imaging (200-400nm): detecting microbial contamination. The testing process includes: Step 1: Image preprocessing: noise reduction; illumination homogenization; geometric correction; Step 2: Feature extraction: color feature analysis; texture feature analysis; morphological feature analysis; Step 3: Cleanliness assessment: establishing a cleanliness assessment model; comprehensive score calculation; grade classification. Further, the cleanliness grade standards are: Grade 5 (Excellent): No visible stains; microbial residue <10 CFU / 100cm²; chemical residue <0.1mg / cm²; Grade 4 (Good): Slight traces but not affecting use; microbial residue 10-50 CFU / 100cm²; chemical residue 0.1-0.5mg / cm²; Grade 3 (Acceptable): Slight visible stains; microbial residue 50-100mg / cm². CFU / 100cm²; Chemical residue 0.5-1mg / cm²; Grade 2 (requires treatment): Obvious stains; Microbial residue 100-500 CFU / 100cm²; Chemical residue 1-2mg / cm²; Grade 1 (unacceptable): Severe stains; Microbial residue >500 CFU / 100cm²; Chemical residue >2mg / cm².

[0094] Furthermore, multiple cleanliness levels are sequentially presented in the key washing areas of the medical fabric, and the difference between two adjacent cleanliness levels is determined based on the comparison between the two adjacent cleanliness levels. A first washing sequence is determined based on the location of the multiple cleanliness levels and the regional morphology of the key washing areas, and a second washing sequence is determined based on the location of the multiple cleanliness levels and the difference between two adjacent cleanliness levels. A gradient washing method for the medical fabric is determined based on the first washing sequence, the second washing sequence, and multiple sub-key washing areas, which takes into account the overall consideration of the first washing sequence, the second washing sequence, and multiple sub-key washing areas, ensuring the accuracy of the gradient washing method for the medical fabric.

[0095] At this point, the system obtains the cleanliness data of all sub-critical washing areas from step S151. The data structure includes: sub-area ID; cleanliness level (level 1-5); center coordinates (x, y); detection details (stain type, residue, number of microorganisms, etc.).

[0096] Cleanliness visualization: On the digital model of medical fabrics, the system displays the cleanliness of each sub-area in sequence according to spatial location; Display method: Color coding: Gradient color from level 1 (red) to level 5 (green); Numerical label: The cleanliness level is displayed in the center of each sub-area; Heat map: The overall cleanliness distribution is displayed; Data sorting rules: The sub-areas are arranged in order from left to right and from top to bottom; For overlapping or boundary areas, a weighted average method is used to determine the display order.

[0097] Neighboring region identification: An 8-neighborhood algorithm is used to identify neighboring regions of each sub-region; for edge regions, only actually existing neighboring regions are considered; Difference calculation: Absolute difference: |Cleanliness A - Cleanliness B|; Relative difference: |Cleanliness A - Cleanliness B| / max(Cleanliness A, Cleanliness B); Total difference = 0.7 × Absolute difference + 0.3 × Relative difference; Difference classification: Minor difference: ≤0.5 level; Moderate difference: 0.6-1.5 level; Significant difference: ≥1.6 level; Difference storage: A difference matrix is ​​established to record the difference of each pair of neighboring regions; it includes difference value, difference type, and direction information.

[0098] Determine the first washing order (based on position and area morphology): Position factor analysis: Spatial position weight: Upper area weight > Lower area weight (considering gravity); Central area weight > Edge area weight (considering importance); Weight calculation formula: W_position = 0.4 × height weight + 0.6 × centrality weight; Area morphology analysis: Shape complexity: Calculate the perimeter-to-area ratio of the area; Concavity / convexity: Identify the concave and convex parts of the area; Connectivity: Analyze the degree of connection between the area and other areas; Order generation algorithm: Use a greedy algorithm, selecting the area with the highest weight each time; Consider the spatial continuity between areas; Output the washing order list.

[0099] Determine the second washing order (based on position and difference): Comprehensive weight calculation: position weight (same as the first order); difference weight: areas with high difference receive higher priority; weight calculation: W_diff = difference × 2.0; comprehensive score: Score = 0.5 × W_position + 0.5 × W_diff; normalize the score; order optimization: consider the movement path optimization of the washing equipment; avoid frequent changes in direction; generate the optimal washing path.

[0100] Washing Method Decision: Sequential Integration: Compare the first and second priorities; determine the final priority using a weighted voting method; Weight Allocation: 60% for the first priority, 40% for the second priority; Gradient Parameter Setting: Set washing intensity according to cleanliness level: Level 1 area: high-intensity wash (water temperature 60℃, spin speed 800rpm); Levels 2-3 area: medium-intensity wash (water temperature 45℃, spin speed 600rpm); Levels 4-5 area: low-intensity wash (water temperature 30℃, spin speed 400rpm); Transition Treatment: Set transition zones between different intensity zones; the transition zones use gradual parameter adjustments; the transition time is dynamically adjusted according to the size of the zone; in this gradient washing method, prioritize the treatment of the dirtiest areas; use gentle treatment for areas that are already relatively clean; optimize overall washing efficiency to ensure that all areas reach the appropriate cleanliness level.

[0101] Please see Figure 7 , Figure 7 This is a schematic diagram of the structural composition of a vision-based medical fabric washing control system according to an embodiment of the present invention; the vision-based medical fabric washing control system includes: The preliminary stain area module 21 is used to determine the usage record of the medical fabric based on the model of the medical fabric, and to determine the preliminary stain area based on the usage record of the medical fabric. The multimodal data module 22 is used to determine the visual inspection pattern of the medical fabric based on the initial stain area, the type of medical fabric, and the area of ​​the medical fabric used; and to determine multiple sub-visual inspection items based on the recognition of the visual inspection pattern. The washing parameter module 23 is used to determine multiple actual stained parts of the medical fabric based on multiple sub-visual detection items, front and back images of the medical fabric, and preliminary stained areas; and to determine multiple washing parameters of the medical fabric during the washing process based on multiple actual stained parts, the type of medical fabric, and the washing mode of the washing equipment. The critical washing area module 24 is used to determine the washing effect level of the medical fabric at each location based on multiple washing parameters, and to determine the critical washing area of ​​the medical fabric based on the washing effect level of the medical fabric at each location and the next use scenario of the medical fabric. The gradient washing mode module 25 is used to determine multiple cleanliness levels based on the detection of key washing areas of medical fabrics, and to determine the difference between two adjacent cleanliness levels. The gradient washing mode of the medical fabric is determined based on the location of the multiple cleanliness levels, the regional morphology of the key washing areas, and the difference between two adjacent cleanliness levels.

[0102] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A method for controlling the washing of medical fabrics based on visual detection, characterized in that, include: The usage records of medical fabrics are determined based on their model, and the initial stained areas are determined based on the usage records of the medical fabrics. The visual inspection mode for medical fabrics is determined based on the initial stained area, the type of medical fabric, and the area of ​​medical fabric used. Multiple sub-visual detection items are determined based on the recognition of this visual detection pattern; Visual inspection modes include: Basic Mode, Standard Mode, Fine Mode, and Professional Mode; Based on multiple sub-visual inspection items, front and back images of the medical fabric, and preliminary stain areas, multiple actual stain portions of the medical fabric are determined; based on multiple actual stain portions, the type of medical fabric, and the washing mode of the washing equipment, multiple washing parameters of the medical fabric during the washing process are determined; the multiple washing parameters include pretreatment parameters, main wash parameters, rinsing parameters, disinfection parameters, and dehydration parameters. The washing effect level of medical fabric at each location is determined based on multiple washing parameters, and the key washing areas of medical fabric are determined based on the washing effect level of medical fabric at each location and the next use scenario of medical fabric. Multiple cleanliness levels are determined based on the detection of key washing areas of medical fabrics, and the difference between two adjacent cleanliness levels is determined. A gradient washing method for medical fabrics is determined based on the location of multiple cleanliness levels, the regional morphology of key washing areas, and the difference between two adjacent cleanliness levels. In this gradient washing method, the dirtiest areas are treated first, and areas that are already relatively clean are treated gently. The overall washing efficiency is optimized to ensure that all areas reach the appropriate cleanliness level.

2. The washing control method for medical fabrics based on vision detection according to claim 1, characterized in that, The process of determining the usage records of medical fabrics based on their model, and identifying the initial stained areas based on these records, includes: The model of the medical fabric is collected, and the past use events of the medical fabric are determined based on the matching of the model of the medical fabric and the medical fabric database. Multiple sub-use events are determined based on the detection of the past use events of the medical fabric, and the use records of the medical fabric are determined based on the identification of the multiple sub-use events. Based on the identification of medical fabric usage records, the stained areas of medical fabrics during each use are determined, and multiple previous stained locations of medical fabrics are marked. The initial stained areas are determined based on the multiple previous stained locations of medical fabrics.

3. The washing control method for medical fabrics based on visual detection according to claim 1, characterized in that, The visual inspection mode of the medical fabric is determined based on the initial stain area, the type of medical fabric, and the area of ​​the medical fabric used. Based on the recognition of this visual detection pattern, multiple sub-visual detection items are determined, including: The current image of the medical fabric is acquired by taking pictures of the medical fabric. The surface texture of the medical fabric is determined by texture recognition of the current image of the medical fabric. The type of medical fabric is determined by the surface texture and model of the medical fabric. The area of ​​use of the medical fabric is determined by area recognition of the current image of the medical fabric. The first detection coefficient is determined based on the initial stain area and the type of medical fabric. The second detection coefficient is determined based on the initial stain area and the area of ​​use of the medical fabric. The visual detection mode of the medical fabric is determined based on the mapping relationship between the first detection coefficient, the second detection coefficient and the detection mode. The visual inspection patterns of medical fabrics are collected, and a corresponding visual inspection procedure table is determined based on the recognition of the visual inspection patterns of medical fabrics. Multiple sub-visual inspection items are determined according to the recognition of the visual inspection procedure table. The inspection content corresponding to each visual inspection item has a different priority. Multiple sub-visual inspection items perform orderly visual inspection of medical fabrics.

4. The washing control method for medical fabrics based on visual detection according to claim 1, characterized in that, The method involves determining multiple actual stained portions of the medical fabric based on multiple sub-visual detection items, front and back images of the medical fabric, and preliminary stained areas. Based on multiple actual stained areas, the type of medical fabric, and the washing mode of the washing equipment, several washing parameters for the medical fabric during the washing process are determined, including: Acquire the current image of the medical fabric, determine the front and back images of the medical fabric based on the division of the current image of the medical fabric, determine multiple stain surfaces based on the visual inspection of the front and back images of the medical fabric, and determine the first stain combination based on multiple sub-visual inspection items and multiple stain surfaces. A second stain combination is determined based on multiple stained surfaces and initial stained areas. Multiple actual stained portions of the medical fabric are determined based on the matching of the first stain combination and the second stain combination, and the location and stain area of ​​the multiple actual stained portions are marked.

5. The washing control method for medical fabrics based on visual detection according to claim 4, characterized in that, The method involves determining multiple actual stained portions of the medical fabric based on multiple sub-visual detection items, front and back images of the medical fabric, and preliminary stained areas. Based on multiple actual stained areas, the type of medical fabric, and the washing mode of the washing equipment, several washing parameters for medical fabrics during the washing process are determined, including: The system collects information on the types of medical fabrics and determines the washing mode of the washing equipment based on multiple operating parameters. It then determines the washing content corresponding to the multiple actual stained parts based on the types of medical fabrics and the washing mode of the washing equipment. Based on the identification of the washing content corresponding to the multiple actual stained parts, it determines the corresponding washing parameters to output multiple washing parameters of the medical fabrics during the washing process.

6. The washing control method for medical fabrics based on visual detection according to claim 1, characterized in that, The process involves determining the washing effect level of the medical fabric at various locations based on multiple washing parameters, and determining the key washing areas of the medical fabric based on the washing effect level at each location and the next usage scenario of the medical fabric, including: Multiple washing parameters and the morphology of the medical fabric during the washing process are collected. Based on the multiple washing parameters and the morphology of the medical fabric during the washing process, multiple washing areas of the medical fabric are determined. In each washing area, the corresponding combination of washing parameters is determined based on the identification of each washing area. Based on the identification of the washing parameter combination, the washing effect level of the medical fabric in that washing area is determined, so as to output the washing effect level of the medical fabric at each location.

7. The washing control method for medical fabrics based on visual detection according to claim 6, characterized in that, The process of determining the washing effect level of medical fabric at various locations based on multiple washing parameters, and determining the key washing areas of medical fabric based on the washing effect level of medical fabric at various locations and the next use scenario of medical fabric, further includes: Collect future usage records of medical textiles, determine the next usage event of medical textiles based on the identification of future usage records, determine multiple usage scenario features based on the detection of the next usage event of medical textiles, and determine the next usage scenario of medical textiles based on multiple usage scenario features. Based on the washing effect level of the medical fabric at various locations and the next use scenario of the medical fabric, multiple parts of the medical fabric of concern are identified, and key washing areas of the medical fabric are constructed based on these multiple parts of concern.

8. The washing control method for medical fabrics based on visual detection according to claim 1, characterized in that, The method involves determining multiple cleanliness levels based on the detection of key washing areas of medical fabrics, and determining the difference between two adjacent cleanliness levels. A gradient washing method for the medical fabrics is then determined based on the location of the multiple cleanliness levels, the morphology of the key washing areas, and the difference between two adjacent cleanliness levels, including: The key washing areas of the medical fabric are collected. Based on the key washing areas and various functional parts of the medical fabric, multiple sub-key washing areas of the medical fabric are determined. The corresponding cleanliness is determined based on the detection of multiple sub-key washing areas, so as to output multiple cleanliness levels of the key washing areas. Each cleanliness level is marked at a different position on the medical fabric.

9. The washing control method for medical fabrics based on visual detection according to claim 8, characterized in that, The method of determining multiple cleanliness levels based on the detection of key washing areas of medical fabrics, determining the difference between two adjacent cleanliness levels, and determining a gradient washing method for medical fabrics based on the location of multiple cleanliness levels, the regional morphology of the key washing areas, and the difference between two adjacent cleanliness levels, further includes: Multiple cleanliness levels are sequentially presented in the key washing areas of medical fabrics, and the difference between two adjacent cleanliness levels is determined by comparing the two adjacent cleanliness levels; the first washing sequence is determined based on the location of the multiple cleanliness levels and the regional morphology of the key washing areas. The second washing sequence is determined based on the location of multiple cleanliness levels and the difference between two adjacent cleanliness levels. The gradient washing method for medical fabrics is determined based on the first washing sequence, the second washing sequence, and multiple sub-key washing areas.

10. A washing control system for medical fabrics based on vision detection, characterized in that, The vision-based medical fabric washing control system is applied to the vision-based medical fabric washing control method as described in any one of claims 1-9, wherein the vision-based medical fabric washing control system comprises: The preliminary stain area module is used to determine the usage record of medical fabrics based on the model of the medical fabrics, and to determine the preliminary stain area based on the usage record of the medical fabrics. The sub-visual inspection item module is used to determine the visual inspection pattern of the medical fabric based on the initial stain area, the type of medical fabric, and the area of ​​the medical fabric used; and to determine multiple sub-visual inspection items based on the recognition of the visual inspection pattern. The washing parameter module is used to determine multiple actual stained parts of the medical fabric based on multiple sub-visual detection items, front and back images of the medical fabric, and preliminary stained areas; and to determine multiple washing parameters of the medical fabric during the washing process based on multiple actual stained parts, the type of medical fabric, and the washing mode of the washing equipment. The critical washing area module is used to determine the washing effect level of medical fabrics at various locations based on multiple washing parameters, and to determine the critical washing areas of medical fabrics based on the washing effect level of medical fabrics at various locations and the next use scenario of medical fabrics. The gradient washing mode module is used to determine multiple cleanliness levels based on the detection of key washing areas of medical fabrics, and to determine the difference between two adjacent cleanliness levels. The gradient washing mode of the medical fabric is determined based on the location of multiple cleanliness levels, the regional morphology of the key washing areas, and the difference between two adjacent cleanliness levels.

Images