A micturition patency assessment system based on micturition conditions and parameters
By designing a urination patency assessment system, which uses a flow guide plate and multimodal sensors to collect urine information, the system solves the problem that urination range and abnormal symptoms cannot be quantified in existing technologies, and achieves accuracy and objectivity in urination patency assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- THE THIRD MEDICAL CENT OF THE CHINESE PEOPLES LIBERATION ARMY GENERAL HOSPITAL
- Filing Date
- 2026-01-07
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies cannot accurately quantify the range of urination and its dynamic changes, cannot distinguish and quantify abnormal urination symptoms, and lack objective detection methods, resulting in inaccurate assessment of urination patency.
Design a urination patency assessment system, including a urine spatial diversion and acquisition device, a multimodal sensing unit, and a central processing and control unit. The system diverts urine through a specific geometric arrangement of holes in a diversion cover and uses multimodal sensors to collect visual and urine volume information to generate multidimensional assessment parameters.
It enables the objective recording of dynamic changes in urination range and urine volume distribution, transforming patients' subjective symptoms into quantifiable data, improving the accuracy and scientific rigor of urinary dysfunction diagnosis, and providing multi-dimensional evidence for treatment.
Smart Images
Figure CN121891013B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical testing technology, specifically to a urination patency assessment system based on urination status and parameters. Background Technology
[0002] Currently, clinical assessment of urinary patency largely relies on uroflowmetry, a technique that provides basic parameters such as maximum urinary flow rate, average urinary flow rate, and voiding time. However, existing techniques based on total urine volume measurement have significant limitations. They only provide holistic, general parameters of the voiding process, failing to reflect crucial spatial information about the dynamic changes in urine during excretion. Specifically, current techniques cannot obtain the important parameter of voiding range and its changes over time. The distance, stability, and presence of attenuation or interruption of the range directly reflect the contractile strength of the bladder detrusor muscle and the patency of the urethra. However, for patients with weak or obstructed voiding, the urine flow often only reaches a short distance or is dribbling, which current techniques cannot objectively quantify.
[0003] Furthermore, current technologies struggle to accurately differentiate and quantify specific urinary abnormalities. For instance, objective detection methods are lacking for dribbling at the end of urination, relying primarily on subjective patient descriptions such as incomplete emptying or dribbling, leading to a lack of precise diagnostic evidence. Simultaneously, the inability to obtain spatial distribution of urine across different distances makes it difficult to assess the kinetic energy differences in urine flow across these distances, which is crucial for determining the location and severity of obstruction. Summary of the Invention
[0004] The purpose of this invention is to provide a urination patency assessment system based on urination status and parameters. The system includes a urine spatial diversion and acquisition device, a multimodal sensing unit, and a central processing and control unit. A flow guide cover with geometrically arranged holes allows for natural spatial diversion of urine. The multimodal sensors simultaneously acquire visual information of the urination range and urine volume information at each distance. The central unit processes the data to generate multidimensional assessment parameters. Therefore, subjective urination symptoms can be transformed into objective data, providing more accurate and comprehensive clinical evidence for assessing functions such as weak urination and urethral obstruction, thus solving the problems mentioned in the background art.
[0005] To achieve the above objectives, the present invention provides the following technical solution: A urination patency assessment system based on urination status and parameters, comprising: A urine spatial diversion and collection device includes a base with an inclined angle, a graded urine volume collection module array supported on the base, and a detachable guide cover that tightly covers the top of the array. The surface of the guide cover has multiple rows of holes, which form an isosceles triangle region with the apex pointing towards the user and the base away from the user. Each row of holes corresponds to an independent graded urine volume collection module below it, which is used to realize the natural diversion and collection of urine based on the distance of urination. A multimodal sensing unit includes an optical detection module for acquiring video information of the urination process and multiple weight sensors for real-time measurement of the accumulated urine volume in each graded urine volume collection module. The optical detection module includes high frame rate cameras respectively installed above and to the side of the aperture array area. The central processing and control unit is signal-connected to the multimodal sensing unit and is used to control the initialization of the evaluation session and the synchronous start of data acquisition. It receives and processes the video information and urine volume information, generates a urination event time series table by aligning with timestamps, and extracts evaluation parameters including urine volume spatial distribution map, urination range change curve and segmented urine flow rate characteristic value based on the time series table to generate a urination patency evaluation report.
[0006] Preferably, each of the graded urine collection modules includes: a sealed housing; a container for receiving urine, located at the top inside the sealed housing; a urine inlet pipe for collecting urine flowing in from the same outlet into the corresponding container; a urine outlet pipe connected to the bottom of the container; and a miniature solenoid valve installed on the urine outlet pipe and controlled by the central processing and control unit for automatically emptying the urine after the assessment.
[0007] Preferably, the central processing and control unit performs system self-test and calibration before data acquisition. The self-test includes checking whether all weight sensors are zeroed and their communication status, and checking whether the camera can start and focus normally. The calibration includes ensuring that the camera's field of view completely covers the entire triangular area of the hole array.
[0008] Preferably, the central processing and control unit performs the determination of the end of urination, and the determination condition is: the readings of all weight sensors do not change significantly within a set number of seconds, and the video image collected by the optical detection module shows that the urine wetted area no longer expands and no new droplets fall.
[0009] Preferably, the central processing and control unit preprocesses the received data, including filtering the weight sensor data to eliminate environmental vibration noise, and performing image enhancement and noise reduction on the camera video stream.
[0010] Preferably, the central processing and control unit performs the following specific actions when extracting evaluation parameters: Based on the urination event timeline, the total amount of urine collected by each row of graded urine collection modules is calculated to form a urine volume spatial distribution diagram. Based on the changes in the position of the leading edge of the urine wetting area dynamically tracked by background difference and color features in the video stream, a curve of urine jet range change is plotted. By analyzing the first derivative of the weight-time curve of the graded urine volume collection module corresponding to a specific distance segment, the segmented urine flow rate characteristic value of that distance segment is calculated.
[0011] Preferably, the central processing and control unit further includes a quantitative analysis module for identifying abnormal urination symptoms, and this module is configured as follows: If, during a period of time that exceeds a set proportion of the total urination time, the urine jet range is consistently below a preset range threshold and the weight accumulation rate of the corresponding distance segment is consistently below a preset rate threshold, then it is determined that there is a symptom of weak urination. When a periodic, pulse-like, minute weight step increase is detected only on the nearest row or two rows of graded urine collection modules within a set time interval, and the video shows discrete droplets falling, it is determined that dribbling symptoms exist, and a pulse sequence detection algorithm is used to count the number of dribbles, the total amount of dribbled urine, and the duration of dribbling.
[0012] Preferably, before evaluation, the system requires the application of a disposable sanitary barrier liner on the surface of the flow guide cover. This liner is made of a high-molecular nonwoven fabric material that allows for rapid liquid penetration but is not prone to splashing, and its size completely covers the array of isosceles triangular holes to form a sanitary barrier.
[0013] Preferably, the parameters included in the generated evaluation report include at least: a spatial distribution diagram of urine volume, a curve showing the change in voiding range, a time-series diagram of the superimposed total urine volume curve, total urine volume, total voiding time, average urine flow rate, maximum voiding range, average range, range fluctuation coefficient, effective voiding volume distribution ratio, percentage of dribbling urine volume, and dribbling duration.
[0014] Preferably, after determining that urination has ended and the user has left, the central processing and control unit controls all the micro solenoid valves to open, draining urine from each container, and monitors the readings of the weight sensor in real time. When the reading is detected to be zero, the central processing and control unit controls the solenoid valves to close, thus completing the device preparation.
[0015] Preferably, assessment parameters, including a spatial distribution map of urine volume, a curve showing the change in voiding range, and segmented urinary flow rate characteristics, are extracted based on the time series table to generate a voiding patency assessment report, including: The time series table extracts assessment parameters, including the spatial distribution map of urine volume, the change curve of urine range, and the segmented urine flow rate characteristic value, and inputs them into the pre-trained urination patency assessment model to generate a urination patency assessment report. The method for constructing the urination patency assessment model includes: Obtain the model training dataset; the training dataset includes urination video information collected by the optical detection module and urine volume information collected by the weight sensor; perform feature extraction on the training dataset to obtain multi-dimensional original features, the multi-dimensional original features include urine volume spatial distribution map, urination range change curve and segmented urine flow rate feature value; Based on the multi-dimensional original features, a three-dimensional sub-model is constructed, which includes a physical feature dimension sub-model, a dynamic time series dimension sub-model, and a distribution pattern dimension sub-model. The construction process of the physical feature dimension sub-model is as follows: calculate the spatial distribution uniformity factor and the effective volume utilization factor based on the urine volume spatial distribution map, map the two factors to the physical feature state index through the Sigmoid function, and construct an age-gender hierarchical benchmark library to determine the physical feature benchmark reference quantity. The construction process of the dynamic time-series dimension sub-model is as follows: the time-series fluctuation factor, peak concentration factor and time-series decay factor are calculated based on the voiding range change curve and segmented urine flow rate characteristic value. The dynamic time-series state index is obtained by weighted summation and normalization, and the dynamic time-series benchmark reference quantity is determined based on the hierarchical benchmark library. The construction process of the distribution pattern dimension sub-model is as follows: the effective distribution ratio factor and the dripping influence factor are calculated based on the segmented urine flow rate characteristic value and dripping related parameters; the distribution pattern state index is obtained through factor trade-off mapping; and the distribution pattern benchmark reference quantity is determined based on the hierarchical benchmark library. A dynamic weight adaptive mechanism is constructed. Based on the urine flow rate curve, the urination process is divided into three stages: initial, middle and final. Sliding windows are divided according to a preset step size and assigned to the corresponding stage. The kernel density estimation function of the three-dimensional state index is pre-constructed using global training data. The overlapping area of the probability density functions of the state index of the three-dimensional sub-model within each window is calculated. The overlapping area matrix is constructed and the calibration coefficient is obtained by feature decomposition. The weight within the window is calculated by combining the basic weight of the stage and the calibration coefficient. The global dynamic weight is obtained by weighted summation according to the proportion of window duration. A multidimensional fusion model is constructed, and the normalized deviation between the state index of the three-dimensional sub-model and the corresponding benchmark reference is calculated. The total smoothness index is obtained by fusion through the weighted Euclidean distance normalization formula. The evaluation parameters are evaluated based on the total smoothness index and the intermediate results of the three-dimensional sub-model. The training dataset is input into the three-dimensional sub-model in batches for training. When the training results all meet the preset requirements, the urination smoothness assessment model is obtained.
[0016] Preferably, the segmented urinary flow rate characteristic value is calculated by analyzing the first derivative of the weight-time curve of the segmented urine volume collection module corresponding to a specific distance segment, including: Based on the anatomical structure of the human urination pathway, n consecutive and non-overlapping distance segments are defined, and a graded urine volume collection module is matched for each distance segment, with the effective area of the collection module corresponding one-to-one with the distance segment; the distance segment includes the initial urination segment, the middle urination segment, and the final urination segment; The weight sensor of the collection module corresponding to each distance segment collects data on the change of total weight over time during urination in real time, and generates the original weight-time curve. The original weight-time curve is subjected to outlier removal and smoothing and noise reduction to obtain the preprocessed weight-time curve; The urine weight curve was calculated based on the pretreated weight-time curve.
[0017] ; in, The urine weight-time curve is for the i-th collection module; The weight-time curve of the i-th collection module after preprocessing; Let be the static initial weight of the i-th collection module before urination begins; The density of urine at each distance segment was calculated based on the urine weight curve; ; in, This is the urine dynamic density-time curve corresponding to the i-th distance segment; Standard urine density; The urine weight-time curve for the i-th module; Let i be the effective collection cross-sectional area of the i-th collection module; This is the density-mass ratio coefficient; Let be the length of the i-th distance segment; Used as a reference density threshold; The instantaneous urinary flow rate corresponding to each distance segment is calculated based on the first derivative of the urine weight curve. ; in, This is the instantaneous urinary flow rate-time curve corresponding to the i-th distance segment; Urine weight curve The first derivative; Based on the instantaneous urinary flow rate, the segmented urinary flow rate characteristic value corresponding to each distance segment is calculated.
[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention uses a specific geometric array of holes on a flow guide plate to physically divert and independently collect urine at different distances during a single urination event. Combined with synchronous visual tracking and weight measurement, it can objectively and continuously record the dynamic changes in the urination range and the distribution of urine volume at different distances. Therefore, it can capture key information that traditional urine flow rate measurement cannot obtain, such as the real-time location of the urine flow landing point, the stability of the range, and the distribution ratio of urine at different distances.
[0019] 2. This invention transforms the patient's subjective description of symptoms such as weak urination, short range, or incomplete dribbling into quantifiable objective data. Based on parameters such as the range change curve and urine volume spatial distribution spectrum generated by the system, doctors can accurately determine the strength of bladder contraction, the attenuation of urinary kinetic energy, and whether there is urethral obstruction. It can also conduct a preliminary analysis of the possible location of the obstruction. This not only greatly improves the accuracy and scientific nature of the diagnosis of urinary dysfunction, but also provides unprecedented detailed and multi-dimensional objective evidence for the evaluation of the effects of surgical or drug treatment. Attached Figure Description
[0020] Figure 1 This is a flowchart of the system detection method of the present invention. Detailed Implementation
[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] To address the limitations of existing uroflowmetry techniques, which only provide overall parameters and cannot quantify spatial dynamic information such as voiding distance, spatial distribution of urine volume, and dribbling symptoms, please refer to [link to relevant documentation]. Figure 1 This embodiment provides the following technical solution: A urination patency assessment system based on urination status and parameters includes a urine spatial diversion acquisition device, a multimodal sensing unit, and a central processing and control unit.
[0023] The physical body of the urine spatial diversion collection device is a base with an appropriate tilt angle. The base supports a graded urine volume collection module array. A detachable guide cover is tightly covered directly above the graded urine volume collection module array. The surface of the guide cover has an array of holes, which form an isosceles triangular area with the apex pointing towards the user and the base away from the user.
[0024] Each row of holes in the hole array corresponds to a graded urine collection module below it, ensuring that urine flowing in from a specific row of holes is only collected by the independent graded urine collection module directly below it, thus achieving natural diversion of urine based on the parameter of urination distance.
[0025] Each graded urine collection module includes a sealed housing, a container for receiving urine, a urine inlet tube, a urine outlet tube, and a solenoid valve. The container is located on top of the sealed housing. The bottoms of the holes in the same row are connected to each other through the urine inlet tube. After the urine enters the urine inlet tube, it is collected into the corresponding container in that row. The bottom of the container is connected to the urine outlet tube, and a miniature solenoid valve is installed on the urine outlet tube.
[0026] The multimodal sensing unit includes an optical detection module and a weight sensor. The optical detection module contains two high-frame-rate cameras with anti-condensation capabilities, which are installed 1.5m directly above and 1m to the side of the aperture array area, respectively, ensuring that the camera's field of view completely covers the entire triangular area of the aperture array to comprehensively acquire video information of the urination process. The weight sensor is located inside each graded urine collection module. When urine flows into the container of the graded urine collection module through the aperture of a specific row above, the weight sensor measures the increase in the weight of the container in real time and continuously, and converts the weight change data into an electrical signal output.
[0027] The optical detection module, weight sensor, and miniature solenoid valve are all wirelessly connected to the central processing and control unit.
[0028] At the start of the assessment process, initialization and patient information binding are performed first. When a user needs to be assessed, the operator enters or retrieves the user's unique identification information in the interactive interface of the central processing and control unit, then creates a new assessment session and binds the assessment session ID to the user information.
[0029] Next, a self-test and calibration are performed. The self-test procedure includes checking whether the weight sensors of all graded urine collection modules are zeroed, whether communication is working properly, and whether the camera can start normally and focus on the perforated array area of the flow guide cover.
[0030] After the self-test and calibration are passed, preparation work is carried out. At this time, the operator lays a disposable sanitary isolation pad flat on the surface of the flow guide cover. The sanitary isolation pad is made of high polymer non-woven fabric, which has the characteristics of rapid liquid penetration but not easy splashing. Its size completely covers the triangular hole array. The sanitary isolation pad forms a certain sanitary barrier to prevent splashing from causing the spread of germs and cross-infection. At the same time, because its liquid resistance is extremely small, it will not interfere with the urine flow.
[0031] After preparation, the user is guided to stand in front of the urine spatial diversion collection device. The position of the user's feet is clearly marked to ensure that the body is facing the center line of the triangular array. After the user is in position, the operator issues a start command through the interactive interface. The central processing and control unit then simultaneously starts the data acquisition process of all graded urine collection modules and cameras. All data streams are stamped with high-precision timestamps.
[0032] When the user begins to urinate naturally, the urine stream formed under the influence of gravity impacts the sanitary isolation pad and quickly permeates through the holes under the sanitary isolation pad into the container of the corresponding graded urine collection module.
[0033] During urination, the multimodal sensing unit collects data in real time. When the urine stream hits a row of holes, the camera continuously records a clear image of the dark area formed by the urine-soaked sanitary pad. This area starts near the apex of the triangle and may extend backward or to the side as the urine pressure and direction change. The visual algorithm built into the central processing and control unit dynamically tracks the leading edge of the wetted area through background subtraction and color feature recognition, thereby estimating the instantaneous urination range in real time, that is, the horizontal distance from the point where the urine stream lands to the starting point. On the other hand, urine flows through the holes into the corresponding graded urine collection modules below. As the weight of the container on the module increases, the module's built-in weight sensor simultaneously collects high-frequency mass data and calculates the cumulative urine volume flowing into that specific module in real time by comparing the difference between the mass of the container and its empty mass during self-testing. Each row of holes and its corresponding graded urine collection module represents a specific, preset distance range. For example, the first row represents the closest distance range (0-5cm), the second row represents a slightly farther distance range (5-10cm), and so on, until the last row represents the farthest distance range.
[0034] As urination progresses, the pressure, direction, and continuity of the urine flow change. In the early stages of urination, the bladder pressure is high, and the urine flow reaches a relatively far number of holes. In the middle stages of urination, the bladder pressure is stable, and the urine flow is relatively concentrated. In the late stages of urination, or for obstructive patients, the urine flow is intermittent and weak, only reaching the nearest few holes, or even dripping.
[0035] The determination of the end of urination is based on a composite condition: the readings of all weight sensors do not change significantly for several consecutive seconds, and the wet area in the camera video image no longer expands and urine no longer drips. At this point, data recording is stopped.
[0036] After the user leaves, the operator removes the used disposable sanitary isolation pad, quickly disinfects the surface of the drainage cover, and opens the solenoid valve to allow urine to drain from the container until the weight sensor detects no urine weight and then closes the valve, preparing for the next user's test.
[0037] After data acquisition is complete, the data processing and analysis phase begins. The central processing and control unit first preprocesses and aligns the raw data. Preprocessing includes filtering the weight sensor data to eliminate noise caused by environmental vibrations and enhancing and denoising the camera video stream. Alignment utilizes a unified timestamp to record the urine flow status completely in the form of spatiotemporally correlated data. By using the timestamp, the system correlates the position of the wet zone leading edge displayed on the camera screen at a certain time t, the urine flow status, the changes in weight sensor readings of each collection module from the 1st to the Nth row, and the rate of change of weight of each module over time. This data is then integrated into a multi-dimensional urination event time series table with a unified time axis.
[0038] Based on the voiding event time series, the central processing and control unit extracts and calculates parameters, including the spatial distribution map of urine volume, the voiding range variation curve, and the segmented urine flow rate characteristic value.
[0039] Specifically, the urine volume spatial distribution map statistically analyzes the total urine volume collected by each row of graded urine volume collection modules during the entire urination process, and plots it in the form of a histogram or curve. For an individual with smooth urination, the peak of the urine volume distribution spectrum usually appears in the mid-to-long distance range, and the distribution is relatively concentrated; while for an individual with weak urination or obstruction, the urine volume may be concentrated in the nearest one or two distance ranges. The urination range variation curve was plotted based on visual data. This curve reflects the force fluctuations during the urination process over time, indicating whether there is a gradual decrease in range or intermittent drops and recoveries in range to indicate intermittent urination. The segmented urine flow rate characteristic value represents the urine flow rate at different distance segments. By analyzing the first derivative of the weight-time curve of the graded urine volume collection module at the farthest distance segment, the urine flow rate at the maximum range can be estimated, thereby effectively analyzing the kinetic energy of the urine reaching the farthest point.
[0040] The system also includes a quantitative analysis module to identify symptoms of weak urination or dribbling.
[0041] Specifically, the quantitative analysis module first sets a range threshold and a weight accumulation rate threshold. If, during a period of time that accounts for more than a set proportion of the total urination time, the urine range is continuously lower than the range threshold and the weight accumulation rate of the corresponding distance segment is also continuously lower than the threshold, then it is determined that there are significant symptoms of weak urination.
[0042] The quantitative analysis module sets a time interval. If, within this time interval, only the row or two rows of weight sensors closest to the user show periodic, pulse-like small increases in weight, and the camera captures discrete droplets falling rather than a continuous stream of urine, then significant dribbling symptoms are identified.
[0043] The quantitative analysis module uses a pulse sequence detection algorithm to accurately count the number of dribbles, the total amount of dribbled urine, and the duration of dribbling, thereby converting the patient's subjective description of urinary weakness or dribbling symptoms into objective data.
[0044] Finally, the system generates an assessment report, which includes at least a spatial distribution diagram of urine volume, a curve showing the variation of voiding range, a time-series overlay of the total urine volume curve, segmented urinary flow rate characteristics, total urine volume, total voiding time, average urinary flow rate, maximum voiding range, average range, range fluctuation coefficient, effective voiding volume distribution ratio, percentage of dribbling urine, and duration of dribbling. Based on the assessment report, doctors can evaluate and determine the user's bladder contraction function, urethral patency, location of obstruction, and assess the effectiveness of surgical or drug treatment.
[0045] Working Principle: Based on the collaborative operation of a urine spatial diversion and collection device, a multimodal sensing unit, and a central processing and control unit. The entire process begins with system initialization. The operator binds patient information to the assessment session and performs self-calibration to ensure the weight sensor is zeroed and the camera starts normally. Subsequently, a disposable sanitary isolation pad is laid to prevent cross-infection, and the user stands in the designated position to begin urination. During urination, the multimodal sensing unit collects data in real time. Urine is diverted through the holes in the diversion cover into the graded urine volume collection modules at corresponding distances. The weight sensor continuously measures the cumulative urine volume of each module, while the camera captures the urine flow range and the expansion of the wetted area. All data is timestamped to ensure spatiotemporal correlation. The end of urination is determined by the stabilization of sensor readings and changes in the video image.
[0046] After data acquisition, the central processing and control unit performs preprocessing, including filtering and noise reduction and image enhancement, and aligns the data based on timestamps to generate a time-series table of voiding events. The system extracts key parameters, such as the spatial distribution map of urine volume, the curve of voiding range variation, and segmented urinary flow rate characteristics, to quantify voiding patterns. The quantitative analysis module identifies symptoms of weak or dribbling voiding through threshold comparisons, such as persistently low range or abnormal weight accumulation rate. Finally, the system integrates all parameters to generate an assessment report, including indicators such as total urine volume and range fluctuation coefficient, to assist doctors in assessing bladder function and urethral patency.
[0047] Based on this time series table, assessment parameters including urine volume spatial distribution map, voiding range variation curve, and segmented urine flow rate characteristic values are extracted to generate a voiding patency assessment report, including: The time series table extracts assessment parameters, including the spatial distribution map of urine volume, the change curve of urine range, and the segmented urine flow rate characteristic value, and inputs them into the pre-trained urination patency assessment model to generate a urination patency assessment report. The method for constructing the urination patency assessment model includes: Obtain the model training dataset; the training dataset includes urination video information collected by the optical detection module and urine volume information collected by the weight sensor; perform feature extraction on the training dataset to obtain multi-dimensional original features, the multi-dimensional original features include urine volume spatial distribution map, urination range change curve and segmented urine flow rate feature value; Based on the multi-dimensional original features, a three-dimensional sub-model is constructed, which includes a physical feature dimension sub-model, a dynamic time series dimension sub-model, and a distribution pattern dimension sub-model. The construction process of the physical feature dimension sub-model is as follows: calculate the spatial distribution uniformity factor and the effective volume utilization factor based on the urine volume spatial distribution map, map the two factors to the physical feature state index through the Sigmoid function, and construct an age-gender hierarchical benchmark library to determine the physical feature benchmark reference quantity. The construction process of the dynamic time-series dimension sub-model is as follows: the time-series fluctuation factor, peak concentration factor and time-series decay factor are calculated based on the voiding range change curve and segmented urine flow rate characteristic value. The dynamic time-series state index is obtained by weighted summation and normalization, and the dynamic time-series benchmark reference quantity is determined based on the hierarchical benchmark library. The construction process of the distribution pattern dimension sub-model is as follows: the effective distribution ratio factor and the dripping influence factor are calculated based on the segmented urine flow rate characteristic value and dripping related parameters; the distribution pattern state index is obtained through factor trade-off mapping; and the distribution pattern benchmark reference quantity is determined based on the hierarchical benchmark library. A dynamic weight adaptive mechanism is constructed. Based on the urine flow rate curve, the urination process is divided into three stages: initial, middle and final. Sliding windows are divided according to a preset step size and assigned to the corresponding stage. The kernel density estimation function of the three-dimensional state index is pre-constructed using global training data. The overlapping area of the probability density functions of the state index of the three-dimensional sub-model within each window is calculated. The overlapping area matrix is constructed and the calibration coefficient is obtained by feature decomposition. The weight within the window is calculated by combining the basic weight of the stage and the calibration coefficient. The global dynamic weight is obtained by weighted summation according to the proportion of window duration. A multidimensional fusion model is constructed, and the normalized deviation between the state index of the three-dimensional sub-model and the corresponding benchmark reference is calculated. The total smoothness index is obtained by fusion through the weighted Euclidean distance normalization formula. The evaluation parameters are evaluated based on the total smoothness index and the intermediate results of the three-dimensional sub-model. The training dataset is input into the three-dimensional sub-model in batches for training. When the training results all meet the preset requirements, the urination smoothness assessment model is obtained.
[0048] In this embodiment, the construction process of the physical feature dimension sub-model is as follows: Based on the urine volume spatial distribution map, a spatial distribution uniformity factor and an effective volume utilization factor are calculated; the two factors are mapped to a physical feature state index using the Sigmoid function; and an age-gender stratified benchmark library is constructed to determine the physical feature benchmark reference quantity, including: Spatial distribution uniformity factor ( ): The degree of concentration of urine volume in the urination area. ; in, This refers to the total area of the urination area. Divide the 5×5 grid into a number of sections (to evenly cover the urination area). For the first Grid urine volume, For the first The distance from the grid to the projection point of the urethral opening. Average distance; The smaller the value, the more concentrated the urine output is in a nearby area, and the more reasonable the distribution.
[0049] Effective volume utilization factor ( ): ; in, Effective urine output (total urine volume with a flow rate ≥ 1 mL / s). Total urine output; Space utilization coefficient; The larger the value, the higher the percentage of effective urination.
[0050] Physical characteristic state index (normalized to [0,1], higher index = better accessibility): ; in, =2.5 ( Negative weights, the smaller the better. =2.0 ( Positive weight (the larger the better). =1.6 (calibration coefficient, fitted from a sample of healthy individuals).
[0051] Construct an age-gender stratified benchmark library (e.g., children, youth, elderly; males, females), with benchmark reference values for each stratum ( ) through this stratified healthy population mean ( ) and standard deviation ( )Sure: .
[0052] In this embodiment, the construction process of the dynamic time-series sub-model is as follows: Based on the voiding range variation curve and segmented urinary flow rate characteristic values, the time-series fluctuation factor, peak concentration factor, and time-series decay factor are calculated; a dynamic time-series state index is obtained through weighted summation and normalization; and a dynamic time-series benchmark reference quantity is determined based on a hierarchical benchmark library, including: Time series fluctuation factor ( ): ; in, The number of time-series segments (divided into 100ms segments, covering the total voiding time). Let be the change in range for segment t. Let t be the change in urinary flow rate during segment t. For maximum range, This represents the maximum urinary flow rate. The smaller the value, the smoother the fluctuation.
[0053] Peak concentration factor ( ): ; in, The time to the occurrence of maximum urinary flow rate. The time when the maximum range is reached. Total urination time; The closer it is to 1, the better the synchronization.
[0054] Time decay factor ( ): ; in, The average urinary flow rate during mid-void. The mean urinary flow rate at the end of urination, ε=0.01 (to avoid a denominator of 0); The closer it is to 1, the more gradual the decay.
[0055] Dynamic time-series state index: ; in, (Volatility factor weight) (Peak synchronization factor weight, the most crucial factor). (Attenuation factor weight).
[0056] Layered benchmark calibration (same logic as physical feature dimension): ; As a benchmark reference value for dynamic time series dimensions; State index for dynamic time series dimension The average value of healthy individuals; State index for dynamic time series dimension The standard deviation.
[0057] In this embodiment, the construction process of the distribution pattern dimension sub-model is as follows: Based on segmented urinary flow rate characteristic values and drip-related parameters, the effective distribution ratio factor and drip influencing factor are calculated; the distribution pattern state index is obtained through factor trade-off mapping; and the distribution pattern benchmark reference quantity is determined based on a hierarchical benchmark library, including: Effective distribution ratio factor ( The percentage of urine volume during the peak flow rate period indicates that urine flow is concentrated during the peak period, which equals unobstructed flow. ; in, This is the ratio of urine volume during the peak urinary flow rate period to total urine volume. The standard deviation of segmented urine volume, (calibration coefficient); The larger the value, the more concentrated the effective distribution.
[0058] Drip Influence Factor : ; in, This represents the percentage of dribbling urine. The duration of the drip. The smaller the value, the less impact the dripping has.
[0059] Distribution pattern state index: .
[0060] Based on a hierarchical benchmark library, a baseline reference value for the distribution pattern was determined (drip is a negative indicator, and the baseline is the mean plus standard deviation of the healthy population): ; This serves as a baseline reference for the distribution pattern dimension. State index for distribution pattern dimension The average value of healthy individuals; State index for distribution pattern dimension The standard deviation.
[0061] In this embodiment, a dynamic weight adaptive mechanism is constructed. Based on the urine flow rate curve, the urination process is divided into three stages: initial, middle, and final. Sliding windows are divided according to a preset step size and assigned to the corresponding stage. The area of overlap of the probability density functions of the three-dimensional sub-model state index within each window is calculated. An overlap area matrix is constructed and eigenvalue decomposition is performed to obtain calibration coefficients. The weights within the window are calculated by combining the stage base weights and calibration coefficients. The global dynamic weights are obtained by weighted summation according to the window duration proportion, including: Division of urination stages Initial stage: Urinary flow rate increases from 0 to 0.5. The time period; Mid-term: Urinary flow rate maintained at 0.5 ~ The time period; End stage: Urinary flow rate from 0.5 The period during which the value drops to 0.
[0062] In-window weight calculation Sliding window partitioning: A 500ms sliding window is used to cover the entire urination process, with each window belonging to a corresponding urination stage; Overlap area calculation: A kernel density estimation function for a three-dimensional state index is pre-constructed using global training data, and the overlap area is calculated for each window. ;in, Let be the probability density function of the i-th sub-model within the window; The state index of the sub-model within the window; The overlapping area of the sub-model state indices; Stage weight allocation: Assign basic weights to each stage (initial stage: =0.4, =0.4, =0.2; Mid-term: =0.2, =0.6, =0.2; End of period: =0.2, =0.3, =0.5).
[0063] In-window weight calibration: Based on the eigendecomposition results of the overlapping area matrix, the basic weights during the calibration phase are: Construct the overlapping area matrix : ; Eigenvalue decomposition takes the eigenvector corresponding to the largest eigenvalue. , , After normalization, the calibration coefficients are obtained. , , ); In-window weight: (i=1,2,3; k is the window number); Global weight fusion involves summing the weights of all windows according to their duration to obtain the global dynamic weight. , , ): ; in, The duration of the k-th window ensures a higher weighting for the core phase.
[0064] In this embodiment, the benchmark deviation is calculated. ;in, Let be the normalization bias of the i-th dimension. The smaller the value, the closer the actual condition is to a healthy state, and the better the patency. It is the exponential mean; Let be the state index of the i-th dimension.
[0065] Overall Smoothness Index Calculation: ; It reflects the dynamic importance of each dimension. The closer S is to 1, the better the smoothness. S∈[0,1], which intuitively reflects the degree of smoothness.
[0066] Assessment parameter mapping: Direct extraction: total urine volume, total voiding time, dribbling duration; Sub-model derivation: mean urinary flow rate, maximum voiding range, mean range; Factor transformation: range fluctuation coefficient ( Temporal sub-model Mapping), effective urine output distribution ratio ( Distribution pattern sub-model Mapping), percentage of dribbling urine ( = Distribution pattern sub-model Mapping).
[0067] The working principle and beneficial effects of the above technical solution are as follows: Constructing a sub-model encompassing three dimensions—physical characteristics, dynamic temporal sequence, and distribution pattern—allows for analysis of urination from different perspectives, enabling a more comprehensive and in-depth assessment of urination patency and avoiding the limitations of single-dimensional assessment. By constructing an age-gender stratified benchmark library to determine benchmark reference values for each dimension, physiological differences among different populations are considered, making the assessment results more consistent with individual circumstances and achieving personalized assessment. A dynamic weight adaptive mechanism is constructed, segmenting the urination process according to the urine flow rate curve and calculating the window weights for different stages to obtain global dynamic weights, which better adapts to the dynamic changes in the urination process, improving the accuracy and flexibility of the assessment. A multi-dimensional fusion overall model is constructed, fusing the state indices of the three-dimensional sub-models with the benchmark reference values to obtain a total patency index, and then back-mapping the assessment parameters, achieving effective integration of multi-dimensional information and providing richer and more comprehensive information for clinical diagnosis. Training the training dataset in batches ensures that the model meets the preset requirements on different data subsets, improving the model's stability and reliability and guaranteeing the credibility of the assessment results.
[0068] By analyzing the first derivative of the weight-time curve of the graded urine volume collection module corresponding to a specific distance segment, the segmented urine flow rate characteristic values are calculated, including: Based on the anatomical structure of the human urination pathway, n consecutive and non-overlapping distance segments are defined, and a graded urine volume collection module is matched for each distance segment, with the effective area of the collection module corresponding one-to-one with the distance segment; the distance segment includes the initial urination segment, the middle urination segment, and the final urination segment; The weight sensor of the collection module corresponding to each distance segment collects data on the change of total weight over time during urination in real time, and generates the original weight-time curve. The original weight-time curve is subjected to outlier removal and smoothing and noise reduction to obtain the preprocessed weight-time curve; The urine weight curve was calculated based on the pretreated weight-time curve. ; in, The urine weight-time curve is for the i-th collection module; The weight-time curve of the i-th collection module after preprocessing; Let be the static initial weight of the i-th collection module before urination begins; The density of urine at each distance segment was calculated based on the urine weight curve; ; in, This is the urine dynamic density-time curve corresponding to the i-th distance segment; Standard urine density; The urine weight-time curve for the i-th module; Let i be the effective collection cross-sectional area of the i-th collection module; This is the density-mass ratio coefficient; Let be the length of the i-th distance segment; Used as a reference density threshold; The instantaneous urinary flow rate corresponding to each distance segment is calculated based on the first derivative of the urine weight curve. ; in, This is the instantaneous urinary flow rate-time curve corresponding to the i-th distance segment; Urine weight curve The first derivative; Based on the instantaneous urinary flow rate, the segmented urinary flow rate characteristic value corresponding to each distance segment is calculated. .
[0069] In this embodiment, ,include: ; in, The peak urinary flow rate corresponding to the i-th distance segment; The time when urine is first detected in the i-th distance segment; This is the time when urine collection ends at the i-th distance segment; ; in, The average urinary flow rate corresponding to the i-th distance segment; Instantaneous urinary flow rate Within the valid time interval Integrals within; ; in, The average slope of the increase in urine flow rate corresponding to the i-th distance segment; Let be the peak urinary flow rate for the i-th distance segment; Let be the initial instantaneous urinary flow rate for the i-th distance segment; The peak urinary flow rate time point for the i-th segment; This is the start detection time; ; in, The standard deviation of instantaneous urinary flow rate corresponding to the i-th distance segment; It is the integral of the square of the difference between the instantaneous urinary flow rate and the average urinary flow rate over the effective time interval; ; in, This is the normalized value of the peak urinary flow rate corresponding to the i-th distance segment; This represents the global maximum urinary flow rate.
[0070] The working principle and beneficial effects of the above technical solution are as follows: Based on the anatomical structure of the human urination pathway, distance segments are defined and matched with graded urine volume collection modules. This allows for targeted collection of data at different urination stages, making subsequent analysis more closely aligned with the actual physiological process of urination and improving the effectiveness and relevance of the data; for the original weight— Outlier removal and smoothing / denoising preprocessing of the time curves remove interference factors, improve data quality, and lay the foundation for accurate subsequent calculations. By calculating the urine weight curve, the influence of the initial static weight of the collection module is eliminated, accurately obtaining the change of urine weight over time at different distance segments, allowing subsequent analysis to focus more on the urine itself. The dynamic density of urine at each distance segment is calculated, taking into account that urine density may change with time and location, making the instantaneous urine flow rate calculation more consistent with reality and improving calculation accuracy. The instantaneous urine flow rate is calculated based on the first derivative of the urine weight curve and dynamic density, which can reflect the urine flow situation at different distance segments in real time, providing more detailed information for analyzing the urination process. The segmented urine flow rate characteristic values corresponding to each distance segment are calculated, covering multiple indicators such as peak urine flow rate and average urine flow rate, which can comprehensively evaluate the urine flow situation at different stages of urination from multiple dimensions, providing rich evidence for the assessment of urination function.
[0071] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.
[0072] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.
Claims
1. A urination patency assessment system based on urination status and parameters, characterized in that, It includes a urine spatial diversion and collection device, a multimodal sensing unit, and a central processing and control unit; The urine spatial diversion and collection device includes a base with an inclined angle, a graded urine volume collection module array supported on the base, and a detachable guide cover that tightly covers the top of the array. The surface of the guide cover has multiple rows of holes, which form an isosceles triangle region with the apex pointing towards the user and the base away from the user. Each row of holes corresponds to an independent graded urine volume collection module below it, which is used to realize the natural diversion and collection of urine based on the distance of urination. The multimodal sensing unit includes an optical detection module for acquiring video information of the urination process and multiple weight sensors for real-time measurement of the accumulated urine volume in each graded urine volume collection module. The optical detection module includes high frame rate cameras respectively installed above and to the side of the aperture array area. The central processing and control unit is signal-connected to the multimodal sensing unit to control the initialization of the evaluation session and the synchronous start of data acquisition. It receives and processes the video information and urine volume information, generates a urination event time series table by aligning with timestamps, and extracts evaluation parameters, including the urine volume spatial distribution map, the urination range change curve, and the segmented urine flow rate characteristic value, based on the time series table to generate a urination patency evaluation report.
2. The urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, Each of the graded urine collection modules includes a sealed housing, a container, a urine inlet pipe, a urine outlet pipe, and a solenoid valve. The container is located on top of the sealed housing and is used to collect urine. The bottoms of the holes in the same row are connected to each other through the urine inlet pipe, which is used to collect the urine flowing into the corresponding container from the holes in the same row. The bottom of the container is connected to the urine outlet pipe, and a miniature solenoid valve is installed on each urine outlet pipe to control the urine to be discharged from the container after the assessment.
3. The urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, The central processing and control unit performs the determination of the end of urination, and the determination condition is as follows: The readings of all weight sensors did not change significantly within a continuously set number of seconds; The video footage captured by the optical detection module showed that the urine-wetted area was no longer expanding and no new droplets were falling.
4. The urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, When extracting evaluation parameters, the central processing and control unit specifically performs the following: Based on the urination event timeline, the total amount of urine collected by each row of graded urine collection modules is calculated to form a urine volume spatial distribution diagram. Based on the changes in the position of the leading edge of the urine wetting area identified in the video stream, a curve of the change in urine jet range is plotted. The segmented urine flow rate characteristic value was calculated by analyzing the first derivative of the weight-time curve of the segmented urine volume collection module corresponding to a specific distance segment.
5. The urination patency assessment system based on urination status and parameters according to claim 4, characterized in that, The central processing and control unit is also equipped with a quantitative analysis module for identifying abnormal urination symptoms. The quantitative analysis module is configured as follows: If, during a period of time that exceeds a set proportion of the total urination time, the urine jet range is consistently below a preset range threshold and the weight accumulation rate of the corresponding distance segment is consistently below a preset rate threshold, then it is determined that there is a symptom of weak urination. When a periodic, pulse-like, minute weight step increase is detected only on the nearest row or two rows of graded urine collection modules within a set time interval, and the video shows discrete droplets falling, it is determined that dribbling symptoms exist, and the number of dribbles, total dribbled urine volume, and dribbling duration are counted.
6. The urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, Before evaluation, the system requires a disposable sanitary isolation liner to be laid on the surface of the flow guide cover. The sanitary isolation liner is made of a high-molecular non-woven fabric material that allows for rapid liquid penetration, and its size completely covers the isosceles triangular hole array.
7. The urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, The assessment report includes at least the following parameters: total urine volume, total voiding time, average urine flow rate, maximum voiding range, average range, range fluctuation coefficient, effective voiding volume distribution ratio, percentage of dribbling urine volume, and dribbling duration.
8. The urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, The central processing and control unit also controls the opening of micro solenoid valves to empty the urine from each graded urine collection module after the assessment, in preparation for the next assessment.
9. A urination patency assessment system based on urination status and parameters according to claim 1, characterized in that, Based on this time series table, assessment parameters including urine volume spatial distribution map, voiding range variation curve, and segmented urine flow rate characteristic values are extracted to generate a voiding patency assessment report, including: The time series table extracts assessment parameters, including the spatial distribution map of urine volume, the change curve of urine range, and the segmented urine flow rate characteristic value, and inputs them into the pre-trained urination patency assessment model to generate a urination patency assessment report. The method for constructing the urination patency assessment model includes: Obtain the model training dataset; the training dataset includes urination video information collected by the optical detection module and urine volume information collected by the weight sensor; perform feature extraction on the training dataset to obtain multi-dimensional original features, the multi-dimensional original features include urine volume spatial distribution map, urination range change curve and segmented urine flow rate feature value; Based on the multi-dimensional original features, a three-dimensional sub-model is constructed, which includes a physical feature dimension sub-model, a dynamic time series dimension sub-model, and a distribution pattern dimension sub-model. The construction process of the physical feature dimension sub-model is as follows: calculate the spatial distribution uniformity factor and the effective volume utilization factor based on the urine volume spatial distribution map, map the two factors to the physical feature state index through the Sigmoid function, and construct an age-gender hierarchical benchmark library to determine the physical feature benchmark reference quantity. The construction process of the dynamic time-series dimension sub-model is as follows: the time-series fluctuation factor, peak concentration factor and time-series decay factor are calculated based on the voiding range change curve and segmented urine flow rate characteristic value. The dynamic time-series state index is obtained by weighted summation and normalization, and the dynamic time-series benchmark reference quantity is determined based on the hierarchical benchmark library. The construction process of the distribution pattern dimension sub-model is as follows: the effective distribution ratio factor and the dripping influence factor are calculated based on the segmented urine flow rate characteristic value and dripping related parameters; the distribution pattern state index is obtained through factor trade-off mapping; and the distribution pattern benchmark reference quantity is determined based on the hierarchical benchmark library. A dynamic weight adaptive mechanism is constructed. Based on the urine flow rate curve, the urination process is divided into three stages: initial, middle and final. Sliding windows are divided according to a preset step size and assigned to the corresponding stage. The kernel density estimation function of the three-dimensional state index is pre-constructed using global training data. The overlapping area of the probability density functions of the state index of the three-dimensional sub-model within each window is calculated. The overlapping area matrix is constructed and the calibration coefficient is obtained by feature decomposition. The weight within the window is calculated by combining the basic weight of the stage and the calibration coefficient. The global dynamic weight is obtained by weighted summation according to the proportion of window duration. A multidimensional fusion model is constructed, and the normalized deviation between the state index of the three-dimensional sub-model and the corresponding benchmark reference is calculated. The total smoothness index is obtained by fusion through the weighted Euclidean distance normalization formula. The evaluation parameters are evaluated based on the total smoothness index and the intermediate results of the three-dimensional sub-model. The training dataset is input into the three-dimensional sub-model in batches for training. When the training results all meet the preset requirements, the urination smoothness assessment model is obtained.
10. A urination patency assessment system based on urination status and parameters according to claim 4, characterized in that, By analyzing the first derivative of the weight-time curve of the graded urine volume collection module corresponding to a specific distance segment, the segmented urine flow rate characteristic values are calculated, including: Based on the anatomical structure of the human urination pathway, n consecutive and non-overlapping distance segments are defined, and a graded urine volume collection module is matched to each distance segment, with the effective area of the collection module corresponding one-to-one with the distance segment; the distance segment includes the initial urination segment, the middle urination segment, and the final urination segment; The weight sensor of the collection module corresponding to each distance segment collects data on the change of total weight over time during urination in real time, and generates the original weight-time curve. The original weight-time curve is subjected to outlier removal and smoothing and noise reduction to obtain the preprocessed weight-time curve; The urine weight curve was calculated based on the pretreated weight-time curve. ； in, The urine weight-time curve is for the i-th collection module; The weight-time curve of the i-th collection module after preprocessing; Let be the static initial weight of the i-th collection module before urination begins; The density of urine at each distance segment was calculated based on the urine weight curve; ； in, This is the urine dynamic density-time curve corresponding to the i-th distance segment; Standard urine density; The urine weight-time curve for the i-th module; Let i be the effective collection cross-sectional area of the i-th collection module; This is the density-mass ratio coefficient; Let be the length of the i-th distance segment; Used as a reference density threshold; The instantaneous urinary flow rate corresponding to each distance segment is calculated based on the first derivative of the urine weight curve. ； in, This is the instantaneous urinary flow rate-time curve corresponding to the i-th distance segment; Urine weight curve The first derivative; Based on the instantaneous urinary flow rate, the segmented urinary flow rate characteristic value corresponding to each distance segment is calculated.