PICC (Peripherally Inserted Central Catheter) ultrasonic puncture positioning system

Abstract

The invention discloses a PICC (Peripherally Inserted Central Catheter) ultrasonic puncture positioning system, and belongs to the technical field of medical instruments. The invention discloses a PICC (Peripherally Inserted Central Catheter) ultrasonic puncture positioning system which comprises a data collection module, an identification model and a positioning module, the PICC positioning method solves the problem that the accuracy and practicability of the prior art are affected by factors such as ambient light and spatial layout, can extract more accurate sound wave features by using the Mel-frequency cepstrum coefficient to perform PICC positioning, effectively identifies and positions the sound wave signals of the PICC, improves the positioning precision, and improves the positioning accuracy. According to the PICC positioning device, interference of background noise on PICC positioning can be effectively reduced, positioning accuracy is improved, it can be ensured that medicine is smoothly conveyed, the treatment effect of the medicine is improved, damage to surrounding tissue is avoided, complications such as phlebitis and infection are reduced, and peripheral veins of a patient are protected against corrosion of chemotherapy medicine.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to a PICC catheterization ultrasonic puncture positioning system. Background Art

[0002] PICC (Peripherally Inserted Central Venous Catheters) uses a catheter to puncture the vein in the arm. The catheter goes directly to the large vein close to the heart, avoiding direct contact between chemotherapy drugs and the arm veins. In addition, the blood flow in the large veins is very fast, which can quickly dilute the chemotherapy drugs and prevent the drugs from irritating the blood vessels. Therefore, it can effectively protect the upper limb veins, reduce the occurrence of phlebitis, relieve the patient's pain, and improve the patient's quality of life.

[0003] A Chinese patent with publication number CN118845214A discloses a method for measuring the length and positioning the tip of a PICC catheter based on mixed reality technology. By combining the patient's three-dimensional vascular model with multiple control models in a database, the length of the PICC catheter can be accurately measured and corrected, thereby improving the accuracy and safety of the catheter. Taking into account the patient's individualized data such as arm circumference, body fat percentage, and vascular elasticity, the length of the PICC catheter is more in line with individual physiological characteristics, thereby improving the treatment effect. Using mixed reality technology, the catheter position can be acquired and adjusted in real time to avoid errors that may occur in traditional methods. Accurate catheter length and tip positioning reduce possible complications during PICC catheterization. Personalized and precise catheterization methods reduce intrusion on patients and improve patients' comfort and satisfaction. Through automated data processing and model matching, the operating procedures of medical staff are simplified and work efficiency is improved.

[0004] In actual use, the above patent uses mixed reality technology to determine the position of the PICC catheter. However, the use of mixed reality technology may be affected by factors such as ambient light and spatial layout, which may affect its accuracy and practicality. Therefore, it does not meet existing needs. We propose a PICC catheter ultrasound puncture positioning system. Summary of the invention

[0005] The purpose of the present invention is to provide a PICC catheter ultrasonic puncture positioning system. The use of Mel-frequency cepstrum coefficients for PICC catheter positioning can extract more accurate sound wave features, effectively identify and locate the sound wave signals of the PICC catheter, thereby improving the positioning accuracy. When processing audio signals, the interference of background noise on PICC catheter positioning can be effectively reduced, and the positioning accuracy can be improved. A high positioning accuracy can be maintained under different environments and conditions. Whether in a noisy environment or in a quiet environment, the Mel-frequency cepstrum coefficients can effectively extract the sound wave features of the PICC catheter. By positioning the PICC catheter, it can ensure the smooth delivery of drugs and improve the therapeutic effect of drugs. In particular, for patients who need long-term infusion or chemotherapy, it can avoid damage to surrounding tissues, reduce the occurrence of complications such as phlebitis and infection, and protect the patient's peripheral veins from corrosion by chemotherapy drugs, thereby solving the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a PICC catheterization ultrasonic puncture positioning system, comprising: A data collection module is used to collect the position of the PICC catheter and the acoustic wave signal generated by the PICC catheter during the insertion process using an implantable sensor, and to process the acoustic wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients, and save the two-dimensional matrix in an image format; A recognition model is used to establish a neural network model for acoustic signal recognition, and to train and optimize the neural network model for acoustic signal recognition; The positioning module is used to obtain the fusion features of different scales and different time points in the Mel frequency cepstral coefficient map, and use the acoustic signal recognition neural network model to analyze the fusion features to obtain the location information of the PICC catheter.

[0007] Preferably, the data collection module comprises: An implantable sensor is placed inside the PICC catheter to obtain the position where the PICC catheter enters and the acoustic wave signal generated during the entry of the PICC catheter; The processing module is used to process the acquired sound wave signal and extract the characteristics of the sound wave signal to obtain a two-dimensional matrix of Mel-frequency cepstral coefficients, and save the two-dimensional matrix in an image format.

[0008] Preferably, the processing module comprises: The feature extraction module is used to filter, emphasize, frame and process the acquired sound wave signal with a window function, and extract the sound wave signal features from the converted sound wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients; The image conversion module is used to save the two-dimensional matrix of Mel-frequency cepstral coefficients into image format, where the rows represent the number of Mel filters and the columns represent the time frames.

[0009] Preferably, the processing module specifically includes: A first-order high-pass filter is used for pre-emphasis processing, the leakage acoustic wave signal after pre-emphasis processing is divided into frames, and each frame is multiplied by a window function for windowing; Perform fast Fourier transform on each frame signal after frame division and windowing to obtain the spectrum of each frame, convert the actual frequency length into Mel frequency length, configure the triangular filter bank and calculate the output of each triangular filter after filtering the signal amplitude spectrum; Perform logarithmic operation on all filter outputs and further perform discrete cosine transform to obtain a two-dimensional matrix of Mel-frequency cepstral coefficients, and save the two-dimensional matrix of Mel-frequency cepstral coefficients in image format.

[0010] Preferably, the identification module includes: A model building module is used to build a neural network model for acoustic signal recognition using a deep learning algorithm of a neural network, and divide the acoustic signal features into a training set and a validation set; A model training module is used to train the acoustic signal recognition neural network model using a training set, and adjust the hyperparameters of the acoustic signal recognition neural network model through a back propagation algorithm; The model evaluation module is used to evaluate the performance of the acoustic signal recognition neural network model using the validation set, adjust the parameters of the acoustic signal recognition neural network model according to the evaluation results, and optimize the performance of the acoustic signal recognition neural network model.

[0011] Preferably, the model building module specifically includes: Use the deep learning algorithm of neural network to establish a neural network model for acoustic signal recognition. The neural network model consists of an input layer, a hidden layer, and an output layer. The data set required by the input layer is derived from the time domain signals and time-frequency spectrum signals of the measured sound waves under different working conditions, and is divided into training sets and validation sets in appropriate proportions; The hidden layer is responsible for extracting and learning the sound wave features, including the convolution layer, feature extraction sampling layer, pooling layer and fully connected layer; The output layer is responsible for exporting the acceleration curve identified as the train sound wave as the recognition result; Preferably, the positioning module includes: Feature fusion module, used to extract fusion features of different scales and different time points in the Mel frequency cepstral coefficient graph; A location determination module is used to input the extracted fusion features into the acoustic wave signal recognition neural network model to analyze the fusion features, and output the location information of the PICC catheter according to the analysis results; The real-time monitoring module is used to monitor the patient's blood flow dynamics parameters in real time and issue corresponding warnings when the patient's blood flow dynamics parameters are abnormal.

[0012] Preferably, the fusion features of the extracted Mel frequency cepstral coefficient map specifically include: Several Mel-frequency cepstral coefficient feature vectors of different scales are concatenated into a high-dimensional vector and input into the compressed time-delay network; The output of the compressed time-delay network is used as the sound wave feature and concatenated with several Mel-frequency cepstral coefficient feature vectors at different time points to obtain a mixed feature. Convolutional neural network is used to process mixed features, and mixed features are learned through convolutional layers. Several Mel-frequency cepstral coefficient features of different scales and time points are fused to obtain the fused features of Mel-frequency cepstral coefficient map.

[0013] Preferably, the PICC catheterization ultrasound puncture positioning system further includes: A model building module, used for building a position marking model according to the position information of the PICC catheter and a pre-configured human body model of the patient; The data collection module also includes: A signal receiving module, used for receiving the acoustic wave signal of the implantable sensor; The signal receiving module is configured as at least two ultrasonic receivers arranged at different positions; a positioning module is configured next to the ultrasonic receiver, and the ultrasonic receivers are respectively configured at the end of a manipulator; The model building module identifies the ultrasonic receiver into the position marking model according to the positioning module when building the position marking model; The robot's controller performs the following operations: Track the position of the end of the PICC catheter on the position marking model and generate a position sequence of the end of the PICC catheter in the human body; Determine a prediction sequence based on the position sequence and a pre-configured position prediction library; Determine the layout area of ​​the receiver corresponding to each prediction position in the prediction sequence according to the preset layout area library; The center point of the intersection area of ​​the layout areas corresponding to each predicted position is used as the target position of the end of the manipulator to control the movement of the manipulator.

[0014] Preferably, the PICC catheterization ultrasonic puncture positioning system further includes: a risk assessment module for real-time assessment and output of the operation risk of the PICC catheter; Among them, the risk assessment module performs the following operations: Obtain the position relationship of the implementer relative to the patient and retrieve the corresponding personnel model according to the implementer's current action; Construct simulation scenarios based on location marking models, location relationships and personnel models; Track the actions of the implementers, identify the risk areas and mark them on the personnel models in the simulation scenario; In the simulation scenario, the moving trajectory of the manipulator is marked and the first risk value is calculated based on the positional relationship between each trajectory point and the risk part. The calculation formula of the first risk value is as follows: In the formula, represents the first risk value; Indicates The trajectory points and risk positions The distance between the sampling points; Shidi The trajectory points and risk positions The preset risk factor corresponding to each sampling point; is the total number of trajectory points; is the total number of sampling points; The moving speed of the PICC catheter end and the position of the predicted sequence are evaluated to obtain the second risk value; the calculation formula of the second risk value is as follows: In the formula, represents the second risk value; Indicates The hazard value of each predicted location; Indicates The preset weight coefficients of the predicted positions; Indicates the movement speed of the PICC catheter tip; is the risk conversion factor of the preset corresponding speed; is the standard speed configured; The first risk value and the second risk value are output.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention utilizes the Mel-frequency cepstrum coefficients to locate the PICC catheter, which can extract more accurate sound wave features, effectively identify and locate the sound wave signals of the PICC catheter, thereby improving the positioning accuracy. When processing the audio signal, it can effectively reduce the interference of background noise on the positioning of the PICC catheter, improve the positioning accuracy, and maintain a high positioning accuracy in different environments and conditions. Whether in a noisy environment or in a quiet environment, the Mel-frequency cepstrum coefficients can effectively extract the sound wave features of the PICC catheter. By locating the PICC catheter, it can ensure the smooth delivery of drugs and improve the therapeutic effect of drugs. In particular, for patients who need long-term infusion or chemotherapy, it can avoid damage to surrounding tissues, reduce the occurrence of complications such as phlebitis and infection, and protect the patient's peripheral veins from corrosion by chemotherapy drugs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a PICC catheterization ultrasonic puncture positioning system of the present invention; Figure 2 The present invention is a schematic diagram of a method of a PICC catheterization ultrasonic puncture positioning system. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In order to solve the problem that the existing patent uses mixed reality technology to determine the location of PICC catheter in actual use, but the use of mixed reality technology may be affected by factors such as ambient light and spatial layout, which may affect its accuracy and practicality, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions A PICC catheterization ultrasonic puncture positioning system, comprising: A data collection module is used to collect the position of the PICC catheter and the acoustic wave signal generated by the PICC catheter during the insertion process using an implantable sensor, and to process the acoustic wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients, and save the two-dimensional matrix in an image format; A recognition model is used to establish a neural network model for acoustic signal recognition, and to train and optimize the neural network model for acoustic signal recognition; The positioning module is used to obtain the fusion features of different scales and different time points in the Mel frequency cepstral coefficient map, and use the acoustic signal recognition neural network model to analyze the fusion features to obtain the location information of the PICC catheter.

[0019] Data collection module, including: An implantable sensor is placed inside the PICC catheter to obtain the position where the PICC catheter enters and the acoustic wave signal generated during the entry of the PICC catheter; The processing module is used to process the acquired sound wave signal and extract the characteristics of the sound wave signal to obtain a two-dimensional matrix of Mel-frequency cepstral coefficients, save the two-dimensional matrix in image format, capture key features and improve analysis accuracy. When processing the audio signal, the signal will be pre-processed, including windowing, denoising, smoothing and other operations to remove interference factors such as noise and burst sounds, which can effectively reduce the interference of background noise on PICC catheter positioning and improve positioning accuracy. It can maintain a high positioning accuracy in different environments and conditions. Whether in a noisy environment or in a quiet environment, the Mel-frequency cepstral coefficients can effectively extract the sound wave characteristics of the PICC catheter, thereby achieving accurate positioning.

[0020] Processing module, including: The feature extraction module is used to filter, emphasize, frame and process the acquired sound wave signal with a window function, and extract the sound wave signal features from the converted sound wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients; The image conversion module is used to save the two-dimensional matrix of Mel-frequency cepstral coefficients into image format, where the rows represent the number of Mel filters and the columns represent the time frames. This matrix can be directly regarded as a grayscale image, where the value of each pixel corresponds to an element value in the Mel-frequency spectrum. Therefore, you only need to save the matrix data of the Mel-frequency spectrum into image format to get a picture representing the Mel-frequency cepstral coefficients. Converting the Mel-frequency cepstral coefficients into images makes the features more intuitive and easier for humans to understand and analyze. Features in image form can be more easily visualized, thereby helping researchers better understand the characteristics of sound wave signals. After converting the Mel-frequency cepstral coefficients into images, the computational efficiency is higher, which can reduce the computational cost, speed up the training and reasoning of the model, and improve the processing efficiency.

[0021] Processing modules, including: A first-order high-pass filter is used for pre-emphasis processing, the leakage acoustic wave signal after pre-emphasis processing is divided into frames, and each frame is multiplied by a window function for windowing; Perform fast Fourier transform on each frame signal after frame division and windowing to obtain the spectrum of each frame, convert the actual frequency length into Mel frequency length, configure the triangular filter bank and calculate the output of each triangular filter after filtering the signal amplitude spectrum; Perform logarithmic operation on all filter outputs and further perform discrete cosine transform to obtain a two-dimensional matrix of Mel-frequency cepstral coefficients, and save the two-dimensional matrix of Mel-frequency cepstral coefficients in image format.

[0022] The acquired sound wave signal is filtered, emphasized, framed and processed by window function to improve the clarity and quality of the signal. The filter can effectively remove high-frequency or low-frequency noise in the signal, thereby improving the clarity and quality of the signal and helping to extract and analyze the characteristics of the sound wave signal.

[0023] Identification module, including: A model building module is used to build a neural network model for acoustic signal recognition using a deep learning algorithm of a neural network, and divide the acoustic signal features into a training set and a validation set; A model training module is used to train the acoustic signal recognition neural network model using a training set, and adjust the hyperparameters of the acoustic signal recognition neural network model through a back propagation algorithm; The model evaluation module is used to evaluate the performance of the acoustic signal recognition neural network model using the validation set, adjust the parameters of the acoustic signal recognition neural network model according to the evaluation results, and optimize the performance of the acoustic signal recognition neural network model.

[0024] Model building modules, including: Use the deep learning algorithm of neural network to establish a neural network model for acoustic signal recognition. The neural network model consists of an input layer, a hidden layer, and an output layer. The data set required by the input layer is derived from the time domain signals and time-frequency spectrum signals of the measured sound waves under different working conditions, and is divided into training sets and validation sets in appropriate proportions; The hidden layer is responsible for extracting and learning the sound wave features, including the convolution layer, feature extraction sampling layer, pooling layer and fully connected layer; The output layer is responsible for exporting the acceleration curve identified as the train sound wave as the recognition result; Positioning module, including: The feature fusion module is used to extract fusion features of different scales and different time points in the Mel frequency cepstral coefficient map, and use the Mel frequency cepstral coefficient to locate the PICC catheter. By converting the audio signal into a set of feature vectors, it can better simulate the characteristics of the human auditory system, thereby extracting more accurate sound wave features, so that the Mel frequency cepstral coefficient has higher robustness when processing sound wave signals, and can effectively identify and locate the sound wave signals of the PICC catheter, thereby improving the positioning accuracy; The position determination module is used to input the extracted fusion features into the acoustic signal recognition neural network model to analyze the fusion features, and output the position information of the PICC catheter according to the analysis results. By positioning the PICC catheter, it can ensure the smooth delivery of drugs and improve the therapeutic effect of drugs, especially for patients who need long-term infusion or chemotherapy, it can avoid damage to surrounding tissues, reduce the occurrence of complications such as phlebitis and infection, and protect the patient's peripheral veins from corrosion by chemotherapy drugs; The real-time monitoring module is used to monitor the patient's blood flow dynamic parameters in real time. When the patient's blood flow dynamic parameters are abnormal, corresponding warnings are issued to achieve continuous monitoring and management of the patient's condition.

[0025] Extract the fusion features of the Mel frequency cepstral coefficient map, including Several Mel-frequency cepstral coefficient feature vectors of different scales are concatenated into a high-dimensional vector and input into the compressed time-delay network; The output of the compressed time-delay network is used as the sound wave feature and concatenated with several Mel-frequency cepstral coefficient feature vectors at different time points to obtain a mixed feature. Convolutional neural network is used to process mixed features, and mixed features are learned through convolutional layers. Several Mel-frequency cepstral coefficient features of different scales and time points are fused to obtain the fused features of Mel-frequency cepstral coefficient map.

[0026] Working principle: When using a PICC catheter ultrasound puncture positioning system of the present invention, according to Figure 1 and Figure 2 , including the following steps: Step 1: using an implantable sensor to collect the position where the PICC catheter enters and the acoustic wave signal generated during the entry process of the PICC catheter, and processing the acoustic wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients, and saving the two-dimensional matrix in an image format; Step 2: Use the deep learning algorithm of the neural network to establish a neural network model for acoustic wave signal recognition, and divide the acoustic wave signal features into a training set and a validation set; Step 3: Use the training set to train the acoustic signal recognition neural network model, adjust the hyperparameters of the acoustic signal recognition neural network model through the back propagation algorithm, use the validation set to evaluate the performance of the acoustic signal recognition neural network model, adjust the parameters of the acoustic signal recognition neural network model according to the evaluation results, and optimize the performance of the acoustic signal recognition neural network model; Step 4: Concatenate several Mel-frequency cepstral coefficient feature vectors of different scales into a high-dimensional vector, and concatenate them with several Mel-frequency cepstral coefficient feature vectors at different time points to obtain mixed features. Through the convolution layer, fuse several Mel-frequency cepstral coefficient features of different scales and different time points to obtain the fused features of the Mel-frequency cepstral coefficient map. Step 5: Input the extracted fusion features into the acoustic signal recognition neural network model to analyze the fusion features, and output the location information of the PICC catheter based on the analysis results.

[0027] In summary, the PICC catheter ultrasonic puncture positioning system of the present invention uses Mel-frequency cepstrum coefficients to locate the PICC catheter. By converting the audio signal into a set of feature vectors, it can better simulate the characteristics of the human auditory system, thereby extracting more accurate sound wave features, so that the Mel-frequency cepstrum coefficients have high robustness when processing sound wave signals, and can effectively identify and locate the sound wave signals of the PICC catheter, thereby improving the positioning accuracy. When processing the audio signal, the signal will be pre-processed, including windowing, denoising, smoothing and other operations to remove interference factors such as noise and burst sounds, which can effectively reduce background noise. The interference of background noise on PICC catheter positioning is improved, and the positioning accuracy is maintained at a high level under different environments and conditions. Whether in a noisy environment or a quiet environment, the Mel-frequency cepstral coefficient can effectively extract the acoustic characteristics of the PICC catheter, thereby achieving accurate positioning. By positioning the PICC catheter, it can ensure the smooth delivery of drugs and improve the therapeutic effect of drugs, especially for patients who need long-term infusion or chemotherapy. It can avoid damage to surrounding tissues, reduce the occurrence of complications such as phlebitis and infection, and protect patients' peripheral veins from corrosion by chemotherapy drugs.

[0028] In order to achieve more intuitive positioning observation, in one embodiment, the PICC catheterization ultrasound puncture positioning system further includes: The model building module is used to build a position marking model according to the position information of the PICC catheter and the pre-configured human body model of the patient; the position marking model marks the movement trajectory of the PICC catheter; The data collection module also includes: A signal receiving module, used for receiving the acoustic wave signal of the implantable sensor; The signal receiving module is configured as at least two ultrasonic receivers arranged at different positions; a positioning module is configured next to the ultrasonic receiver, and the ultrasonic receivers are respectively configured at the end of a manipulator; The model building module identifies the ultrasonic receiver into the position marking model according to the positioning module when building the position marking model; The robot's controller performs the following operations: Tracking the position of the end of the PICC catheter on the position marking model, and generating a position sequence of the end of the PICC catheter in the human body; the position sequence is composed of the coordinates of the positions of the end of the PICC catheter of a preset number (any one of 2 to 10); for example, the first position in the position sequence is the position of the most recent PICC catheter, followed by the positions of the PICC catheter at the previous moment; Determine the prediction sequence according to the position sequence and the pre-configured position prediction library; the position sequence in the position prediction library is associated with the prediction sequence in a one-to-one correspondence; Determine the layout area of ​​the receiver corresponding to each predicted position in the prediction sequence according to a preset layout area library; the predicted position in the layout area library is associated with the layout area in a one-to-one correspondence; The center point of the intersection area of ​​the layout area corresponding to each predicted position is used as the target position of the end of the manipulator to control the movement of the manipulator. The position of the receiver is adjusted by tracking the position of the PICC catheter to further ensure the effectiveness and accuracy of signal reception; In order to realize real-time risk analysis and remind the implementers to pay attention to the risks, the PICC catheterization ultrasound puncture positioning system also includes: a risk assessment module, which is used to evaluate and output the operation risks of the PICC catheter in real time; Among them, the risk assessment module performs the following operations: Obtain the position relationship of the implementer relative to the patient and retrieve the corresponding personnel model according to the implementer's current action; Construct simulation scenarios based on location marking models, location relationships and personnel models; Track the actions of the implementers, determine the risky parts and mark them on the personnel model in the simulation scene; the risky parts are first determined by determining the distance between each part of the implementer and the end of the robot; when the distance is less than or equal to the preset threshold and the distance gradually decreases, the part is judged as a risky part; In the simulation scenario, the moving trajectory of the manipulator is marked and the first risk value is calculated based on the positional relationship between each trajectory point and the risk part. The calculation formula of the first risk value is as follows: In the formula, represents the first risk value; Indicates The trajectory points and risk positions The distance between the sampling points; Shidi The trajectory points and risk positions The preset risk factor corresponding to each sampling point; is the total number of trajectory points; is the total number of sampling points; the sampling rules for risk parts are pre-configured, that is, the risk parts are sampled according to the sampling rules to obtain sampling points; for example, the sampling rules for the hand are the end, the joint of the phalanges, and the joint of the wrist, etc.; in addition, the trajectory points are obtained by sampling the moving trajectory according to the pre-configured trajectory sampling rules; The moving speed of the PICC catheter end and the position of the predicted sequence are evaluated to obtain the second risk value; the calculation formula of the second risk value is as follows: In the formula, represents the second risk value; Indicates The hazard value of each predicted location; Indicates The preset weight coefficients of the predicted positions; Indicates the movement speed of the PICC catheter tip; is the risk conversion factor of the preset corresponding speed; is the standard speed configured; The first risk value and the second risk value are output. The first risk value indicates the interference of the implementation personnel's actions on the positioning system, and the second risk value indicates that there is a risk in the puncture action of the PICC catheter. Marking both can remind the implementation personnel to perform the operation more attentively.

[0029] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0030] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A PICC catheterization ultrasonic puncture positioning system, characterized in that: include: A data collection module is used to collect the position of the PICC catheter and the acoustic wave signal generated by the PICC catheter during the insertion process using an implantable sensor, and to process the acoustic wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients, and save the two-dimensional matrix in an image format; A recognition model is used to establish a neural network model for acoustic signal recognition, and to train and optimize the neural network model for acoustic signal recognition; The positioning module is used to obtain the fusion features of different scales and different time points in the Mel frequency cepstral coefficient map, and use the acoustic signal recognition neural network model to analyze the fusion features to obtain the location information of the PICC catheter.

2. A PICC catheterization ultrasonic puncture positioning system according to claim 1, characterized in that: The data collection module comprises: An implantable sensor is placed inside the PICC catheter to obtain the position where the PICC catheter enters and the acoustic wave signal generated during the entry of the PICC catheter; The processing module is used to process the acquired sound wave signal and extract the characteristics of the sound wave signal to obtain a two-dimensional matrix of Mel-frequency cepstral coefficients, and save the two-dimensional matrix in an image format.

3. A PICC catheterization ultrasonic puncture positioning system according to claim 2, characterized in that: The processing module comprises: The feature extraction module is used to filter, emphasize, frame and process the acquired sound wave signal with a window function, and extract the sound wave signal features from the converted sound wave signal to obtain a two-dimensional matrix of Mel frequency cepstrum coefficients; The image conversion module is used to save the two-dimensional matrix of Mel-frequency cepstral coefficients into image format, where the rows represent the number of Mel filters and the columns represent the time frames.

4. A PICC catheterization ultrasonic puncture positioning system according to claim 3, characterized in that: The processing module specifically includes: A first-order high-pass filter is used for pre-emphasis processing, the leakage acoustic wave signal after pre-emphasis processing is divided into frames, and each frame is multiplied by a window function for windowing; Perform fast Fourier transform on each frame signal after frame division and windowing to obtain the spectrum of each frame, convert the actual frequency length into Mel frequency length, configure the triangular filter bank and calculate the output of each triangular filter after filtering the signal amplitude spectrum; Perform logarithmic operation on all filter outputs and further perform discrete cosine transform to obtain a two-dimensional matrix of Mel-frequency cepstral coefficients, and save the two-dimensional matrix of Mel-frequency cepstral coefficients in image format.

5. The PICC catheterization ultrasonic puncture positioning system according to claim 1, characterized in that: The identification module comprises: A model building module is used to build a neural network model for acoustic signal recognition using a deep learning algorithm of a neural network, and divide the acoustic signal features into a training set and a validation set; A model training module is used to train the acoustic signal recognition neural network model using a training set, and adjust the hyperparameters of the acoustic signal recognition neural network model through a back propagation algorithm; The model evaluation module is used to evaluate the performance of the acoustic signal recognition neural network model using the validation set, adjust the parameters of the acoustic signal recognition neural network model according to the evaluation results, and optimize the performance of the acoustic signal recognition neural network model.

6. A PICC catheterization ultrasonic puncture positioning system according to claim 5, characterized in that: The model building module specifically includes: Use the deep learning algorithm of neural network to establish a neural network model for acoustic signal recognition. The neural network model consists of an input layer, a hidden layer, and an output layer. The data set required by the input layer is derived from the time domain signals and time-frequency spectrum signals of the measured sound waves under different working conditions, and is divided into training sets and validation sets in appropriate proportions; The hidden layer is responsible for extracting and learning the sound wave features, including the convolution layer, feature extraction sampling layer, pooling layer and fully connected layer; The output layer is responsible for exporting the acceleration curve identified as the train sound wave as the recognition result.

7. The PICC catheterization ultrasonic puncture positioning system according to claim 1, characterized in that: The positioning module comprises: Feature fusion module, used to extract fusion features of different scales and different time points in the Mel frequency cepstral coefficient graph; A location determination module is used to input the extracted fusion features into the acoustic wave signal recognition neural network model to analyze the fusion features, and output the location information of the PICC catheter according to the analysis results; The real-time monitoring module is used to monitor the patient's blood flow dynamics parameters in real time and issue corresponding warnings when the patient's blood flow dynamics parameters are abnormal.

8. The PICC catheterization ultrasonic puncture positioning system according to claim 7, characterized in that: The fusion features of the extracted Mel frequency cepstral coefficient map specifically include Several Mel-frequency cepstral coefficient feature vectors of different scales are concatenated into a high-dimensional vector and input into the compressed time-delay network; The output of the compressed time-delay network is used as the sound wave feature and concatenated with several Mel-frequency cepstral coefficient feature vectors at different time points to obtain a mixed feature. Convolutional neural network is used to process mixed features, and mixed features are learned through convolutional layers. Several Mel-frequency cepstral coefficient features of different scales and time points are fused to obtain the fused features of Mel-frequency cepstral coefficient map.

9. The PICC catheterization ultrasonic puncture positioning system according to claim 1, characterized in that: Also includes: A model building module, used for building a position marking model according to the position information of the PICC catheter and a pre-configured human body model of the patient; The data collection module also includes: A signal receiving module, used for receiving the acoustic wave signal of the implantable sensor; The signal receiving module is configured as at least two ultrasonic receivers arranged at different positions; a positioning module is configured next to the ultrasonic receiver, and the ultrasonic receivers are respectively configured at the end of a manipulator; The model building module identifies the ultrasonic receiver into the position marking model according to the positioning module when building the position marking model; The robot's controller performs the following operations: Track the position of the end of the PICC catheter on the position marking model and generate a position sequence of the end of the PICC catheter in the human body; Determine a prediction sequence based on the position sequence and a pre-configured position prediction library; Determine the layout area of ​​the receiver corresponding to each prediction position in the prediction sequence according to the preset layout area library; The center point of the intersection area of ​​the layout areas corresponding to each predicted position is used as the target position of the end of the manipulator to control the movement of the manipulator.

10. A PICC catheterization ultrasonic puncture positioning system according to claim 9, characterized in that: Also includes: Risk assessment module, used to conduct real-time assessment and output of PICC catheter operation risks; Among them, the risk assessment module performs the following operations: Obtain the position relationship of the implementer relative to the patient and retrieve the corresponding personnel model according to the implementer's current action; Construct simulation scenarios based on location marking models, location relationships and personnel models; Track the actions of the implementers, identify the risk areas and mark them on the personnel models in the simulation scenario; In the simulation scenario, the moving trajectory of the manipulator is marked and the first risk value is calculated based on the positional relationship between each trajectory point and the risk part. The calculation formula of the first risk value is as follows: In the formula, represents the first risk value; Indicates The trajectory points and risk positions The distance between the sampling points; Shidi The trajectory points and risk positions The preset risk factor corresponding to each sampling point; is the total number of trajectory points; is the total number of sampling points; The moving speed of the PICC catheter end and the position of the predicted sequence are evaluated to obtain the second risk value; the calculation formula of the second risk value is as follows: In the formula, represents the second risk value; Indicates The hazard value of each predicted location; Indicates The preset weight coefficients of the predicted positions; Indicates the movement speed of the PICC catheter tip; is the risk conversion factor of the preset corresponding speed; is the standard speed configured; The first risk value and the second risk value are output.

Images