Medical instrument air tightness detection system and detection method

Abstract

The invention provides a medical instrument airtightness detection system and method, and the method comprises the steps: collecting an image of a to-be-detected medical instrument based on a data collection module, and determining the type of the to-be-detected medical instrument according to the collected image of the medical instrument based on a medical instrument recognition module. A to-be-detected medical instrument is placed in the sealed cavity module, the sealed cavity module is filled with gas with the corresponding pressure through the pressure control module according to the type of the to-be-detected medical instrument, and pressure change data and surrounding environment information at the corresponding moment are collected through the data collection module; the collected pressure change data is preprocessed based on the data preprocessing module, leakage judgment is performed through the airtightness detection module according to the category of the to-be-detected medical instrument, the surrounding environment information and the preprocessed pressure change data, and a judgment result is output based on the output module. According to the invention, airtightness leakage detection of the medical instrument can be realized more sensitively, the detection precision is high, and the use is convenient.

Description

Technical Field

[0001] The present invention relates to the technical field of medical device detection, and in particular to a medical device air tightness detection system and detection method. Background Art

[0002] The airtightness of medical devices is an important indicator to ensure their safety and effectiveness, especially in medical devices involving implantation, closed transportation or fluid control, where airtightness requirements are particularly strict. Currently, existing airtightness testing equipment usually has the following problems:

[0003] 1. The detection accuracy is insufficient and cannot meet the precise measurement requirements of tiny leaks;

[0004] 2. Low degree of automation, more manual operations and low efficiency;

[0005] 3. The adaptability to the testing environment is poor, and it is difficult to meet the complex and diverse testing needs of medical devices.

[0006] Therefore, it is necessary to design a medical device air tightness detection system and detection method. Summary of the invention

[0007] In order to overcome the deficiencies of the prior art, an object of the present invention is to provide a medical device airtightness detection system and detection method.

[0008] To achieve the above object, the present invention provides the following solutions:

[0009] The present invention provides a medical device airtightness detection system, comprising: a sealing cavity module, a pressure control module, a data acquisition module, a medical device identification module, a data preprocessing module, an airtightness detection module and an output module;

[0010] Medical equipment is arranged inside the sealed cavity module;

[0011] The pressure control module is connected to the sealed cavity module and is used to restart the sealed cavity and perform pressure detection;

[0012] The data acquisition module is used to collect pressure changes inside the sealed cavity, images of medical devices and surrounding environment information;

[0013] The medical device identification module is used to identify the medical device based on the image of the medical device;

[0014] The data preprocessing module is used to preprocess the pressure change data inside the sealed cavity;

[0015] The airtight detection module is used to make leakage judgment based on the pre-processed pressure change data and surrounding environment information according to the medical device identification result;

[0016] The output module is used to output the judgment result.

[0017] Preferably, the sealed cavity module includes a cavity made of a transparent pressure-resistant material and a quick sealing joint arranged on the cavity, the pressure control module includes an air source, a pressure regulating device and a pressure sensor, the cavity is connected to the pressure regulating device through a quick sealing joint, the pressure regulating device is connected to the air source, the pressure sensor is arranged inside the cavity, and the pressure sensor is connected to the data acquisition module.

[0018] The present invention also provides a method for detecting air tightness of a medical device, comprising:

[0019] Step 100: collecting images of the medical device to be detected based on the data collection module, and determining the category of the medical device to be detected based on the collected images of the medical device based on the medical device recognition module;

[0020] Step 200: placing the medical device to be tested inside the sealed cavity module, filling the sealed cavity module with gas of corresponding pressure through the pressure control module according to the type of the medical device to be tested, and collecting pressure change data and surrounding environment information at the corresponding time through the data acquisition module;

[0021] Step 300: preprocessing the collected pressure change data based on the data preprocessing module;

[0022] Step 400: using the airtight detection module to make a leakage judgment based on the type of the medical device to be detected, the surrounding environment information and the pre-processed pressure change data;

[0023] Step 500: Output the judgment result based on the output module.

[0024] Preferably, in step 100, the category of the medical device to be detected is determined based on the collected image of the medical device by the medical device identification module, specifically:

[0025] Construct a medical device recognition model based on the improved ResNet34 network structure;

[0026] Train the medical device recognition model based on the preset data set;

[0027] The medical device to be detected is input into the trained medical device recognition model to obtain the category of the medical device to be detected.

[0028] Preferably, the medical device recognition model is divided into 9 parts, namely, convolution layer, maximum pooling layer Maxpool, Stage 1, Stage 2, SE, Stage 3, Stage 4, average pooling layer Avgpool and fully connected layer FC.

[0029] Preferably, in step 300, the collected pressure change data is preprocessed based on the data preprocessing module, specifically:

[0030] The pressure change data is obtained and denoised based on the IWOA-VMD algorithm.

[0031] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0032] The present invention provides a medical device air tightness detection system and detection method, the method includes collecting images of medical devices to be detected based on a data acquisition module, determining the category of the medical devices to be detected based on the collected images of the medical devices based on a medical device identification module, placing the medical devices to be detected inside a sealed cavity module, filling the sealed cavity module with gas of corresponding pressure through a pressure control module according to the category of the medical devices to be detected, collecting pressure change data and surrounding environment information at the corresponding moment through a data acquisition module, preprocessing the collected pressure change data based on a data preprocessing module, making leakage judgments based on the category of the medical devices to be detected, surrounding environment information and preprocessed pressure change data through an airtight detection module, and outputting the judgment results based on an output module. The present invention can more sensitively realize the airtightness detection leakage of medical devices, has high detection accuracy, and is easy to use. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0034] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0035] Figure 2 Identify model structure diagrams for medical devices;

[0036] Figure 3 Schematic diagram of the decomposition of the convolution kernel. DETAILED DESCRIPTION

[0037] 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.

[0038] The purpose of the present invention is to provide a medical device air tightness detection system and detection method, which can more sensitively realize the medical device air tightness leakage detection, have high detection accuracy, and are easy to use.

[0039] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] The present invention provides a medical device airtightness detection system, comprising: a sealing cavity module, a pressure control module, a data acquisition module, a medical device identification module, a data preprocessing module, an airtightness detection module and an output module;

[0041] The sealed cavity module is provided with a medical device inside to place the medical device to be tested in a sealed environment to prevent external interference;

[0042] The pressure control module is connected to the sealed cavity module and is used to restart the sealed cavity and perform pressure detection;

[0043] The data acquisition module is used to collect pressure changes inside the sealed cavity, images of medical devices and surrounding environment information;

[0044] The medical device identification module is used to identify the medical device based on the image of the medical device;

[0045] The data preprocessing module is used to preprocess the pressure change data inside the sealed cavity;

[0046] The airtight detection module is used to make leakage judgment based on the pre-processed pressure change data and surrounding environment information according to the medical device identification result;

[0047] The output module is used to output the judgment result.

[0048] It should be noted that the present invention also includes an actuator module, including an automatic mechanical arm and a clamping device, for realizing automatic placement and sealing operations of medical devices.

[0049] The sealed cavity module includes a cavity made of a transparent pressure-resistant material and a quick sealing joint arranged on the cavity. The pressure control module includes an air source, a pressure regulating device and a pressure sensor. The cavity is connected to the pressure regulating device through a quick sealing joint, and the pressure regulating device is connected to the air source. The pressure sensor is arranged inside the cavity, and the pressure sensor is connected to the data acquisition module.

[0050] All the devices mentioned above can be existing devices in the prior art. The present invention does not limit their models, and they only need to achieve specific functions.

[0051] like Figure 1 As shown, the present invention also provides a method for detecting air tightness of a medical device, comprising:

[0052] Step 100: collecting images of the medical device to be detected based on the data collection module, and determining the category of the medical device to be detected based on the collected images of the medical device based on the medical device recognition module;

[0053] Step 200: placing the medical device to be tested inside the sealed cavity module, filling the sealed cavity module with gas of corresponding pressure through the pressure control module according to the type of the medical device to be tested, and collecting pressure change data and surrounding environment information at the corresponding time through the data acquisition module;

[0054] Step 300: preprocessing the collected pressure change data based on the data preprocessing module;

[0055] Step 400: using the airtight detection module to make a leakage judgment based on the type of the medical device to be detected, the surrounding environment information and the pre-processed pressure change data;

[0056] Step 500: Output the judgment result based on the output module.

[0057] In step 100, the category of the medical device to be detected is determined based on the collected image of the medical device by the medical device identification module, specifically:

[0058] 1. Build a medical device recognition model based on the improved ResNet34 network structure, specifically:

[0059] In order to meet the application requirements of high-precision recognition of medical device images, an improved residual network model, SE-ResNet36, was constructed based on the ResNet34 network. The overall architecture of the network model is as follows: Figure 2 As shown, the network structure is divided into 9,parts, namely, convolutional layer, maximum pooling layer Maxpool, Stage1, Stage2, SE, Stage3, Stage4, average pooling layer Avgpool and fully connected layer FC;

[0060] The SE-ResNet36 network model is based on ResNet34. It decomposes the first convolution layer in the original network model, splits a convolution layer with a convolution kernel of 7 into three convolution layers with a convolution kernel of 3, and adds an SE attention module after Stage2 to focus on the relationship between channels, so that the model can automatically learn the importance of different channel features. The overall network model architecture of SE-ResNet36 and its internal parameters are shown in Table 1;

[0061] Table 1 The overall network model architecture and internal parameters of SE-ResNet36

[0062]

[0063] The coarse-grained features extracted by the 7×7 large-size convolution kernel of the first layer of ResNet34 are not conducive to the recognition of medical device analogies, affecting the recognition effect of the network model on medical devices. In order to more accurately identify medical devices, it is necessary to extract more effective subtle features. Therefore, when designing the network model, the first convolution layer with a convolution kernel of 7 is split and connected in series by three convolution layers with a convolution kernel of 3 to replace the original large-size convolution kernel, such as Figure 3 As shown;

[0064] The SE attention module is proposed in the SENet network model. It is a type of channel attention module, which can make the network pay more attention to channel information and enhance the model learning ability. The SE module has the characteristics of plug-and-play and can be directly added to other convolutional neural networks. It can effectively improve the recognition accuracy of the model. The SE module can be divided into two processes: compression (Squeeze) and excitation (Excitation). First, the Squeeze operation is used to compress the feature map, and then the channel feature learning is performed on it. Then the Excitation operation is performed to give the feature map different channel weights from the perspective of the channel, and finally a feature map with channel attention is obtained. The SE attention mechanism enables the model to strengthen the channel features of the input feature map, highlight important features, and suppress unimportant features, thereby improving the performance of the model;

[0065] 2. Train the medical device recognition model based on a preset data set, where pictures of various medical devices can be directly collected and image enhancement can be performed on them, for example, image rotation processing, color dithering processing and Gaussian noise processing can be performed. After enhancement, the medical device images are annotated based on the annotation software to obtain the preset data set;

[0066] 3. Input the medical device to be detected into the trained medical device recognition model to obtain the category of the medical device to be detected.

[0067] In step 200, the medical device to be detected is placed inside the sealed cavity module, and the gas of corresponding pressure is filled into the sealed cavity module through the pressure control module according to the type of the medical device to be detected, and the pressure change data and the surrounding environment information at the corresponding time are collected through the data acquisition module, specifically:

[0068] The medical device to be tested is placed inside the sealed cavity module, and the gas of corresponding pressure is filled into the sealed cavity module through the pressure control module according to the type of the medical device to be tested. Different pressure stabilization times, gas types, pressures, etc. are set for each medical device based on practical experience. After the filling is completed, the set pressure is maintained for a certain period of time, and the pressure change inside the sealed cavity module is detected by the pressure sensor. In addition, it is necessary to detect the surrounding environmental information during the detection as additional information.

[0069] In step 300, the collected pressure change data is preprocessed based on the data preprocessing module, specifically:

[0070] The pressure change data is obtained, and the collected pressure change data is used as the original time series data, and denoising is performed on it based on the IWOA-VMD algorithm. The specific steps include:

[0071] Initialize the VMD algorithm parameter optimization range, and find the parameter combination corresponding to the minimum value of the composite fitness function according to the IWOA algorithm;

[0072] Substitute the optimal parameter combination into the VMD decomposition model to decompose the original time series data into a series of IMFs;

[0073] The multi-scale permutation entropy-variance contribution rate selection principle is used to calculate and identify different IMFs, classify them into noise-dominated components and effective signal components, and perform NLM denoising on the noise-dominated components to obtain the denoised components.

[0074] Reconstruct the denoised component and the effective signal component to obtain the denoised original time series data;

[0075] A detailed introduction to the variational mode decomposition (VMD) algorithm;

[0076] As a very active branch in the field of adaptive signal decomposition, VMD algorithm has played a pivotal role in the field of signal processing in recent years. The core idea of ​​VMD algorithm is to adaptively decompose the input signal f(t) into k k nearby modal functions u k (t), while minimizing u kThe bandwidth and of (t). The VMD algorithm decomposition process is described as follows:

[0077] Let the kth modal function u k The expression of (t) is:

[0078] u k (t) = A k (t)cos[φ k (t)]

[0079] In the formula, A k (t) is u k (t) and A k (t)≥0,φ k (t) is u k The instantaneous phase of (t) is a non-decreasing function.

[0080] For k modal functions u k (t) Perform Hilbert transform calculation to obtain the corresponding single-sided spectrum. The relevant analytical signal is:

[0081]

[0082] Where δ(t) is the unit pulse function and * represents convolution.

[0083] By multiplying the analytical signal by Get the baseband signal:

[0084]

[0085] In order to obtain the bandwidth of each modal function, signal demodulation is achieved through Gaussian smoothing. The constrained variational model is as follows:

[0086]

[0087] Where K is the preset number of decomposed modes.

[0088] In order to solve the above constrained variational model, the Lagrange multiplier λ and the quadratic penalty factor α are introduced, and the above formula is transformed into an unconstrained variational model as shown below:

[0089]

[0090] The modal function {u k}、Center frequency {ω k} and the Lagrange multiplier λ are continuously updated and iterated, and the formula is as follows:

[0091]

[0092] Where τ is the time step.

[0093] When the following conditions are met, the ADMM algorithm converges and the above update iteration process terminates:

[0094]

[0095] Where c is the convergence parameter, usually ε = 1 × 10 -6 .

[0096] A detailed introduction to the whale optimization algorithm (WOA):

[0097] The whale optimization algorithm is a new swarm intelligence optimization algorithm based on iterative search of humpback whale populations. It models its predation behavior to achieve optimization of complex optimization problems. According to its algorithm characteristics, it is divided into three stages: searching for prey, surrounding prey, and spiral bubble net predation.

[0098] (1) Searching for prey

[0099] In this stage, according to the random walk mechanism, each whale participating in the hunt iteratively updates its position based on each other's current position. According to the t-th search position, when the coefficient |A|>1, the whale group randomly searches for high-quality prey in the global range, and its t+1-th position update expression is:

[0100] D=|C·x rand (t)-x(t)|

[0101] x(t+1)=x rand (t)-A·D

[0102] Among them, t is the current iteration number, x rand (t) is the arbitrary position of the whale, x(t) is the current position, and:

[0103] A=2ar1-a

[0104] C=2r2

[0105] Among them, r1,r2∈rand[0,1],a=2-2t / T, and T is the number of selections. According to the linear decreasing property of a, A gradually decreases with a. When |A|≤1, the algorithm then enters the stage of surrounding the prey.

[0106] (2) Encirclement of prey

[0107] After discovering the target prey, the whale group will gradually narrow its search range and update the search position according to the best individual in the current population. The corresponding expression is:

[0108] D1=|C·x best (t)-x(t)|

[0109] x(t+1)=x best (t)-A·D

[0110] In the formula, x best (t) is the best individual position in the current whale group.

[0111] (3) Spiral bubble net predation stage

[0112] Whales have a unique hunting method, using the bubble net they spit out to spiral upward and shrink to surround and approach the target. The mathematical model is as follows:

[0113] D2=|x best (t)-x(t)|

[0114] x(t+1)=D2e bl cos(2πl)+x best (t)

[0115] Where b is a constant coefficient and l is a random number in the interval [-1, 1].

[0116] The hunting phase is divided by setting random probability. When the random probability p<0.5, the whale updates the best individual position; when p≥0.5, the whale gradually approaches the target. During the approach process, |A| decreases continuously, and the optimal solution is obtained when |A|=0. The overall hunting mechanism is described as follows:

[0117]

[0118] Next, we introduce the improvement strategy of the whale optimization algorithm:

[0119] The traditional whale optimization algorithm adjusts global search and local development through coefficient A. The value of A is mainly related to a. However, the linearly changing convergence factor a has disadvantages in early search capability and later iteration speed. In order to improve its impact on algorithm performance, a piecewise nonlinear convergence factor m is proposed without affecting the overall change trend. Its expression is as follows:

[0120]

[0121] During the entire iteration process, m shows a nonlinear decreasing trend. In the early stage, the larger parameter m gradually decreases, which can improve the global exploration capability; in the later stage, the smaller parameter m can increase the convergence speed of the algorithm.

[0122] In addition, considering the impact of inertia weight on the algorithm's optimization ability, an adaptive weight strategy is proposed here. The formula is as follows:

[0123]

[0124] So the position update formula becomes:

[0125]

[0126] The traditional WOA algorithm is prone to fall into local optimality. Here, the Cauchy operator is introduced to give the optimal individual Cauchy perturbation. The Cauchy operator can generate larger perturbations to make the algorithm escape from the local optimality, and can also generate smaller perturbations to increase the convergence speed.

[0127] Therefore, the present invention uses the following Cauchy variation formula to update the best individual position as follows:

[0128] x new (t) = x best (t)·[1+Cauchy(0,1)]

[0129] In the formula, x new (t) is the new value of the current optimal value after disturbance, and Cauchy(0,1) is the Cauchy operator.

[0130] The fitness function of the IWOA-VMD algorithm is:

[0131] Energy fan can characterize the energy distribution of the signal in the frequency domain. The smaller its value is, the higher the energy proportion of the corresponding component is. The energy calculation formula of each component is:

[0132]

[0133] Where N is the signal length. The energy probability density formula obtained by normalizing it is:

[0134]

[0135] The energy calculation formula is:

[0136]

[0137] In practical applications, if energy fanning is used as a single fitness function, it is easy to decompose the signal into simple harmonic waves. This is because the signal is over-decomposed in pursuit of energy concentration. In order to improve the over-decomposition of the signal, kurtosis is introduced to describe the waveform characteristics of the signal. Kurtosis is a dimensionless parameter that characterizes the degree of data spikes and is often used for fault diagnosis of nonlinear signals. Its expression is as follows:

[0138]

[0139] In the formula, μ and σ represent the mean and variance of the signal respectively; E(x-μ) 4 represents the fourth-order mathematical expectation.

[0140] According to the characteristics of wireless signals, a composite fitness function is constructed by combining energy fanning and kurtosis. By iteratively searching for the minimum value of the composite fitness function, the (K, α) parameter combination with the best decomposition effect is found. The fitness function calculation formula is as follows:

[0141] F=1 / Ku+0.1×H E

[0142] The principle of selecting the optimal modal component of multi-scale permutation entropy-variance contribution rate is introduced:

[0143] Multi-scale permutation entropy (MPE) introduces permutation entropy of different scales and comprehensively considers the signal characteristics at multiple scales in a weighted average manner to describe the randomness and complexity of the signal. The basic principle of MPE is as follows:

[0144] 1. For the time series x={x1,x2,…,x L Multi-scale coarse-graining

[0145]

[0146] Where s is the scale factor, is a multi-scale time series. When s is set to 1, For the original time series, the permutation entropy can be calculated.

[0147] 2. Time series Reconstructing the phase space

[0148]

[0149] Where τ is the time delay factor, m is the embedding dimension, To reconstruct the sequence.

[0150] 3. Reconstruction Arrange in ascending order, and you can get m! permutations and combinations, and the number of times each permutation occurs is N t , the corresponding probability is Right now:

[0151]

[0152]

[0153] 4. Scale factor s Arrange the calculation and normalize it

[0154]

[0155] As a data statistics theory based on the Gaussian hypothesis, principal component analysis can reduce the redundancy and invalid features of data components, eliminate data correlation to highlight the advantages and useful information of the original sample data. Variance contribution rate is a commonly used method for selecting principal components in principal component analysis. It mainly represents the proportion of IMF energy in the signal energy before decomposition. The larger the variance contribution rate, the stronger the IMF's ability to explain the original signal. Therefore, the IMF can be ranked according to the size of the variance contribution rate to evaluate the importance of the IMF. The calculation formula of variance contribution rate is as follows:

[0156]

[0157] In the formula, D(IMF k ) represents the variance of the kth IMF, V k represents the variance contribution rate of the kth IMF.

[0158] A detailed introduction to NLM algorithm denoising:

[0159] Based on the weighted average operation of similar structures in the image to improve the signal-to-noise ratio, the non-local mean filtering algorithm achieves the purpose of image denoising. The NLM algorithm screens all the information in the image, retains the obvious main features, and removes obviously irrelevant redundant components. By analyzing the similarity characteristics, it is found that the idea of ​​the NLM algorithm can also be applied to the field of signal filtering. Using the NLM algorithm for filtering, find as many similar values ​​as possible within the search range, and use the weighted coefficient to obtain the estimated value for weighted averaging to make it close to the true value in the signal, thereby achieving the purpose of noise reduction. The principle of the NLM algorithm is as follows:

[0160] Assume that the noisy signal y(t) is the superposition of the original signal u(t) and the external interference noise n(t), as shown in the following formula:

[0161] y(t)=u(t)+n(t)

[0162] The NLM algorithm estimates the original signal u(t) by calculating the weighted average K(t) of all similar blocks in y(t). The calculation process is as follows:

[0163]

[0164] Where ω(t, s) refers to the similarity between two search blocks centered at t and s respectively, N(t) represents the set of all points in the target search area centered at the target sample point t, Z(t) is the normalization constant, representing the sum of the similarities of all search blocks in the search domain, and ω(t, s) also needs to satisfy and These two basic conditions, and then calculate ω(t, s) as:

[0165]

[0166] Where L Δ is the neighborhood block centered at s, L Δ = 2P + 1. The parameter P affects the number of similar structural blocks found. △ is the search block centered at t. Take K as half of the length of the △ region. The parameter K affects the amount of calculation and calculation time. λ is the filter bandwidth parameter, which affects the smoothness of the filtered signal.

[0167] The selection of the three key parameters P, K, and λ greatly affects the denoising effect of the algorithm on one-dimensional signals. If the similar block parameter P is too small, it is easy to find more similar blocks within the search range. If P is too large, no similar blocks can be found. In the present invention, P is set to 10 < P < 20. Theoretically, the search block parameter K can search the entire domain, but this will increase the calculation difficulty of the algorithm and the calculation efficiency is low. The literature points out that when the signal length N > 4000, take 0.25N < K < 0.35N. The search range contains more signal information and reaches a balance in terms of the amount of calculation and calculation time. When N < 4000, take K = 0.5N. The search range can cover the entire domain and most of the signal information can be obtained. If the bandwidth parameter A is too small, it will cause noise fluctuations, which will interfere with the weighted average of similar blocks and the filtering effect is poor; if λ is too large, it will misjudge the similarity of search blocks, the signal is too smooth, and the filtering effect is too strong. In the present invention, when 0.3σ < λ < 0.6σ (σ is the standard deviation of the noisy signal), the filtering effect is better.

[0168] In step 400, the airtight detection module makes a leakage judgment according to the category of the medical device to be detected, the surrounding environment information, and the preprocessed pressure change data. Specifically:

[0169] AI models can be used to analyze the category of medical devices to be tested, surrounding environment information, and pre-processed pressure change data. For example, time series analysis models such as LSTM and Transformer are used to model the pressure change trend, and the category of medical devices to be tested and the surrounding environment information are used as additional dimensions of the model. The model is trained with a large amount of historical detection data to learn the characteristic patterns of different types of leaks (such as small leaks, intermittent leaks, and large leaks). During the detection process, the model analyzes the pressure change data in real time, and makes intelligent judgments on the leakage rate, leakage location, and leakage type to obtain the judgment results.

[0170] In step 500, the judgment result is output based on the output module, specifically:

[0171] The test results are presented in graphical form through the output module, including pressure change curve, leakage rate value and leakage type determination result.

[0172] The present invention can also be configured for self-learning and optimization:

[0173] For example, by updating the AI ​​model in the cloud, we can continuously learn and optimize the model parameters from new detection data to improve detection accuracy. We can also further enhance the model's ability to recognize special situations by labeling special detection samples (such as abnormal leakage cases).

[0174] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0175] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A medical device air tightness detection system, characterized in that: include: Sealing cavity module, pressure control module, data acquisition module, medical device identification module, data preprocessing module, airtight detection module and output module; Medical equipment is arranged inside the sealed cavity module; The pressure control module is connected to the sealed cavity module and is used to restart the sealed cavity and perform pressure detection; The data acquisition module is used to collect pressure changes inside the sealed cavity, images of medical devices and surrounding environment information; The medical device identification module is used to identify the medical device based on the image of the medical device; The data preprocessing module is used to preprocess the pressure change data inside the sealed cavity; The airtight detection module is used to make leakage judgment based on the pre-processed pressure change data and surrounding environment information according to the medical device identification result; The output module is used to output the judgment result.

2. The system according to claim 1, characterized in that The sealed cavity module includes a cavity made of a transparent pressure-resistant material and a quick sealing joint arranged on the cavity. The pressure control module includes an air source, a pressure regulating device and a pressure sensor. The cavity is connected to the pressure regulating device through a quick sealing joint, and the pressure regulating device is connected to the air source. The pressure sensor is arranged inside the cavity, and the pressure sensor is connected to the data acquisition module.

3. A method for detecting air tightness of medical devices, characterized in that: include: Step 100: collecting images of the medical device to be detected based on the data collection module, and determining the category of the medical device to be detected based on the collected images of the medical device based on the medical device recognition module; Step 200: placing the medical device to be tested inside the sealed cavity module, filling the sealed cavity module with gas of corresponding pressure through the pressure control module according to the type of the medical device to be tested, and collecting pressure change data and surrounding environment information at the corresponding time through the data acquisition module; Step 300: preprocessing the collected pressure change data based on the data preprocessing module; Step 400: using the airtight detection module to make a leakage judgment based on the type of the medical device to be detected, the surrounding environment information and the pre-processed pressure change data; Step 500: Output the judgment result based on the output module.

4. The method according to claim 3, characterized in that In step 100, the category of the medical device to be detected is determined based on the collected image of the medical device by the medical device identification module, specifically: Construct a medical device recognition model based on the improved ResNet34 network structure; Train the medical device recognition model based on the preset data set; The medical device to be detected is input into the trained medical device recognition model to obtain the category of the medical device to be detected.

5. The method according to claim 4, characterized in that The medical device recognition model is divided into 9 parts, namely convolution layer, maximum pooling layer Maxpool, Stage 1, Stage 2, SE, Stage 3, Stage 4, average pooling layer Avgpool and fully connected layer FC.

6. The method according to claim 3, characterized in that In step 300, the collected pressure change data is preprocessed based on the data preprocessing module, specifically: The pressure change data is obtained and denoised based on the IWOA-VMD algorithm.

Images