Intelligent detection system for cleaning state of tracheal cannula medical instrument

Abstract

The invention relates to the technical field of tracheal cannulas, in particular to a tracheal cannula medical instrument cleaning state intelligent detection system which comprises the steps that a first outer wall image set and a first inner wall image set of a cleaned tracheal cannula are acquired through an image acquisition module; the image recognition module recognizes a first outer wall image set and a first inner wall image set by constructing a first image recognition model, extracts outer wall image texture features and inner wall image texture features, obtains a first outer wall pollution coefficient and a second inner wall pollution coefficient, and further obtains a comprehensive pollution coefficient; the cleaning state analysis module evaluates a first cleaning state according to the comprehensive pollution coefficient and a preset threshold value; the cleaning judgment module judges whether further cleaning is needed or not, and if yes, second continuous cleaning time is obtained by combining the initial comprehensive pollution coefficient, the comprehensive pollution coefficient, a preset threshold value and the first cleaning time; and if not, cleaning is ended. The intelligent detection capability of the cleaning state of the trachea cannula can be effectively improved, so that the medical safety is ensured.

Description

Technical Field

[0001] This invention relates to the field of tracheostomy tube technology, specifically to an intelligent detection system for the cleaning status of tracheostomy tube medical devices. Background Technology

[0002] Metal endotracheal cannulas mainly consist of an outer tube and an inner tube, and are widely used in hospitals. If the outer and inner tubes of metal endotracheal cannulas are not cleaned promptly after use, they can easily lead to infection risks, thereby affecting medical safety. Therefore, effective monitoring of the cleaning status of metal endotracheal cannulas is of great significance for ensuring medical safety.

[0003] Traditional methods for detecting the cleaning status of metal tracheal cannulas do not fully consider the comprehensive analysis of the outer and inner tubes of the tracheal cannulas and lack intelligent detection capabilities, which limits the accuracy of detecting the cleaning status of the tracheal cannulas.

[0004] To address this, an intelligent detection system for the cleaning status of tracheostomy tube medical devices is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent detection system for the cleaning status of tracheostomy cannulas. This invention relates to the field of tracheostomy cannulas technology, specifically an intelligent detection system for the cleaning status of tracheostomy cannulas, comprising: acquiring a first set of images of the outer wall and a first set of images of the inner wall of the cleaned tracheostomy cannulas through an image acquisition module; an image recognition module recognizing the first set of images of the outer wall and the first set of images of the inner wall by constructing a first image recognition model, extracting texture features of the outer wall and inner wall images to obtain a first outer wall contamination coefficient and a second inner wall contamination coefficient, and thus obtaining a comprehensive contamination coefficient; a cleaning status analysis module evaluating a first cleaning status based on the comprehensive contamination coefficient and a preset threshold; and a cleaning judgment module determining whether further cleaning is needed. If so, a second cleaning time is obtained by combining the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time; otherwise, the cleaning is terminated. This invention can effectively improve the intelligent detection capability of the cleanliness status of the tracheostomy cannulas, thereby ensuring medical safety.

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

[0007] A smart detection system for the cleaning status of a tracheostomy cannula medical device includes:

[0008] The image acquisition module acquires images of the outer wall of the outer tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube after cleaning; thereby obtaining the first set of images of the outer wall and the first set of images of the inner wall of the cleaned tracheostomy tube.

[0009] The image recognition module constructs a first image recognition model to recognize a first outer wall image set and a first inner wall image set. The first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer. The preprocessing layer preprocesses the first outer wall image set and the first inner wall image set to obtain a second outer wall image set and a second inner wall image set. The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set to obtain texture features of the outer wall image and the inner wall image. The image feature analysis obtains the first outer wall pollution coefficient and the second inner wall pollution coefficient based on the texture features of the outer wall image and the inner wall image, respectively. Finally, a comprehensive pollution coefficient is obtained based on the first outer wall pollution coefficient and the second inner wall pollution coefficient.

[0010] The cleaning status analysis module is used to determine the first cleaning status based on the comprehensive contamination coefficient and a preset threshold.

[0011] The cleaning judgment module is used to determine whether cleaning needs to continue based on the first cleaning state. If not, the cleaning ends; if so, the second cleaning time is obtained based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time.

[0012] Preferably, the first outer wall image set consists of the outer wall image of the outer tube and the outer wall image of the inner tube; the first inner wall image set consists of the inner wall image of the outer tube and the inner wall image of the inner tube.

[0013] Preferably, the first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer;

[0014] The preprocessing layer performs preprocessing operations on the first outer wall image set and the first inner wall image set to obtain the second outer wall image set and the second inner wall image set; the preprocessing operations include image denoising, grayscale conversion and geometric correction.

[0015] The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set through a deep convolutional network to obtain the texture features of the outer wall image and the texture features of the inner wall image.

[0016] The image feature analysis is used to obtain the first outer wall contamination coefficient and the second inner wall contamination coefficient based on the texture features of the outer wall image and the texture features of the inner wall image, respectively; and to obtain the comprehensive contamination coefficient based on the first outer wall contamination coefficient and the second inner wall contamination coefficient.

[0017] The output layer is used to output the comprehensive pollution coefficient.

[0018] Preferably, the outer wall image texture features include outer wall contrast, outer wall brightness, and outer wall uniformity; each outer wall contaminant region is extracted based on the outer wall image texture features, and a first outer wall contamination coefficient is obtained based on the pixel area of ​​each outer wall contaminant region; the inner wall image texture features include inner wall contrast, inner wall brightness, and inner wall uniformity; each inner wall contaminant region is extracted based on the inner wall image texture features, and a second inner wall contamination coefficient is obtained based on the pixel area of ​​each inner wall contaminant region.

[0019] Preferably, the comprehensive pollution coefficient is obtained by weighting the first outer wall pollution coefficient and the second inner wall pollution coefficient.

[0020] Preferably, the first cleaning state includes a clean state and a contaminated state.

[0021] Preferably, the initial comprehensive pollution coefficient is obtained by acquiring images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube that have not been cleaned, thereby obtaining a first set of outer wall images and a first set of inner wall images of the tracheostomy tube that have not been cleaned, and inputting them into a first image recognition model for recognition; the first cleaning time represents the time that the tracheostomy tube has been cleaned.

[0022] Preferably, the second wash time is:

[0023] ;

[0024] in, Indicates the second wash time; Indicates the overall pollution coefficient; Indicates a preset threshold; This represents the initial comprehensive pollution coefficient; Indicates the first cleaning time; This represents the adjustment factor.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] 1. This invention acquires images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube to form a first outer wall image set and a first inner wall image set. It then uses a first image recognition model to identify the contaminant regions by extracting texture features from the outer and inner wall images using a deep convolutional network. This process yields the first outer wall contamination coefficient and the second inner wall contamination coefficient, ultimately resulting in a comprehensive contamination coefficient. The first image recognition model effectively identifies the contamination status of the inner and outer tubes of the tracheostomy tube, significantly improving the intelligent detection capability of the tracheostomy tube's cleaning status and thus effectively ensuring medical safety.

[0027] 2. This invention combines comprehensive identification of the inner and outer wall images of the tracheostomy tube with the inner and outer wall images of the tracheostomy tube to obtain the comprehensive contamination coefficient of the tracheostomy tube, providing model support for comprehensively improving the intelligent detection capability of the tracheostomy tube cleaning status, thereby effectively ensuring medical safety.

[0028] 3. Based on the comprehensive contamination coefficient obtained by the first image recognition model, and based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient and the first cleaning time, the second cleaning time is obtained. This provides effective data support for further precise cleaning of the tracheal cannula, effectively improves the intelligent detection capability of the cleaning status of the tracheal cannula, and thus effectively ensures medical safety. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the structure of an intelligent detection system for the cleaning status of a tracheostomy cannula medical device provided in an embodiment of the present invention;

[0030] Figure 2 This is a schematic diagram of the structure of a first image recognition model provided in an embodiment of the present invention;

[0031] Figure 3 This is a flowchart illustrating an intelligent detection system for the cleaning status of a tracheostomy cannula medical device, provided as an embodiment of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] Example 1

[0034] Tracheostomy tube A is a metal tracheostomy tube. In order to improve the intelligent detection capability of the cleanliness status of tracheostomy tube A and thus ensure medical safety, an intelligent detection system for the cleaning status of tracheostomy tube medical device was applied.

[0035] Reference Figure 1 A schematic diagram of a smart detection system for the cleaning status of a tracheostomy cannula medical device provided in an embodiment of the present invention includes:

[0036] The image acquisition module acquires images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube after cleaning; thereby obtaining the first set of images of the outer wall and the first set of images of the inner wall of the tracheostomy tube after cleaning; the outer wall images of the outer tube and the outer wall images of the inner tube are acquired by a high-resolution camera; the inner wall images of the outer tube and the inner wall images of the inner tube are acquired by a miniature endoscope.

[0037] Furthermore, the first outer wall image set consists of the outer wall image of the outer tube and the outer wall image of the inner tube; the first inner wall image set consists of the inner wall image of the outer tube and the inner wall image of the inner tube.

[0038] The image recognition module constructs a first image recognition model to recognize a first outer wall image set and a first inner wall image set. The first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer. The preprocessing layer preprocesses the first outer wall image set and the first inner wall image set to obtain a second outer wall image set and a second inner wall image set. The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set to obtain texture features of the outer wall image and the inner wall image. The image feature analysis obtains the first outer wall contamination coefficient and the second inner wall contamination coefficient based on the texture features of the outer wall image and the inner wall image, respectively. Finally, a comprehensive contamination coefficient is obtained based on the first outer wall contamination coefficient and the second inner wall contamination coefficient. Figure 2 This is a schematic diagram of the structure of a first image recognition model provided in an embodiment of the present invention;

[0039] Furthermore, the first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer;

[0040] The preprocessing layer performs preprocessing operations on the first outer wall image set and the first inner wall image set to obtain the second outer wall image set and the second inner wall image set; the preprocessing operations include image denoising, grayscale conversion and geometric correction.

[0041] The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set through a deep convolutional network to obtain the texture features of the outer wall image and the texture features of the inner wall image.

[0042] The image feature analysis is used to obtain the first outer wall contamination coefficient and the second inner wall contamination coefficient based on the texture features of the outer wall image and the texture features of the inner wall image, respectively; and to obtain the comprehensive contamination coefficient based on the first outer wall contamination coefficient and the second inner wall contamination coefficient.

[0043] The output layer is used to output the comprehensive pollution coefficient.

[0044] Furthermore, the outer wall image texture features include outer wall contrast, outer wall brightness, and outer wall uniformity; each outer wall contaminant region is extracted based on the outer wall image texture features, and a first outer wall contamination coefficient is obtained based on the pixel area of ​​each outer wall contaminant region; the inner wall image texture features include inner wall contrast, inner wall brightness, and inner wall uniformity; each inner wall contaminant region is extracted based on the inner wall image texture features, and a second inner wall contamination coefficient is obtained based on the pixel area of ​​each inner wall contaminant region;

[0045] Furthermore, the various contaminants on the outer wall include blood, sputum, drugs, and bacteria; the first outer wall contamination coefficient is obtained by quoting the sum of the pixel areas of each contaminant region on the outer wall with the total pixel area of ​​the outer wall image.

[0046] Furthermore, the first outer wall contamination coefficient is:

[0047] ;

[0048] in, Indicates the first outer wall contamination coefficient; Indicates the type of pollutant; Indicates the first The total pixel area of ​​each pollutant; This represents the total pixel area of ​​the outer wall image;

[0049] The various contaminants on the inner wall include blood, sputum, drugs, and bacteria; the second inner wall contamination coefficient is obtained by quoting the sum of the pixel areas of each contaminant region on the inner wall with the total pixel area of ​​the inner wall image.

[0050] The second inner wall contamination coefficient is:

[0051] ;

[0052] in, Indicates the second inner wall contamination coefficient; Indicates the type of pollutant; Indicates the first The total pixel area of ​​each pollutant; This represents the total pixel area of ​​the inner wall image;

[0053] Furthermore, the comprehensive pollution coefficient is obtained by weighting the first outer wall pollution coefficient and the second inner wall pollution coefficient, specifically as follows:

[0054] ;

[0055] in, Indicates the overall pollution coefficient; Indicates the weight of the first outer wall contamination coefficient; Indicates the first outer wall contamination coefficient; Indicates the weight of the second inner wall contamination coefficient; This represents the second inner wall contamination coefficient.

[0056] This embodiment acquires images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube to form a first outer wall image set and a first inner wall image set. These images are then recognized using a first image recognition model. A deep convolutional network is used to extract texture features from the outer and inner wall images, thereby extracting various contaminant regions and obtaining the first outer wall contamination coefficient and the second inner wall contamination coefficient, ultimately yielding a comprehensive contamination coefficient. The first image recognition model effectively identifies the contamination status of the inner and outer tubes of the tracheostomy tube, significantly improving the intelligent detection capability of the tracheostomy tube's cleaning status and thus effectively ensuring medical safety.

[0057] This embodiment combines comprehensive identification of the inner and outer wall images of the tracheostomy tube with the inner and outer wall images of the tracheostomy tube to obtain the comprehensive contamination coefficient of the tracheostomy tube. This provides model support for comprehensively improving the intelligent detection capability of the tracheostomy tube cleaning status, thereby effectively ensuring medical safety.

[0058] The cleaning status analysis module is used to determine the first cleaning status based on the comprehensive contamination coefficient and a preset threshold.

[0059] Furthermore, the first cleaning state includes a clean state and a contaminated state, and the specific judgment process is as follows:

[0060] ;

[0061] in, Indicates the first cleaning state; Indicates a clean state; Indicates a state of pollution; Indicates a preset threshold; Indicates the overall pollution coefficient;

[0062] The cleaning judgment module is used to determine whether cleaning needs to continue based on the first cleaning state. If not, the cleaning ends; if so, the second cleaning time is obtained based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time.

[0063] Furthermore, the initial comprehensive pollution coefficient is obtained by acquiring images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube that have not been cleaned, thereby obtaining a first set of outer wall images and a first set of inner wall images of the tracheostomy tube that have not been cleaned, and inputting them into a first image recognition model for recognition.

[0064] Furthermore, the second wash time is:

[0065] ;

[0066] in, Indicates the second wash time; Indicates the overall pollution coefficient; Indicates a preset threshold; This represents the initial comprehensive pollution coefficient; Indicates the first cleaning time; This represents the adjustment factor.

[0067] In summary, such as Figure 3 This is a flowchart illustrating an intelligent detection system for the cleaning status of a tracheostomy cannula medical device, provided in an embodiment of the present invention.

[0068] This embodiment uses the comprehensive contamination coefficient obtained by the first image recognition model, and the second cleaning time is obtained based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient and the first cleaning time. This provides effective data support for further precise cleaning of the tracheal cannula, effectively improves the intelligent detection capability of the cleaning status of the tracheal cannula, and thus effectively ensures medical safety.

[0069] This embodiment acquires a first set of images of the outer wall and a first set of images of the inner wall of the cleaned tracheostomy cannula through an image acquisition module; an image recognition module identifies the first set of images of the outer wall and the inner wall by constructing a first image recognition model, extracts the texture features of the outer wall and the inner wall images to obtain the first outer wall contamination coefficient and the second inner wall contamination coefficient, and then obtains the comprehensive contamination coefficient; a cleaning status analysis module evaluates the first cleaning status based on the comprehensive contamination coefficient and a preset threshold; a cleaning judgment module determines whether further cleaning is needed, and if so, it combines the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time to obtain the second cleaning time. This invention can effectively improve the intelligent detection capability of the tracheostomy cannula's cleanliness status, thereby ensuring medical safety.

[0070] To verify the effectiveness of the intelligent detection system for cleaning status of tracheostomy cannulas provided in this embodiment, the cleaning status of multiple A tracheostomy cannulas was detected by different systems. The average accuracy of the judgment on whether multiple A tracheostomy cannulas need to continue cleaning was calculated by statistical analysis, and the comparison results are shown in Table 1.

[0071] Table 1. Average accuracy of different systems in identifying multiple A-type endotracheal cannulas

[0072]

[0073] System 1 is an intelligent detection system for the cleaning status of a tracheostomy cannula medical device provided in this embodiment; System 2 is based on System 1 but does not consider contamination analysis of the outer tube; System 3 is based on System 1 but does not consider contamination analysis of the inner tube.

[0074] As shown in Table 1, the intelligent detection system for the cleaning status of tracheostomy tube medical devices provided in this embodiment has a certain degree of effectiveness.

[0075] Example 2

[0076] The B-type tracheostomy tube is a metal tracheostomy tube. In order to improve the intelligent detection capability of the cleanliness status of the B-type tracheostomy tube and thus ensure medical safety, an intelligent detection system for the cleaning status of tracheostomy tube medical devices was applied.

[0077] Reference Figure 1 A schematic diagram of a smart detection system for the cleaning status of a tracheostomy cannula medical device provided in an embodiment of the present invention includes:

[0078] The image acquisition module acquires images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube after cleaning; thereby obtaining the first set of images of the outer wall and the first set of images of the inner wall of the tracheostomy tube after cleaning; the outer wall images of the outer tube and the outer wall images of the inner tube are acquired by a high-resolution camera; the inner wall images of the outer tube and the inner wall images of the inner tube are acquired by a miniature endoscope.

[0079] Furthermore, the first outer wall image set consists of the outer wall image of the outer tube and the outer wall image of the inner tube; the first inner wall image set consists of the inner wall image of the outer tube and the inner wall image of the inner tube.

[0080] The image recognition module constructs a first image recognition model to recognize a first outer wall image set and a first inner wall image set. The first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer. The preprocessing layer preprocesses the first outer wall image set and the first inner wall image set to obtain a second outer wall image set and a second inner wall image set. The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set to obtain texture features of the outer wall image and the inner wall image. The image feature analysis obtains the first outer wall contamination coefficient and the second inner wall contamination coefficient based on the texture features of the outer wall image and the inner wall image, respectively. Finally, a comprehensive contamination coefficient is obtained based on the first outer wall contamination coefficient and the second inner wall contamination coefficient. Figure 2 This is a schematic diagram of the structure of a first image recognition model provided in an embodiment of the present invention;

[0081] Furthermore, the first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer;

[0082] The preprocessing layer performs preprocessing operations on the first outer wall image set and the first inner wall image set to obtain the second outer wall image set and the second inner wall image set; the preprocessing operations include image denoising, grayscale conversion and geometric correction.

[0083] The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set through a deep convolutional network to obtain the texture features of the outer wall image and the texture features of the inner wall image.

[0084] The image feature analysis is used to obtain the first outer wall contamination coefficient and the second inner wall contamination coefficient based on the texture features of the outer wall image and the texture features of the inner wall image, respectively; and to obtain the comprehensive contamination coefficient based on the first outer wall contamination coefficient and the second inner wall contamination coefficient.

[0085] The output layer is used to output the comprehensive pollution coefficient.

[0086] Furthermore, the outer wall image texture features include outer wall contrast, outer wall brightness, and outer wall uniformity; each outer wall contaminant region is extracted based on the outer wall image texture features, and a first outer wall contamination coefficient is obtained based on the pixel area of ​​each outer wall contaminant region; the inner wall image texture features include inner wall contrast, inner wall brightness, and inner wall uniformity; each inner wall contaminant region is extracted based on the inner wall image texture features, and a second inner wall contamination coefficient is obtained based on the pixel area of ​​each inner wall contaminant region;

[0087] Furthermore, the various contaminants on the outer wall include blood, sputum, drugs, and bacteria; the first outer wall contamination coefficient is obtained by quoting the sum of the pixel areas of each contaminant region on the outer wall with the total pixel area of ​​the outer wall image.

[0088] Furthermore, the first outer wall contamination coefficient is:

[0089] ;

[0090] in, Indicates the first outer wall contamination coefficient; Indicates the type of pollutant; Indicates the first The total pixel area of ​​each pollutant; This represents the total pixel area of ​​the outer wall image;

[0091] The various contaminants on the inner wall include blood, sputum, drugs, and bacteria; the second inner wall contamination coefficient is obtained by quoting the sum of the pixel areas of each contaminant region on the inner wall with the total pixel area of ​​the inner wall image.

[0092] The second inner wall contamination coefficient is:

[0093] ;

[0094] in, This indicates the second inner wall contamination coefficient; Indicates the type of pollutant; Indicates the first The total pixel area of ​​each pollutant; This represents the total pixel area of ​​the inner wall image;

[0095] Furthermore, the comprehensive pollution coefficient is obtained by weighting the first outer wall pollution coefficient and the second inner wall pollution coefficient, specifically as follows:

[0096] ;

[0097] in, Indicates the overall pollution coefficient; Indicates the weight of the first outer wall contamination coefficient; Indicates the first outer wall contamination coefficient; Indicates the weight of the second inner wall contamination coefficient; This represents the second inner wall contamination coefficient.

[0098] This embodiment acquires images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube to form a first outer wall image set and a first inner wall image set. These images are then recognized using a first image recognition model. A deep convolutional network is used to extract texture features from the outer and inner wall images, thereby extracting various contaminant regions and obtaining the first outer wall contamination coefficient and the second inner wall contamination coefficient, ultimately yielding a comprehensive contamination coefficient. The first image recognition model effectively identifies the contamination status of the inner and outer tubes of the tracheostomy tube, significantly improving the intelligent detection capability of the tracheostomy tube's cleaning status and thus effectively ensuring medical safety.

[0099] This embodiment combines comprehensive identification of the inner and outer wall images of the tracheostomy tube with the inner and outer wall images of the tracheostomy tube to obtain the comprehensive contamination coefficient of the tracheostomy tube. This provides model support for comprehensively improving the intelligent detection capability of the tracheostomy tube cleaning status, thereby effectively ensuring medical safety.

[0100] The cleaning status analysis module is used to determine the first cleaning status based on the comprehensive contamination coefficient and a preset threshold.

[0101] Furthermore, the first cleaning state includes a clean state and a contaminated state;

[0102] Furthermore, the first cleaning state includes a clean state and a contaminated state, and the specific judgment process is as follows:

[0103] ;

[0104] in, Indicates the first cleaning state; Indicates a clean state; Indicates a state of pollution; Indicates a preset threshold; Indicates the overall pollution coefficient;

[0105] The cleaning judgment module is used to determine whether cleaning needs to continue based on the first cleaning state. If not, the cleaning ends; if so, the second cleaning time is obtained based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time.

[0106] Furthermore, the initial comprehensive pollution coefficient is obtained by acquiring images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube of the tracheostomy tube that have not been cleaned, thereby obtaining a first set of outer wall images and a first set of inner wall images of the tracheostomy tube that have not been cleaned, and inputting them into a first image recognition model for recognition; the first cleaning time represents the time that the tracheostomy tube has been cleaned.

[0107] Furthermore, the second wash time is:

[0108] ;

[0109] in, Indicates the second wash time; Indicates the overall pollution coefficient; Indicates a preset threshold; This represents the initial comprehensive pollution coefficient; Indicates the first cleaning time; This represents the adjustment factor.

[0110] In summary, such as Figure 3 This is a flowchart illustrating an intelligent detection system for the cleaning status of a tracheostomy cannula medical device, provided in an embodiment of the present invention.

[0111] This embodiment uses the comprehensive contamination coefficient obtained by the first image recognition model, and the second cleaning time is obtained based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient and the first cleaning time. This provides effective data support for further precise cleaning of the tracheal cannula, effectively improves the intelligent detection capability of the cleaning status of the tracheal cannula, and thus effectively ensures medical safety.

[0112] This embodiment acquires a first set of images of the outer wall and a first set of images of the inner wall of the cleaned tracheostomy cannula through an image acquisition module; an image recognition module identifies the first set of images of the outer wall and the first set of images of the inner wall by constructing a first image recognition model, extracts the texture features of the outer wall image and the inner wall image, obtains the first outer wall contamination coefficient and the second inner wall contamination coefficient, and then obtains the comprehensive contamination coefficient; a cleaning status analysis module evaluates the first cleaning status based on the comprehensive contamination coefficient and a preset threshold; a cleaning judgment module determines whether further cleaning is needed. If so, it combines the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time to obtain the second cleaning time; otherwise, the cleaning is terminated. This invention can effectively improve the intelligent detection capability of the cleanliness status of the tracheostomy cannula, thereby ensuring medical safety.

[0113] To verify the effectiveness of the intelligent detection system for the cleaning status of a tracheostomy cannula provided in this embodiment, the cleaning status of multiple B-type tracheostomy cannulas was detected using different systems. The average accuracy of the judgment on whether multiple B-type tracheostomy cannulas need to continue cleaning was calculated, and the comparison results are shown in Table 2. System 1 is the intelligent detection system for the cleaning status of a tracheostomy cannula provided in this embodiment; System 2 is based on System 1 without considering the contamination analysis of the outer tube; and System 3 is based on System 1 without considering the contamination analysis of the inner tube.

[0114] Table 2. Average accuracy of different systems in identifying multiple B-type endotracheal cannulas

[0115]

[0116] As shown in Table 2, the intelligent detection system for the cleaning status of tracheostomy tube medical devices provided in this embodiment has a certain degree of effectiveness.

[0117] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent detection system for the cleaning status of a tracheostomy cannula medical device, characterized in that, include: The image acquisition module acquires images of the outer wall of the outer tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube after cleaning; thereby obtaining the first set of images of the outer wall and the first set of images of the inner wall of the cleaned tracheostomy tube. The image recognition module constructs a first image recognition model to recognize a first outer wall image set and a first inner wall image set. The first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer. The preprocessing layer preprocesses the first outer wall image set and the first inner wall image set to obtain a second outer wall image set and a second inner wall image set. The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set to obtain texture features of the outer wall image and the inner wall image. The image feature analysis obtains the first outer wall pollution coefficient and the second inner wall pollution coefficient based on the texture features of the outer wall image and the inner wall image, respectively. Finally, a comprehensive pollution coefficient is obtained based on the first outer wall pollution coefficient and the second inner wall pollution coefficient. The cleaning status analysis module is used to determine the first cleaning status based on the comprehensive contamination coefficient and a preset threshold. The cleaning judgment module is used to determine whether cleaning needs to continue based on the first cleaning state. If not, the cleaning ends; if so, the second cleaning time is obtained based on the initial comprehensive contamination coefficient, the comprehensive contamination coefficient, the preset threshold, and the first cleaning time.

2. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 1, characterized in that: The first outer wall image set consists of the outer wall image of the outer tube and the outer wall image of the inner tube; the first inner wall image set consists of the inner wall image of the outer tube and the inner wall image of the inner tube.

3. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 2, characterized in that: The first image recognition model includes an input layer, a preprocessing layer, an image feature extraction layer, an image feature analysis layer, and an output layer; The preprocessing layer performs preprocessing operations on the first outer wall image set and the first inner wall image set to obtain the second outer wall image set and the second inner wall image set; the preprocessing operations include image denoising, grayscale conversion and geometric correction. The image feature extraction layer extracts features from the second outer wall image set and the second inner wall image set through a deep convolutional network to obtain the texture features of the outer wall image and the texture features of the inner wall image. The image feature analysis is used to obtain the first outer wall contamination coefficient and the second inner wall contamination coefficient based on the texture features of the outer wall image and the texture features of the inner wall image, respectively; and to obtain the comprehensive contamination coefficient based on the first outer wall contamination coefficient and the second inner wall contamination coefficient. The output layer is used to output the comprehensive pollution coefficient.

4. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 3, characterized in that: The outer wall image texture features include outer wall contrast, outer wall brightness, and outer wall uniformity; each outer wall contaminant region is extracted based on the outer wall image texture features, and a first outer wall contamination coefficient is obtained based on the pixel area of ​​each outer wall contaminant region; the inner wall image texture features include inner wall contrast, inner wall brightness, and inner wall uniformity; each inner wall contaminant region is extracted based on the inner wall image texture features, and a second inner wall contamination coefficient is obtained based on the pixel area of ​​each inner wall contaminant region.

5. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 4, characterized in that: The comprehensive pollution coefficient is obtained by weighting the first outer wall pollution coefficient and the second inner wall pollution coefficient.

6. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 5, characterized in that: The first cleaning state includes a clean state and a contaminated state.

7. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 6, characterized in that: The initial comprehensive pollution coefficient is obtained by acquiring images of the outer wall of the tracheostomy tube, the inner wall of the outer tube, the outer wall of the inner tube, and the inner wall of the inner tube that have not been cleaned, thereby obtaining a first set of outer wall images and a first set of inner wall images of the tracheostomy tube that have not been cleaned, and inputting them into a first image recognition model for recognition; the first cleaning time represents the time that the tracheostomy tube has been cleaned.

8. The intelligent detection system for the cleaning status of a tracheostomy cannula medical device according to claim 7, characterized in that: The second wash cycle is: ； in, Indicates the second wash time; Indicates the overall pollution coefficient; Indicates a preset threshold; This represents the initial comprehensive pollution coefficient; Indicates the first cleaning time; This represents the adjustment factor.

Images