Trachea cannula displacement real-time monitoring method and device

Abstract

The invention discloses a trachea cannula displacement real-time monitoring method and device, and the method comprises the steps: S1, obtaining an internal image of an airway of a patient when a trachea cannula is located at a proper position, and taking the internal image as a standard image; s2, performing image analysis on the standard image to obtain standard state information of the airway tissue of the patient; s3, acquiring a current frame image in the airway of the patient; s4, carrying out displacement state analysis on the current frame image according to the standard state information of the airway tissue of the patient, and obtaining the current state information of the airway tissue of the patient and the similarity percentage between the current state information of the airway tissue of the patient and the standard state information; and S5, setting a displacement threshold value, comparing the similarity percentage with the displacement threshold value, judging that the trachea cannula is displaced when the similarity percentage is lower than the set threshold value, and otherwise, returning to execute S3. According to the invention, the real-time monitoring of the displacement of the trachea cannula is realized, a doctor can be reminded to process when the displacement is monitored, the calculation complexity is reduced, and the real-time monitoring capability is improved.

Description

Technical Field

[0001] The invention belongs to the field of medical devices, and in particular relates to a real-time monitoring method and device for endotracheal tube displacement. Background Art

[0002] The visual double-lumen endotracheal tube is a clinical medical auxiliary device commonly used in thoracic lung surgery. It is mainly used to perform general anesthesia on patients before surgery, establish a respiratory pathway during surgery, provide respiratory support for patients, and relieve airway obstruction after surgery. In actual use, the endotracheal tube should be inserted into the patient's airway at an appropriate depth and position, and its position should be kept relatively fixed throughout the use process to ensure that it plays a good role in airway occlusion and respiratory support.

[0003] However, in actual surgery, doctors often need to change the patient's position to facilitate the operation, and the change in the patient's position can easily cause the endotracheal tube to shift or even fall out of the patient's airway. If the doctor fails to detect the displacement and dislocation of the endotracheal tube in time and restore it to the correct position, it may cause the patient's lung lobe to expand, interfere with the surgical operation, and even cause the patient to suffer from hypoxia, respiratory failure, or cardiopulmonary arrest and other dangerous conditions. Therefore, it is very necessary to monitor the displacement of the endotracheal tube in real time and remind the doctor to deal with it in time when displacement occurs.

[0004] There are several methods for monitoring the displacement of endotracheal tubes:

[0005] (1) A method for monitoring endotracheal tube displacement based on X-ray scanning images and convolutional neural networks. First, an X-ray scanning device is used to scan the patient's chest image. The scanned image is then input into a trained convolutional neural network. Finally, the convolutional neural network identifies the position of the endotracheal tube in the patient's airway in the image. Based on the identification results, corresponding prompts are given to remind the doctor to take action.

[0006] (2) A method for monitoring the displacement of an endotracheal tube based on a CMOS sensor or a Holter sensor and a distance detection algorithm. A Chinese utility model with publication number CN202113456U discloses an endotracheal tube with a real-time displacement monitoring system. The system monitors the position changes of the endotracheal tube in real time by setting a magnetic source at the front end of the endotracheal tube and installing multiple Holter sensors outside the tube.

[0007] (3) Tracheal tube displacement monitoring method based on artificial marking, external camera and convolutional neural network. The Chinese patent with patent number TWI842635B discloses a method for monitoring the displacement of medical tubes through artificial intelligence through labels. First, some specific artificial marks are added to the external tube part of the tracheal tube and the patient's face, and a monitoring camera is installed next to the patient's operating table. Figure 1As shown, the surgical workspace and the above-mentioned markings are photographed in real time, and finally a convolutional neural network is used to analyze the distance and position changes between the various marks in the captured images to monitor the displacement of the endotracheal tube.

[0008] Although the above methods all achieve the displacement monitoring of the endotracheal tube, they require additional X-ray image acquisition devices or external cameras to collect image samples. These additional external image acquisition devices not only increase the cost of the method implementation, but also cause certain interference to the surgical process. In addition, the computational complexity of using deep convolutional neural networks to judge the position of the endotracheal tube is high, the real-time performance is poor, and it requires the support of a hardware platform with high computing power, which further increases the cost and power consumption of the equipment, will increase medical expenses, and increase the economic burden on patients. Adding artificial markings and installing CMOS sensors or Holter sensors on the endotracheal tube requires changing the appearance and structure of the endotracheal tube, increasing the structural complexity of the endotracheal tube, and easily causing safety hazards. Therefore, the displacement monitoring method of the endotracheal tube should be further studied in the direction of low cost and high computational efficiency. Summary of the invention

[0009] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art, reduce the computational complexity, improve the real-time monitoring capability, and provide a method and device for real-time monitoring of endotracheal tube displacement.

[0010] In order to achieve the above object, the present invention adopts the following technical solution:

[0011] A real-time monitoring method for endotracheal tube displacement comprises the following steps:

[0012] S1: Obtain the image of the patient's airway when the endotracheal tube is in the proper position and use it as the standard image;

[0013] S2: performing image analysis on the standard image to obtain standard state information of the patient's airway tissue, wherein the standard state information includes position, shape and area information of the patient's airway tissue region in the standard image;

[0014] S3: Acquire the current frame image inside the patient's airway;

[0015] S4: performing a displacement state analysis on the current frame image according to the standard state information of the patient's airway tissue, obtaining the current state information of the patient's airway tissue, and the similarity percentage between the current state information of the patient's airway tissue and the standard state information, wherein the current state information includes the position, shape and area information of the patient's airway tissue region in the current frame image;

[0016] S5: Set a displacement threshold, compare the similarity percentage with the displacement threshold, and when the similarity percentage is lower than the set threshold, it is determined that the endotracheal tube has shifted, otherwise, return to execute S3.

[0017] The present invention analyzes the standard image and the displacement state of the current frame image, analyzes the area where the patient's airway tissue is located in the standard image, and compares the current frame image with the standard image, and calculates the similarity percentage between the area where the patient's airway tissue is located in the current frame image and the area where the patient's airway tissue is located in the standard image. The difference between the real-time position of the tracheal tube and the standard position is determined according to the similarity percentage. When the difference exceeds a certain threshold, it is determined that the tracheal tube has shifted, and the doctor is reminded to handle it, thereby shortening the image processing time and improving the image processing efficiency and real-time monitoring capability.

[0018] Preferably, the step of performing image analysis on the standard image in S2 includes:

[0019] S2.1: Separate the three-channel RGB color standard image into its red channel image, green channel image and blue channel image;

[0020] S2.2: Set the scale factor of the green channel and the scale factor of the blue channel to obtain the standard grayscale image of the patient's airway tissue area in the standard image. The calculation expression is:

[0021] S_gray=S_r–p×S_g–q×S_b

[0022] Wherein, S_gray is the standard grayscale image, S_r is the red channel image of the standard image, S_g is the green channel image of the standard image, S_b is the blue channel image of the standard image, p is the scale factor of the green channel, and q is the scale factor of the blue channel;

[0023] S2.3: Calculate the mean and variance of non-zero pixels in the standard grayscale image, and calculate the binarization threshold of the standard grayscale image according to the mean and variance of the non-zero pixels, and the expression is:

[0024] thresh=mean+sqrt(var) / 2

[0025] Among them, thresh is the binarization threshold, mean is the mean of non-zero pixels in the standard grayscale image, var is the variance of non-zero pixels in the standard grayscale image, and sqrt() represents the square root operation;

[0026] S2.4: Binarize the standard grayscale image into a standard binary image according to a binarization threshold, and then obtain the position and shape of the patient's airway tissue area in the standard image;

[0027] S2.5: Count the number of non-zero pixels in the standard binary image to obtain the area of ​​the patient's airway tissue region in the standard image.

[0028] Preferably, the step of obtaining the position and shape of the patient's airway tissue area in the standard image in step 2.4 is:

[0029] Pixels in the standard grayscale image whose pixel values ​​are less than the binarization threshold are assigned a value of 0, and pixels whose pixel values ​​are greater than the binarization threshold are assigned a value of 255, to obtain a standard binary image. The non-zero pixel area with a pixel value of 255 in the standard binary image is used to represent the position and shape of the patient's airway tissue area in the standard image.

[0030] Preferably, the step of analyzing the displacement state in S4 comprises:

[0031] S4.1: performing channel separation on the three-channel RGB color current frame image to obtain its red channel image, green channel image and blue channel image;

[0032] S4.2: Obtain the current grayscale image of the patient's airway tissue area in the current frame image, and its calculation expression is:

[0033] F_gray=F_r–p×F_g–q×F_b

[0034] Among them, F_gray is the current grayscale image, F_r is the red channel image of the current frame image, F_g is the green channel image of the current frame image, and F_b is the blue channel image of the current frame image.

[0035] S4.3: Binarize the current grayscale image according to the binarization threshold to obtain a current binary image, and then obtain the position and shape of the patient's airway tissue area in the current frame image;

[0036] S4.4: Count the number of non-zero pixels in the current binary image to obtain the area of ​​the patient's airway tissue region in the current frame image;

[0037] S4.5: Perform element-wise AND operation on the standard binary image and the current binary image to obtain a similar binary image.

[0038] S4.6: Count the number of non-zero pixels in similar binary images;

[0039] S4.7: Calculate the similarity percentage, which is expressed as:

[0040] percent=(nonzero_A / (nonzero_S+nonzero_F-nonzero_A))×100%

[0041] Among them, percent is the similarity percentage, nonzero_A is the number of nonzero pixels in the similar binary image, nonzero_S is the number of nonzero pixels in the standard binary image, and nonzero_F is the number of nonzero pixels in the current binary image.

[0042] Preferably, the step of obtaining the position and shape of the patient's airway tissue area in the current frame image in step S4.3 is:

[0043] Pixels in the current grayscale image whose pixel values ​​are less than the binarization threshold are assigned a value of 0, and pixels whose pixel values ​​are greater than the binarization threshold are assigned a value of 255 to obtain a current binary image. The non-zero pixel area with a pixel value of 255 in the current binary image is used to represent the position and shape of the patient's airway tissue area in the current frame image.

[0044] Preferably, the steps of element AND operation in step S4.5 are:

[0045] When the pixel values ​​corresponding to the same position of the standard binary image and the current binary image are both equal to 255, the pixel at the corresponding position in the similar binary image is assigned a value of 255; otherwise, the pixel at the corresponding position in the similar binary image is assigned a value of 0.

[0046] Preferably, the method further comprises step S6: when it is determined that the endotracheal tube is displaced, providing a displacement prompt.

[0047] The real-time monitoring method for endotracheal tube displacement of the present invention realizes the monitoring function of endotracheal tube displacement, and can remind doctors to handle it when endotracheal tube displacement is detected, thereby reducing the risk of patients' lung lobe expansion, hypoxia, respiratory failure or cardiopulmonary arrest caused by endotracheal tube displacement. The traditional image processing algorithm is used to replace the convolutional neural network, with low hardware resource occupancy and good real-time performance, which solves the problems of high computational complexity, poor real-time performance, dependence on high-performance hardware platform and high power consumption of the existing technical solutions.

[0048] A real-time monitoring device for endotracheal tube displacement, used to implement the real-time monitoring method for endotracheal tube displacement, comprising an endotracheal tube, an image acquisition unit and an image processing host, wherein the image acquisition unit is connected to the image processing host;

[0049] The image acquisition unit is used to acquire the internal image of the patient's airway in real time;

[0050] The image processing host includes a memory, a processor and a human-computer interaction interface. The memory is used to store images acquired by the image acquisition unit and program instructions of the real-time monitoring method of endotracheal tube displacement; the processor is used to perform image analysis on standard images and perform displacement state analysis on the current frame image based on standard state information of the patient's airway tissue; the human-computer interaction interface is used to set parameters, display images of the inside of the patient's airway in real time, and set standard images.

[0051] Preferably, the image acquisition unit adopts a camera integrated inside the endotracheal tube.

[0052] Preferably, the parameters set through the human-computer interaction interface include one or more of a shift threshold, a scaling factor of a green channel, and a scaling factor of a blue channel.

[0053] The present invention uses a low-power micro camera to replace a complex and expensive in vitro image acquisition device, thereby solving the problem of high cost in the prior art solutions. It does not require the use of additional sensors and pipeline markers, thus avoiding changes in the existing appearance and structure of the endotracheal tube, and solving the problem in the prior art solutions that changes in the appearance and structure of the endotracheal tube are prone to potential safety hazards, which is beneficial to ensuring product safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0055] Figure 1 is a schematic diagram of a method for monitoring the displacement of medical tubes by artificial intelligence through tags;

[0056] Figure 2 is a method block diagram of the method of the present invention;

[0057] Figure 3 is a flow chart of the method of the present invention;

[0058] Figure 4 is a flow chart of standard image analysis in the method of the present invention;

[0059] Figure 5 is a flow chart of the shift state analysis in the method of the present invention;

[0060] Figure 6 is a schematic diagram of the method of the present invention;

[0061] Figure 7 It is a schematic diagram of the device of the present invention. DETAILED DESCRIPTION

[0062] like Figure 2-6 As shown, the present invention provides a real-time monitoring method for endotracheal tube displacement, comprising the following steps:

[0063] S1: Acquire an image of the patient's airway interior when the endotracheal tube is in a suitable position, and use it as a standard image S. In this embodiment, the standard image S is a three-channel RGB color image with a resolution of 1280×720; the total number of pixels of the image is 921600;

[0064] S2: Perform image analysis on the standard image S to obtain standard state information of the patient's airway tissue, where the standard state information includes position, shape and area information of the patient's airway tissue region in the standard image S.

[0065] like Figure 4 As shown, the specific steps include:

[0066] S2.1: Separate the three-channel RGB color standard image S to obtain its red channel image S_r, green channel image S_g and blue channel image S_b;

[0067] S2.2: Set the scale factor of the green channel and the scale factor of the blue channel to obtain the standard grayscale image S_gray of the patient's airway tissue area in the standard image S. The calculation expression is:

[0068] S_gray=S_r–p×S_g–q×S_b

[0069] Wherein, p is the scaling factor of the green channel, and q is the scaling factor of the blue channel; in this embodiment, both p and q are 0.5. Figure 6 As shown, the gray-white area in the standard grayscale image S_gray is a non-zero pixel area, which represents the approximate area where the patient's airway tissue in the standard image S is located.

[0070] S2.3: Calculate the mean and variance of non-zero pixels in the standard grayscale image S_gray, and calculate the binarization threshold of the standard grayscale image S_gray based on the mean and variance of non-zero pixels. The expression is:

[0071] thresh=mean+sqrt(var) / 2

[0072] Wherein, thresh is the binarization threshold, mean is the mean of non-zero pixels in the standard grayscale image S_gray, var is the variance of non-zero pixels in the standard grayscale image S_gray, and sqrt() represents a square root operation. In this embodiment, mean=71.75, var=1969.5844, and thresh=93.94.

[0073] S2.4: According to the binarization threshold, the standard grayscale image S_gray is binarized into a standard binary image S_bin, and then the position and shape of the patient's airway tissue area in the standard image S are obtained. The specific steps are:

[0074] The pixels whose pixel values ​​in the standard grayscale image S_gray are less than the binarization threshold are assigned a value of 0, and the pixels whose pixel values ​​are greater than the binarization threshold are assigned a value of 255, thereby obtaining the standard binary image S_bin. The non-zero pixel area with a pixel value of 255 in the standard binary image S_bin is used to represent the position and shape of the patient's airway tissue area in the standard image S.

[0075] S2.5: Count the number of non-zero pixels in the standard binary image S_bin, that is, the number of pixels with a pixel value equal to 255 nonzero_S=240253, to obtain the area of ​​the patient's airway tissue region in the standard image S.

[0076] S3: Acquire a current frame image F of the patient's airway; the acquired current frame image is a three-channel RGB color image with a resolution of 1280×720; the total number of pixels of the image is 921600.

[0077] S4: Performing a displacement state analysis on the current frame image F according to the standard state information of the patient's airway tissue, obtaining the current state information of the patient's airway tissue, and the similarity percentage between the current state information of the patient's airway tissue and the standard state information, wherein the current state information includes the position, shape and area information of the patient's airway tissue region in the current frame image F. Figure 5 As shown, the specific steps are:

[0078] S4.1: performing channel separation on the three-channel RGB color current frame image F to obtain its red channel image, green channel image and blue channel image;

[0079] S4.2: Obtain the current grayscale image F_gray of the patient's airway tissue area in the current frame image F, and its calculation expression is:

[0080] F_gray=F_r–p×F_g–q×F_b

[0081] Among them, F_gray is the current grayscale image F_gray, F_r is the red channel image of the current frame image F, F_g is the green channel image of the current frame image F, and F_b is the blue channel image of the current frame image F.

[0082] like Figure 6 As shown, the gray-white area in the current grayscale image F_gray is a non-zero pixel area, which represents the approximate area where the patient's airway tissue in the current frame image F is located.

[0083] S4.3: Binarize the current grayscale image F_gray according to the binarization threshold to obtain the current binary image F_bin, and then obtain the position and shape of the patient's airway tissue area in the current frame image F. The specific steps are:

[0084] The pixels whose pixel values ​​in the current grayscale image F_gray are less than the binarization threshold are assigned a value of 0, and the pixels whose pixel values ​​are greater than the binarization threshold are assigned a value of 255, thereby obtaining the current binary image F_bin. The non-zero pixel region with a pixel value of 255 in the current binary image F_bin is used to represent the position and shape of the patient's airway tissue region in the current frame image F.

[0085] S4.4: Count the number of non-zero pixels in the current binary image F_bin, that is, the number of pixels with a pixel value equal to 255, nonzero_F=127004, to obtain the area of ​​the patient's airway tissue region in the current frame image F;

[0086] S4.5: Perform an element-wise AND operation on the standard binary image S_bin and the current binary image F_bin to obtain a similar binary image A_bin. In this embodiment, the steps of the element-wise AND operation in step S4.5 are:

[0087] When the pixel values ​​corresponding to the same position of the standard binary image S_bin and the current binary image F_bin are both equal to 255, the pixel at the corresponding position in the similar binary image A_bin is assigned 255; otherwise, the pixel at the corresponding position in the similar binary image A_bin is assigned 0.

[0088] S4.6: Count the number of non-zero pixels in the similar binary image A_bin, that is, count the number of pixels with a pixel value equal to 255, nonzero_A=99863;

[0089] S4.7: Calculate the similarity percentage, which is expressed as:

[0090] percent=(nonzero_A / (nonzero_S+nonzero_F-nonzero_A))×100%

[0091] Wherein, percent is the similarity percentage, nonzero_A is the number of nonzero pixels in the similar binary image A_bin, nonzero_S is the number of nonzero pixels in the standard binary image S_bin, and nonzero_F is the number of nonzero pixels in the current binary image F_bin. In this embodiment, the similarity percentage percent=(99863 / (240253+127004-99863))×100%=37.35% can be calculated.

[0092] In this embodiment, when a large displacement occurs in the endotracheal tube, the relative position between the endotracheal tube and the patient's airway tissue changes, and the endotracheal tube will cause a certain blockage to the airway tissue area of the patient, resulting in a large difference between the area where the patient's airway tissue is located in the current frame image and the area where the patient's airway tissue is located in the standard image. At this time, the calculated similarity percentage is relatively low.

[0093] S5: Set a displacement threshold, compare the similarity percentage with the displacement threshold. When the similarity percentage is lower than the set threshold, it is determined that the endotracheal tube has shifted. Otherwise, return to execute S3. In this embodiment, the set displacement threshold T is 68%. According to the similarity percentage percent = 37.35%, it can be determined that percent < T, that is, the endotracheal tube has shifted. In this embodiment, the endotracheal tube displacement threshold T is used to adjust the sensitivity of the displacement state analysis and is an adjustable parameter set manually. Its adjustment range is 30% - 90%. In this embodiment, the default value of T is 68%, and the doctor can fine-tune the T value according to needs. The larger the value of T, the higher the sensitivity of the displacement state analysis, and the more subtle displacements can be monitored.

[0094] Step S6: When it is determined that the endotracheal tube has shifted, give a displacement prompt. According to the judgment result of S5, give a displacement prompt to remind the doctor to handle it.

[0095] The real-time monitoring method for endotracheal tube displacement of the present invention realizes the function of monitoring endotracheal tube displacement, can remind the doctor to handle it when the endotracheal tube displacement is detected, and reduces the risks such as patient lung lobe inflation, hypoxia, respiratory failure or cardiopulmonary arrest caused by endotracheal tube displacement. Using traditional image processing algorithms instead of convolutional neural networks, it has a low hardware resource occupancy rate and good real-time performance, and solves the problems of high computational complexity, poor real-time performance, dependence on high-performance hardware platforms, and high power consumption in the existing technical solutions; by analyzing the standard image and analyzing the displacement state of the current frame image, analyzing the area where the patient's airway tissue is located in the standard image and comparing the current frame image with the standard image, calculating the similarity percentage between the area where the patient's airway tissue is located in the current frame image and the area where the patient's airway tissue is located in the standard image. Determine the difference between the real-time position and the standard position of the endotracheal tube according to the similarity percentage. When the difference exceeds a certain threshold, it is determined that the endotracheal tube has shifted, and the doctor is reminded to handle it, shortening the image processing time and improving the image processing efficiency and real-time monitoring ability.

[0096] Such as Figure 7As shown, the present invention also provides a real-time monitoring device for endotracheal tube displacement, which is used to implement a real-time monitoring method for endotracheal tube displacement, including a endotracheal tube, an image acquisition unit and an image processing host, wherein the image acquisition unit and the image processing host are connected via a USB to Type-C connection cable; the image acquisition unit is used to acquire real-time images of the patient's airway.

[0097] The image processing host includes a memory, a processor and a human-computer interaction interface. The memory is used to store images acquired by the image acquisition unit and program instructions of the real-time monitoring method for endotracheal tube displacement; the processor is used to perform image analysis on the standard image S and perform displacement state analysis on the current frame image F according to the standard state information of the patient's airway tissue; the human-computer interaction interface is used to set parameters, display images of the inside of the patient's airway in real time, and set the standard image S.

[0098] In this embodiment, the image acquisition unit uses a camera integrated inside the endotracheal tube.

[0099] In this embodiment, the parameters set through the human-computer interaction interface include one or more of a shift threshold, a scale factor of a green channel, and a scale factor of a blue channel.

[0100] In this embodiment, when used, the doctor first inserts the endotracheal tube into the appropriate position of the patient's airway, and the image acquisition unit acquires the image of the inside of the patient's airway at that position. After the doctor presses the standard image setting button on the human-computer interaction interface of the image processing host, the image of the inside of the patient's airway at that position is used as the standard image S, and the image processing host performs image analysis on the standard image S. After that, the image acquisition unit acquires the image of the inside of the patient's airway in real time, and after setting the parameters through the human-computer interaction interface, the image processing host performs displacement state analysis, and displays the image of the inside of the patient's airway in real time through the human-computer interaction interface. When the displacement state analysis is completed, the displacement state analysis result is output. When the image processing host determines that displacement has occurred, it outputs displacement prompt information to remind the doctor to handle it.

[0101] The present invention uses a low-power micro camera to replace a complex and expensive in vitro image acquisition device, thereby solving the problem of high cost in the prior art solutions. It does not require the use of additional sensors and pipeline markers, thus avoiding changes in the existing appearance and structure of the endotracheal tube, and solving the problem in the prior art solutions that changes in the appearance and structure of the endotracheal tube are prone to potential safety hazards, which is beneficial to ensuring product safety.

[0102] The above are only preferred embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope defined by the claims.

Claims

1. A real-time monitoring method for endotracheal tube displacement, characterized in that: The following steps are involved: S1: Obtain the image of the patient's airway when the endotracheal tube is in the proper position and use it as the standard image; S2: performing image analysis on the standard image to obtain standard state information of the patient's airway tissue, wherein the standard state information includes position, shape and area information of the patient's airway tissue region in the standard image; S3: Acquire the current frame image inside the patient's airway; S4: performing a displacement state analysis on the current frame image according to the standard state information of the patient's airway tissue, obtaining the current state information of the patient's airway tissue, and the similarity percentage between the current state information of the patient's airway tissue and the standard state information, wherein the current state information includes the position, shape and area information of the patient's airway tissue region in the current frame image; S5: Set a displacement threshold, compare the similarity percentage with the displacement threshold, and when the similarity percentage is lower than the set threshold, it is determined that the endotracheal tube is displaced, otherwise, return to execute S3.

2. The real-time monitoring method for endotracheal tube displacement according to claim 1, characterized in that: The step of performing image analysis on the standard image in S2 comprises: S2.1: performing channel separation on the three-channel RGB color standard image to obtain its red channel image, green channel image and blue channel image; S2.2: Set the scale factor of the green channel and the scale factor of the blue channel to obtain the standard grayscale image of the patient's airway tissue area in the standard image. The calculation expression is: S_gray=S_r–p×S_g–q×S_b Wherein, S_gray is the standard grayscale image, S_r is the red channel image of the standard image, S_g is the green channel image of the standard image, S_b is the blue channel image of the standard image, p is the scale factor of the green channel, and q is the scale factor of the blue channel; S2.3: Calculate the mean and variance of non-zero pixels in the standard grayscale image, and calculate the binarization threshold of the standard grayscale image according to the mean and variance of the non-zero pixels, and the expression is: thresh=mean+sqrt(var) / 2 Among them, thresh is the binarization threshold, mean is the mean of non-zero pixels in the standard grayscale image, var is the variance of non-zero pixels in the standard grayscale image, and sqrt() represents the square root operation; S2.4: Binarize the standard grayscale image into a standard binary image according to a binarization threshold, and then obtain the position and shape of the patient's airway tissue area in the standard image; S2.5: Count the number of non-zero pixels in the standard binary image to obtain the area of ​​the patient's airway tissue region in the standard image.

3. The real-time monitoring method for endotracheal tube displacement according to claim 2, characterized in that: The steps for obtaining the position and shape of the patient's airway tissue area in the standard image described in step 2.4 are: Pixels in the standard grayscale image whose pixel values ​​are less than the binarization threshold are assigned a value of 0, and pixels whose pixel values ​​are greater than the binarization threshold are assigned a value of 255, to obtain a standard binary image. The non-zero pixel area with a pixel value of 255 in the standard binary image is used to represent the position and shape of the patient's airway tissue area in the standard image.

4. The real-time monitoring method for endotracheal tube displacement according to claim 1, characterized in that: The step of shift state analysis in S4 comprises: S4.1: performing channel separation on the three-channel RGB color current frame image to obtain its red channel image, green channel image and blue channel image; S4.2: Obtain the current grayscale image of the patient's airway tissue area in the current frame image, and its calculation expression is: F_gray=F_r–p×F_g–q×F_b Among them, F_gray is the current grayscale image, F_r is the red channel image of the current frame image, F_g is the green channel image of the current frame image, and F_b is the blue channel image of the current frame image. S4.3: Binarize the current grayscale image according to the binarization threshold to obtain a current binary image, and then obtain the position and shape of the patient's airway tissue area in the current frame image; S4.4: Count the number of non-zero pixels in the current binary image to obtain the area of ​​the patient's airway tissue region in the current frame image; S4.5: Perform element-wise AND operation on the standard binary image and the current binary image to obtain a similar binary image. S4.6: Count the number of non-zero pixels in similar binary images; S4.7: Calculate the similarity percentage, which is expressed as: percent=(nonzero_A / (nonzero_S+nonzero_F-nonzero_A))×100% Among them, percent is the similarity percentage, nonzero_A is the number of nonzero pixels in the similar binary image, nonzero_S is the number of nonzero pixels in the standard binary image, and nonzero_F is the number of nonzero pixels in the current binary image.

5. The real-time monitoring method for endotracheal tube displacement according to claim 4, characterized in that: The steps of obtaining the position and shape of the patient's airway tissue area in the current frame image in step S4.3 are: Pixels in the current grayscale image whose pixel values ​​are less than the binarization threshold are assigned a value of 0, and pixels whose pixel values ​​are greater than the binarization threshold are assigned a value of 255 to obtain a current binary image. The non-zero pixel area with a pixel value of 255 in the current binary image is used to represent the position and shape of the patient's airway tissue area in the current frame image.

6. The real-time monitoring method for endotracheal tube displacement according to claim 4, characterized in that: The steps of the element AND operation in step S4.5 are: When the pixel values ​​corresponding to the same position of the standard binary image and the current binary image are both equal to 255, the pixel at the corresponding position in the similar binary image is assigned a value of 255; otherwise, the pixel at the corresponding position in the similar binary image is assigned a value of 0.

7. The real-time monitoring method for endotracheal tube displacement according to claim 1, characterized in that: The method further comprises step S6: when it is determined that the endotracheal tube is displaced, providing a displacement prompt.

8. A real-time monitoring device for endotracheal tube displacement, used to implement the real-time monitoring method for endotracheal tube displacement according to any one of claims 1 to 7, characterized in that: It includes an endotracheal cannula, an image acquisition unit and an image processing host, wherein the image acquisition unit is connected to the image processing host; The image acquisition unit is used to acquire the internal image of the patient's airway in real time; The image processing host includes a memory, a processor and a human-computer interaction interface. The memory is used to store images acquired by the image acquisition unit and program instructions of the real-time monitoring method of endotracheal tube displacement; the processor is used to perform image analysis on standard images and perform displacement state analysis on the current frame image based on standard state information of the patient's airway tissue; the human-computer interaction interface is used to set parameters, display images of the inside of the patient's airway in real time, and set standard images.

9. The real-time monitoring device for endotracheal tube displacement according to claim 8, characterized in that: The image acquisition unit adopts a camera integrated inside the endotracheal tube.

10. The real-time monitoring device for endotracheal tube displacement according to claim 8, characterized in that: The parameters set through the human-computer interaction interface include one or more of a shift threshold, a scaling factor of a green channel, and a scaling factor of a blue channel.

Images