Liquid level detection and automatic warning system in medical infusion process

Abstract

The invention discloses a liquid level detection and automatic warning system in a medical infusion process. The system comprises a weight detection module, a backboard module, a temperature sensing module, an image acquisition and recognition module, a data processing and judging module and a communication and notification module, wherein the weight detection module and the backboard module are arranged at an infusion hanging position, and the temperature sensing module, the image acquisition and recognition module, the data processing and judging module and the communication and notification module are matched with the backboard module. The data processing and judging module is configured to judge the liquid level state in the infusion process based on the weight change, the liquid level recognition result or the temperature change characteristic, and select or switch different detection modes according to the detection environment state or the detection condition limitation condition; the communication and notification module is configured to output corresponding warning information to a nurse station when it is detected that infusion is about to be finished or abnormity occurs. The infusion monitoring system can meet infusion monitoring requirements in public infusion areas, night wards and shielding conditions, reliable monitoring of the infusion state is achieved on the premise that rest of patients is not disturbed, and safety and nursing efficiency in the infusion process are improved.

Description

Technical Field

[0001] This invention relates to the field of medical information technology and intelligent monitoring, specifically to a liquid level detection and automatic warning system for medical infusion processes. Background Technology

[0002] With the increasing demand for medical services, intravenous infusion, as one of the most common and widely used methods of drug administration in clinical treatment, is widely used in various medical scenarios such as outpatient infusion halls, inpatient wards, and emergency rooms. In actual medical practice, nurses usually need to manage multiple patients receiving infusions simultaneously, keep abreast of changes in the fluid levels of each infusion container, and replace or remove the needle when the infusion is about to end or when abnormalities occur, in order to prevent air from entering the blood vessels or causing other safety risks.

[0003] Current intravenous infusion management primarily relies on manual rounds and experience-based judgment. Nurses periodically check the infusion stands, visually observing the fluid level in the bottles or bags, and deciding whether to intervene based on changes in the level. While this method can meet basic needs when the number of patients receiving infusions is small and environmental conditions are good, it becomes difficult to monitor the infusion status of each patient promptly and accurately in scenarios such as crowded infusion halls, limited nighttime lighting in wards, or patients using curtains to obscure their surroundings. This carries the risk of omissions or delayed interventions.

[0004] To reduce manual labor, some existing technologies propose installing drip rate detection or simple alarm devices on infusion lines to determine whether the infusion is abnormal by detecting changes in drip rate or the condition of the tubing. However, such devices usually only reflect the local condition of the tubing and cannot directly reflect the actual liquid level in the infusion container. Their accuracy in determining when the infusion is about to end is limited, and they are easily affected by factors such as tubing bends and patient movement.

[0005] In addition, some technical solutions attempt to capture images and identify fluid levels in infusion bottles using cameras. However, these solutions typically rely on visible light imaging, which requires specific ambient lighting conditions. In low-light environments such as hospital wards at night or when bed curtains obscure the view, cameras struggle to obtain clear and stable images, leading to decreased accuracy in fluid level identification and even malfunction. Furthermore, continuous use of visible light for supplemental lighting or frequent shooting can disrupt patient rest and negatively impact the medical experience.

[0006] Furthermore, the applicability of detection methods varies significantly across different infusion scenarios. For example, in public areas such as infusion halls, the location of infusion stands is relatively fixed and the ambient lighting conditions are good, making them suitable for centralized inspection and image recognition. However, in wards, especially at night or under conditions of privacy, infusion containers are often not directly observable, and relying on a single detection method is insufficient to meet the needs for continuous and reliable monitoring.

[0007] Therefore, existing technologies generally suffer from the following problems: First, the infusion level detection method is singular, highly dependent on environmental conditions, and limited in applicable scenarios; Second, there is insufficient support for infusion monitoring under non-ideal conditions such as nighttime or obstruction. Third, the lack of a comprehensive judgment and redundant verification mechanism for multiple detection information leads to a high risk of false alarms or missed alarms. Fourth, the linkage between the infusion status assessment and the nurse station is not flexible enough, making it difficult to provide tiered reminders based on different scenarios and risk levels.

[0008] To address the aforementioned issues, there is an urgent need for an infusion level detection and automatic warning system that can be applied to different infusion scenarios and work reliably under both visible and invisible conditions, in order to improve the safety and nursing efficiency of the infusion process. Summary of the Invention

[0009] The purpose of this invention is to provide a liquid level detection and automatic warning system for medical infusion processes, so as to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution: a liquid level detection and automatic warning system for medical infusion processes, characterized in that it comprises: The infusion attachment module includes a hook for suspending an infusion bottle or infusion bag, and a weighing sensor unit mounted on the hook for acquiring real-time weight information of the infusion container. An information identification and binding module is installed on the infusion pole, including a barcode scanning and identification device, which is used to obtain the drug category, specifications and patient association information when the infusion container is attached, and bind it with the initial weight information obtained by the weighing sensing unit; A liquid level detection module is used to detect the liquid state inside an infusion container. The liquid level detection module includes at least one or more of the following detection methods: a detection method based on weight change, a liquid level detection method based on image recognition, a liquid level detection method based on temperature difference, and a detection method based on thermal conductivity. The pipeline status detection module is installed on the infusion rod and is used to detect the flow status of the infusion pipeline; The data processing and judgment module is used to analyze or fuse the data obtained by the liquid level detection module and the pipeline status detection module, and to judge the infusion status according to the preset threshold or model. The warning output module outputs warning information when it detects that the infusion level is nearing the end or an abnormal state occurs; The communication and notification module is used to send the warning information to the workstation or mobile terminal to achieve remote prompting and handling linkage.

[0011] Preferably, the liquid level detection module includes a weight change detection module. The weight change detection module continuously collects the weight change of the infusion container during the infusion process based on the weighing sensor unit set on the infusion attachment module. Combined with the initial weight information obtained by the information identification and binding module, the module records the change in infusion volume and transmits it to the data processing and judgment module for comparison with the initial weight information.

[0012] Preferably, the liquid level detection module includes an image acquisition and recognition module. The image acquisition and recognition module uses a high-definition camera to acquire video of the infusion container and analyzes the acquired images to identify the liquid level height inside the infusion container. In the inspection state, the image acquisition and recognition module continuously monitors the infusion process and sends the analyzed images or liquid level recognition results to the operation station through the communication and notification module for remote viewing and management by staff.

[0013] Preferably, the infusion attachment module has a back plate module on the back of the hook used to suspend the infusion container. After the infusion container is attached, the back plate module is located on the rear side of the infusion container and forms a relatively fixed positional relationship with the infusion container.

[0014] Preferably, when the image acquisition and recognition module identifies the liquid level of the infusion container, the backplate module is provided with color markings, scale markings, or visual calibration structures. The image acquisition and recognition module uses the markings to perform image analysis on the relative positional relationship between the liquid in the infusion container and the backplate to obtain the liquid level height information in the infusion container.

[0015] Preferably, the backplate module is equipped with a heating unit to form a preset temperature field for the infusion container during the infusion process; the image acquisition and recognition module has an infrared image acquisition function, which acquires infrared images of the infusion container and performs image analysis on the difference in temperature distribution between the liquid area and the non-liquid area to obtain the liquid level height information in the infusion container.

[0016] Preferably, the backplate module is provided with a heating unit and a temperature sensor corresponding to the heating unit. Under preset heating conditions, by detecting the temperature change characteristics or temperature loss rate difference of the backplate module in the area in contact with or close to the infusion container, it is determined whether there is liquid at the corresponding position of the infusion container, thereby obtaining the liquid level status information in the infusion container.

[0017] Preferably, the pipeline status detection module includes a pipeline alarm installed on the infusion rod. The upper end of the infusion pipeline is clamped in the pipeline alarm during use to detect the flow status or abnormal status of the infusion pipeline, and the detection result is used as the trigger condition for the warning output module.

[0018] Preferably, the back panel module is provided with a luminous display structure or a self-emissive display structure to form a recognizable background in low-light or nighttime environments, so that the image acquisition and recognition module can identify the liquid level of the infusion container under different lighting conditions.

[0019] Preferably, the data processing and judgment module is configured to judge the liquid level status and pipeline status during the infusion process, wherein the judgment of the liquid level status is based on at least one or more of the following detection methods: Detection methods based on changes in the weight of infusion containers; Liquid level detection method based on image acquisition and recognition; Infrared image liquid level detection method based on temperature field formed by backplane module; Liquid level detection method based on the temperature dissipation characteristics of the backplane module; The warning output module is configured to output warning information based on the liquid level status or pipeline status; The communication and notification module is configured to send the warning information to a workstation or mobile terminal.

[0020] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention provides a liquid level detection and automatic warning system for medical infusion processes. By setting a weight detection module, a backplate module, and corresponding temperature sensing, image acquisition and recognition modules at the infusion attachment position, and combining them with a data processing and judgment module to fuse and analyze multiple detection results, it achieves continuous monitoring and automatic judgment of the liquid level status, infusion progress, and abnormal conditions during the infusion process. This system can flexibly select or switch detection methods according to different usage environment conditions. In well-lit public infusion areas, image recognition can be used for centralized inspection, while in nighttime or low-light conditions... In the ward, infrared recognition and gentle heating of the back panel create stable recognition conditions. In situations where direct observation is not possible, such as when the patient is covered by a bed curtain, indirect fluid level determination is achieved through weight changes and temperature characteristics. The communication and notification module pushes tiered alarm information to the nurses' station, thereby improving the continuity and reliability of infusion monitoring without disturbing the patient's rest. At the same time, by recording and learning data from the entire infusion process, the judgment parameters and alarm thresholds in subsequent infusion processes are gradually optimized. This solves the problems of existing infusion monitoring methods, such as reliance on manual inspection, limited detection methods, difficulty in accurately determining the infusion status at night or under obstructed conditions, and the tendency for missed or false alarms. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0022] This invention provides a technical solution: a liquid level detection and automatic warning system for medical infusion processes, characterized in that it includes: The infusion attachment module includes a hook for suspending an infusion bottle or bag, and a weighing sensor unit mounted on the hook for acquiring real-time weight information of the infusion container. The information identification and binding module is installed on the infusion pole and includes a barcode scanning device, which is used to obtain the drug category, specifications and patient association information when the infusion container is attached, and bind it with the initial weight information obtained by the weighing sensor unit; The liquid level detection module is used to detect the liquid state inside the infusion container. The liquid level detection module includes at least one or more of the following detection methods: detection method based on weight change, liquid level detection method based on image recognition, liquid level detection method based on temperature difference, and detection method based on thermal conductivity. The pipeline status detection module is installed on the infusion rod and is used to detect the flow status of the infusion pipeline; The data processing and judgment module is used to analyze or fuse the data obtained by the liquid level detection module and the pipeline status detection module, and to judge the infusion status according to the preset threshold or model. The warning output module outputs warning information when it detects that the infusion level is nearing the end or an abnormal state occurs; The communication and notification module is used to send warning information to workstations or mobile terminals to achieve remote prompting and coordinated handling.

[0023] This invention provides a liquid level detection and automatic warning system for medical infusion processes. It continuously monitors the liquid state within the infusion container and the status of the infusion tubing during the infusion process, and automatically outputs warning information when the infusion level approaches completion or when an abnormality occurs. The system uses a conventional infusion pole as its mounting base and integrates multiple detection units at the infusion container attachment point to achieve full-process monitoring of the infusion procedure.

[0024] In this embodiment, the system includes an infusion connection module, an information identification and binding module, a liquid level detection module, a pipeline status detection module, a data processing and judgment module, an alarm output module, and a communication and notification module. The modules work together to detect and alert on the infusion status.

[0025] The infusion attachment module is located at the upper end of the infusion rod and includes a hook structure for suspending the infusion bottle or bag, as well as a weighing sensor unit mounted on the hook. After the infusion container is attached to the hook, the weighing sensor unit can acquire the weight information of the infusion container in real time and continuously collect weight change data during the infusion process, providing basic data support for subsequent infusion level determination.

[0026] The information identification and binding module is installed on the infusion pole, near the attachment point of the infusion container, and includes a barcode scanning device. At the start of the infusion, the barcode scanning device identifies the markings on the infusion container to obtain the drug's type, specifications, and patient information. This information is then bound to the initial weight information collected by the weighing sensor unit, thereby establishing a basic information record corresponding to a single infusion process.

[0027] The liquid level detection module is used to detect the liquid state inside the infusion container. In this embodiment, the liquid level detection module supports at least one or more of the following detection methods: detection based on changes in the weight of the infusion container, liquid level detection based on image acquisition and recognition, liquid level detection based on temperature differences, and liquid level detection based on thermal conductivity. By using these different detection methods individually or in combination, reliable detection of the infusion liquid level can be achieved under different environmental conditions and usage scenarios.

[0028] The tubing status detection module is installed on the infusion pole to monitor the flow status of the infusion tubing. During infusion, if the infusion tubing becomes blocked, clamped, or the fluid flow is abnormal, the tubing status detection module can detect the corresponding status change in a timely manner and transmit the detection results to the data processing and judgment module.

[0029] The data processing and judgment module is used to process and analyze the data obtained by the liquid level detection module and the pipeline status detection module. Based on preset judgment rules or thresholds, it judges the infusion liquid level status and pipeline status to determine whether the infusion is in a normal state, about to end, or abnormal state.

[0030] The warning output module outputs corresponding warning messages when it detects that the infusion level is nearing completion or an abnormal state occurs, to remind medical staff to take timely action. Warning messages can be output in the form of sound, light, or other means.

[0031] The communication and notification module is used to send warning information to workstations or mobile terminals, enabling medical staff to receive infusion status prompts remotely, thereby achieving remote prompting and coordinated handling of the infusion process.

[0032] Specifically, the liquid level detection module includes a weight change detection module. Based on the weighing sensor unit set on the infusion attachment module, the weight change detection module continuously collects the weight change of the infusion container during the infusion process. Combined with the initial weight information obtained by the information identification and binding module, the module records the change in infusion volume and transmits it to the data processing and judgment module for comparison with the initial weight information.

[0033] In this embodiment, the infusion attachment module, serving as the support and foundation detection unit for the infusion container, is located at the upper end of the infusion rod and is used to suspend the infusion bottle or bag. The attachment module adopts a hook structure conventionally used in medical infusion scenarios, allowing the infusion container to hang in a natural downward state, thereby ensuring the stability of the infusion process.

[0034] The hook is equipped with a weighing sensor unit, which can be a strain gauge load cell, pressure sensor, or other weight detection device suitable for medical environments. The sensor can be connected to the hook base via a connecting rod, slide bar, or other means, with a load-bearing component on the outside of the detection end. Changes in pressure outside the load-bearing component are used to output a corresponding signal, thus recording the weight change. After the infusion container is attached, the weighing sensor unit collects the overall weight of the infusion container and uses this weight information as the initial weight data for the current infusion process.

[0035] During the infusion process, as the fluid is continuously delivered, the amount of fluid in the infusion container gradually decreases. The weighing sensor unit continuously collects the weight changes of the infusion container in real time, generating weight change data corresponding to time. The data processing and judgment module receives the weight change data and, based on the difference between the initial weight and the real-time weight, determines the remaining infusion volume or the progress of the infusion.

[0036] By employing a weight-change-based detection method, continuous monitoring of infusion fluid levels can be achieved without additional modifications to the infusion container. This detection method is unaffected by lighting conditions, obstructions, or differences in the appearance of the infusion container. Even when ward curtains are drawn, nighttime lighting is insufficient, or the infusion container is partially obscured, it can still reliably acquire infusion status information, serving as an important basic means of infusion fluid level detection.

[0037] In practical applications, the weight change-based detection method can be used as an independent liquid level detection method, or as an auxiliary or backup detection method for other detection methods. When image detection or temperature detection is limited, the infusion process can still be judged through weight change data, thereby improving the reliability of the system in complex operating environments.

[0038] Specifically, the liquid level detection module includes an image acquisition and recognition module. The image acquisition and recognition module uses a high-definition camera to capture video of the infusion container and analyzes the acquired images to identify the liquid level height inside the infusion container. In the inspection mode, the image acquisition and recognition module continuously monitors the infusion process and sends the analyzed images or liquid level recognition results to the operation station through the communication and notification module for remote viewing and management by staff.

[0039] In this embodiment, the liquid level detection module includes an image acquisition and recognition module for visually detecting the liquid state inside the infusion container. The image acquisition and recognition module can employ a high-definition camera positioned near the infusion pole hook to capture close-range video of the infusion container, ensuring the camera can fully cover the outer contour and internal liquid area of ​​the container without interfering with the infusion operation.

[0040] During the infusion process, a high-definition camera continuously captures video of the infusion container, acquiring image data reflecting the distribution of liquid within the container. The image acquisition and recognition module analyzes and processes the acquired images, determining the current liquid level in the infusion container by identifying the boundary between liquid and non-liquid areas, and continuously tracks changes in the liquid level as the infusion progresses.

[0041] In some implementations, the image acquisition and recognition module can also be used in conjunction with surveillance cameras installed in public areas such as wards, infusion areas, or corridors. When the public area surveillance cameras meet the required field of view coverage, they capture video of the entire infusion area. The image acquisition and recognition module extracts the image area containing the infusion container from the public surveillance footage and performs liquid level identification analysis on that image area, thereby enabling monitoring of the infusion level without adding additional bedside cameras.

[0042] In practical applications, the image acquisition and recognition module can receive video data from cameras near the infusion pole hook or from surveillance cameras in public areas, depending on the camera source. It then continuously or periodically monitors the infusion process according to a preset inspection mode. The analyzed images or fluid level recognition results are sent to the operation station via the communication and notification module, allowing medical staff to remotely view the real-time fluid level of the infusion container on a workstation or mobile terminal, thereby improving the flexibility and coverage of infusion monitoring.

[0043] By combining bedside cameras with public area monitoring, it is possible to achieve image-based monitoring of the infusion process under different ward layouts and equipment configurations, providing a stable source of visual information for subsequent comprehensive judgment in conjunction with other detection methods.

[0044] Specifically, the infusion attachment module has a back plate module on the back of the hook used to suspend the infusion container. After the infusion container is attached, the back plate module is located on the rear side of the infusion container and forms a relatively fixed positional relationship with the infusion container.

[0045] In this embodiment, a backplate module is provided on the back of the hook of the infusion attachment module. After the infusion container is suspended, the backplate module is naturally located behind the infusion container, forming a relatively stable positional relationship with it. The backplate module has a flat plate structure, and its size and shape are adapted to the shape of common infusion bottles or bags, so that the backplate can cover the main liquid distribution area of ​​the infusion container without contacting it or only slightly approaching it. By fixing the backplate module to the back of the hook, changes in the detection benchmark caused by shaking, rotation, or manual adjustment of the infusion container are avoided, thus structurally ensuring the stability and consistency of the subsequent detection process.

[0046] In practical use, the backplate module serves as a fixed reference background behind the infusion container, ensuring that the container remains within a relatively defined spatial range, thus providing a stable spatial reference for level detection. The backplate module does not affect the connection of the infusion tubing or the delivery of the fluid, nor does it interfere with the routine operations of medical personnel. It can stably participate in the monitoring of the infusion process over a long period while maintaining consistent infusion operating habits. Through this structural arrangement, the backplate module can serve as a background reference for visual inspection as well as a mounting and operational platform for other detection methods, providing a unified structural foundation for the implementation of various subsequent level detection methods.

[0047] Specifically, when the image acquisition and recognition module identifies the liquid level of the infusion container, the backplate module is equipped with color markings, scale markings, or visual calibration structures. The image acquisition and recognition module uses the markings to perform image analysis on the relative positional relationship between the liquid in the infusion container and the backplate in order to obtain the liquid level height information in the infusion container.

[0048] In this embodiment, the backplate module is used in conjunction with the image acquisition and recognition module to identify the liquid level inside the infusion container. The backplate surface is provided with a marking structure for visual recognition. This marking structure may include color blocks, linear scales, pattern markings, or combinations thereof. The markings are arranged on the backplate surface in a predetermined manner, forming a stable, clear, and easily distinguishable background feature in the camera's captured image. When the infusion container is suspended on the hook, the liquid area inside the container and the backplate markings are presented within the same field of view, with a clear difference between the liquid and non-liquid areas in terms of light transmittance, reflectivity, or color overlay effect.

[0049] During image acquisition, the camera captures an image containing the infusion container and its backplate markings. The image acquisition and recognition module locates the backplate marking area in the image and uses it as a spatial reference to compare and analyze the position of the liquid boundary within the infusion container. By determining the positional relationship between the liquid boundary and the backplate markings—for example, whether the liquid boundary is above or below a certain scale line, or the proportion of a color block it covers—the module identifies the infusion level. As the infusion process continues, the corresponding position of the liquid boundary on the backplate markings continuously changes. The image acquisition and recognition module continuously tracks this change, thus reflecting the dynamic changes in the infusion level.

[0050] By setting clear visual markings on the backplate surface, the liquid level recognition process is independent of differences in the scale, material, or appearance of the infusion container itself, avoiding the instability caused by different sizes or transparency of infusion bottles or bags. The backplate markings provide a unified and fixed reference system for image recognition, ensuring consistency and reliability of liquid level detection results across different infusion containers, shooting angles, and usage environments. In one specific configuration, the backplate module features vertically spaced linear graduations on its surface. These graduations are colored to contrast sharply with the background, ensuring clear edge details in the camera's view. With the infusion container suspended, the liquid area within the container is positioned in front of the graduations. The liquid surface boundary, formed by the liquid-air interface, is within the same field of view as the backplate graduations in the image. After acquiring the image, the image acquisition and recognition module first identifies the distribution of the backplate graduations in the image. Using these graduations as a reference coordinate system, it compares and determines the location of the liquid surface boundary. By identifying the corresponding graduation interval or position, it identifies the liquid level within the infusion container. As the infusion continues, the corresponding position of the liquid surface boundary on the graduations gradually decreases. The image acquisition and recognition module continuously records this positional change, reflecting the trend of the infusion level.

[0051] In one embodiment, the surface of the backplate module is configured with a regularly arranged grid-like marking structure. The grids maintain a fixed spacing in both the horizontal and vertical directions, creating a stable and uniform visual reference area on the monitoring screen. Several monitoring cameras are pre-positioned at fixed locations within the infusion area. The installation position of each camera corresponds to the arrangement position of the infusion pole, ensuring that the infusion hooks are located in a relatively fixed spatial area on the monitoring screen. During system initialization, the image acquisition and recognition module analyzes the monitoring screen, identifies and records the corresponding position area of ​​each infusion hook in the screen, and associates and stores this position area with the specific infusion position number, thereby establishing a correspondence between the infusion position and the monitoring screen.

[0052] In actual operation, the system sequentially inspects multiple infusion stations according to the daily infusion schedule or the order of personnel currently receiving infusions. The image acquisition and recognition module polls and switches between image areas corresponding to each infusion station at preset inspection time intervals. When switching to a specific infusion station, it acquires the current image of that station and extracts the image area containing the infusion container and the grid markings on the backplate. By analyzing this image area, it identifies the coverage area of ​​the liquid within the infusion container within the grid markings and compares the current coverage area with the coverage area recorded during the previous inspection, calculating the displacement change of the liquid boundary per unit time.

[0053] Based on the liquid level change data obtained during multiple inspections, the system can generate a liquid level change trend corresponding to time and evaluate the infusion process accordingly. When the liquid level change rate of a certain infusion level slows down significantly, accelerates abnormally, or approaches the preset end position, the system can mark the infusion level as entering a key monitoring state, increase its priority in the inspection queue, or shorten its inspection interval to obtain the latest status of the infusion level more frequently.

[0054] During inspections, if a valid image cannot be obtained for an infusion station due to personnel movement, obstructions, or temporary image blur, the system can temporarily skip that station and re-acquire the corresponding image in the next round of inspections. This avoids a single anomaly interfering with the overall inspection process. For newly started or added infusion stations, the system can automatically add the station to the inspection sequence after detecting a new infusion container being attached, and establish a corresponding liquid level change record from the initial inspection time, thereby ensuring the continuity and integrity of inspection management.

[0055] By combining the grid markings on the back panel with fixed monitoring cameras, the system achieves centralized inspection and management of multiple infusion sites. This allows the system to continuously track changes in the liquid level of multiple infusion containers without relying on independent bedside cameras, and to judge and warn of the infusion process in conjunction with the time dimension.

[0056] In one embodiment for nighttime use, the backplate module surface is provided with luminescent or fluorescent markings. The luminescent markings absorb ambient light energy during the day or when the lights are on, and continue to emit light when the light level decreases or is turned off, ensuring the backplate remains a clearly identifiable visual background in low-light environments. The luminescent markings can be distributed on the backplate surface in the form of linear scales, grid markings, or combinations thereof, allowing the infusion container to maintain a clear outline contrast with the backplate in nighttime environments.

[0057] At night or when the ward lights are off, the fixed-position monitoring cameras continue to poll and collect data on each infusion level according to the daytime inspection logic. Because the luminous backplate still emits light even without external illumination, the monitoring image clearly shows the marked areas and their relative positions. The image acquisition and recognition module determines the fluid level by identifying the coverage relationship between the luminous markings and the liquid area inside the infusion container. Compared to daytime inspections, nighttime inspections do not require additional lighting, avoiding the disruption of patients' rest caused by strong light.

[0058] During inspections, when the overall brightness of the monitoring screen is low or local brightness fluctuates, the image acquisition and recognition module can prioritize location analysis based on the luminous marker area, rather than relying on the appearance features of the infusion container itself, thereby improving the stability of nighttime liquid level recognition. For situations where the luminous effect gradually weakens or the luminous brightness is insufficient, the system can continuously judge the liquid level change trend by combining the results of the previous inspection, avoiding misjudgments caused by a single image quality degradation.

[0059] By setting up a luminous backplate, the image acquisition and recognition module can still centrally monitor multiple infusion levels according to the established inspection process at night, in low light, or in areas with insufficient local lighting. This achieves a consistent connection between daytime and nighttime infusion level detection methods, improving the continuity and reliability of infusion monitoring in all-weather usage scenarios.

[0060] Specifically, the backplate module is equipped with a heating unit to create a preset temperature field for the infusion container during the infusion process; the image acquisition and recognition module has an infrared image acquisition function, which acquires infrared images of the infusion container and performs image analysis on the differences in temperature distribution between the liquid area and the non-liquid area to obtain the liquid level height information in the infusion container.

[0061] In this embodiment, the backplate module is equipped with a heating unit, which is located inside the backplate or near its surface. This heating unit is used to create a relatively stable preset temperature environment for the infusion container during infusion. During operation, the heating unit slowly and evenly heats the backplate, ensuring that the side of the infusion container closest to the backplate is within a controllable temperature field, thus preventing interference from ambient temperature fluctuations on the test results. The heating temperature is controlled within a safe range suitable for medical settings and will not affect the infusion medication itself.

[0062] In one specific implementation, when the system is in a low-light environment, when visible light image recognition is unstable, or when the transparency of the infusion container is low, the heating unit of the backplate module is activated to create a stable preset temperature field on the backplate. The heating unit operates in a low-power, uniform heating manner, keeping the surface temperature of the backplate relatively constant for a period of time, thereby creating a stable thermal environment background behind the infusion container. After the infusion container is suspended in front of the backplate, its internal liquid and non-liquid regions exhibit different temperature response characteristics under the influence of this temperature field due to differences in heat capacity and thermal conductivity.

[0063] During the heating process, the image acquisition and recognition module performs infrared imaging on the infusion container using infrared image acquisition. The infrared imaging range covers the corresponding area between the infusion container and the backplate. Because the liquid area has a larger heat capacity and a slower rate of temperature change, while the non-liquid area responds to the heating of the backplate more quickly, there is a significant difference in temperature distribution between the liquid and non-liquid areas in the infrared image, manifested as a temperature gradient or the presence of a temperature boundary.

[0064] The image acquisition and recognition module identifies areas within the infusion container where the temperature changes slowly based on the temperature distribution in different regions of the infrared image, classifying these areas as liquid regions. Simultaneously, areas where the temperature changes rapidly or is close to the temperature of the backplate are classified as non-liquid regions. By analyzing the vertical distribution of liquid and non-liquid regions, the boundary between them is determined, thereby obtaining the liquid level information within the infusion container.

[0065] During infusion, as the liquid volume gradually decreases, the temperature response area corresponding to the liquid region gradually shrinks, and the temperature boundary between the liquid and non-liquid regions in the infrared image changes over time. By continuously tracking this boundary position, the system can stably acquire the liquid level change trend and independently determine the infusion liquid level status even when visible light image recognition is limited or ambient light is insufficient.

[0066] In the above embodiment, the heating unit of the backplate module uses a low-power heating method during operation, resulting in a slight temperature rise on the backplate surface. This temperature rise is only sufficient to meet the infrared recognition requirements for temperature difference detection and will not cause discomfort to the infusion container or the patient. By controlling the heating power and operating time, the side of the infusion container near the backplate is kept in a mild and stable temperature environment, which helps to reduce the impact of ambient temperature fluctuations on the infusion process.

[0067] In practical use, low-amplitude heating can alleviate the cold stimulation caused by low ambient temperature of the infusion fluid to a certain extent, allowing patients to maintain a more comfortable experience during infusion. At the same time, this heating method exists only as an auxiliary detection condition and is not intended to change the temperature of the infusion fluid, thus avoiding any impact on drug stability and infusion safety. Therefore, while meeting the needs of fluid level identification, it also takes into account the comfort and safety of patients during use.

[0068] Specifically, the backplate module is equipped with a heating unit and a temperature sensor corresponding to the heating unit. Under preset heating conditions, by detecting the temperature change characteristics or temperature loss rate difference of the backplate module in the area in contact with or close to the infusion container, it can determine whether there is liquid at the corresponding position of the infusion container, thereby obtaining the liquid level status information in the infusion container.

[0069] In this embodiment, temperature sensors are distributed along the height of the backplate module, covering the height range where liquid changes may occur within the infusion container during infusion. Preferably, the temperature sensors are arranged at equal intervals along the vertical direction of the backplate, for example, multiple temperature sensors are arranged at intervals of 20mm to 40mm along the height of the backplate, so that each sensor corresponds to a height detection range within the infusion container, thereby enabling the acquisition of temperature change characteristics at different height positions within the infusion container. This distribution method allows for vertical determination of the liquid distribution within the infusion container without relying on optical recognition.

[0070] The temperature sensor is a high-sensitivity sensor suitable for detecting low temperature differences. Its operating temperature range covers the temperature variation range of common ward ambient temperatures and low-power heating conditions, such as 15℃ to 35℃. Under the preset heating conditions formed by the backplate heating unit, the surface temperature of the backplate only changes slightly relative to the ambient temperature. This change is usually controlled within the range of 2℃ to 6℃ to ensure that there is no impact on the safety of the infusion container and the patient, while still forming a temperature difference that can be detected by the temperature sensor.

[0071] Regarding recognition accuracy, the temperature sensor's measurement accuracy is preferably within the range of ±0.1℃ to ±0.3℃, enabling it to distinguish subtle temperature changes between liquid and non-liquid areas during heating or cooling. At the corresponding height position where liquid is present, the temperature change curve exhibits a slow rise and stable fluctuation; while at the height position where no liquid is present, the temperature change curve more closely resembles the heating or cooling characteristics of the backplate itself, with a significantly increased rate of change. The data processing and judgment module improves the stability and anti-interference capability of liquid level recognition by comparing and analyzing the temperature change rates collected by each temperature sensor, rather than relying solely on the absolute temperature value at a single moment.

[0072] In practical applications, when individual temperature sensors deviate from the normal range due to installation errors, local contact differences, or environmental airflow, the system can cross-compare the detection results from temperature sensors at adjacent heights to correct or eliminate abnormal data, preventing single-point anomalies from affecting the overall liquid level judgment. Through the aforementioned sensor distribution method, temperature setting strategy, and recognition accuracy control, the liquid level recognition scheme based on backplate heating and temperature detection possesses good feasibility and engineering reliability. Specifically, the pipeline status detection module includes a pipeline alarm installed on the infusion rod. The upper end of the infusion pipeline is clamped in the pipeline alarm during use to detect the flow status or abnormal status of the infusion pipeline, and the detection result is used as the trigger condition for the warning output module.

[0073] In this embodiment, the pipeline status detection module includes a pipeline alarm mounted on the infusion rod. The pipeline alarm is positioned at the upper end of the infusion pipeline, allowing the pipeline to naturally pass through or be clamped within the detection channel of the pipeline alarm during infusion. The pipeline alarm utilizes an existing, mature infusion pipeline detection structure, monitoring the flow status of the liquid within the infusion pipeline by partially clamping or closely monitoring it during use.

[0074] Under normal infusion conditions, the fluid in the infusion tubing flows continuously. The tubing alarm can detect physical changes corresponding to this fluid flow, such as periodic deformation of the tubing, pressure changes, or changes in state caused by flow, and identify these changes as normal flow conditions. When an abnormality occurs in the infusion tubing, such as tubing kinks, blockages, clamps not being opened, the fluid almost running out, or the infusion ending, the fluid flow state in the infusion tubing changes significantly. The tubing alarm detects the corresponding abnormal characteristic signals and identifies this as an abnormal state.

[0075] The pipeline alarm sends detected signals of pipeline flow status or abnormal status to the data processing and judgment module, serving as one of the trigger conditions for the warning output module. When the pipeline alarm detects an abnormal status persisting for more than a preset time threshold, the system determines that there is an abnormal risk in the infusion pipeline and outputs an appropriate level of warning information through the warning output module to remind medical staff to handle the situation promptly. By cooperating with the aforementioned fluid level detection module, the system can not only determine changes in the infusion fluid level but also promptly identify potential pipeline abnormalities during the infusion process, thereby improving the completeness and safety of infusion process monitoring.

[0076] In a preferred embodiment, the pipeline alarm uses a detection method based on pipeline pressure changes to monitor pipeline status. The alarm forms a repeatable, lightweight clamping structure on the upper section of the infusion pipeline, and a micro-pressure / strain detection unit is arranged in the clamping area to acquire the pressure change characteristics generated during the liquid flow process. During normal infusion, the liquid in the infusion pipeline is in a state of continuous dripping or continuous flow. The pipeline pressure signal corresponding to the clamping area remains within a stable range and exhibits small periodic fluctuations. For example, the equivalent pipeline pressure signal is stable in the range of 2–6 kPa for a long time, and the fluctuation amplitude is less than 0.5 kPa.

[0077] When the infusion tubing experiences kinking, clamping, or partial blockage, fluid flow is obstructed, causing a significant change in the tubing pressure signal in the clamped area, which continuously deviates from the normal range. For example, the equivalent tubing pressure signal may rise to 8–15 kPa and remain there for more than 5–10 seconds. When the infusion is about to end or cavitation occurs, the flow within the tubing changes from liquid to a gas-liquid mixture or intermittent flow, resulting in significant irregular fluctuations or a rapid decrease in the tubing pressure signal. For example, the equivalent tubing pressure signal may drop from its original stable range to below 1 kPa, accompanied by increased fluctuation amplitude. The data processing and judgment module compares and judges the absolute value range, rate of change, and duration of the tubing pressure signal. A sustained high-pressure state is identified as an abnormal blockage / clamping, while a low-pressure state with abnormal fluctuations is identified as a risk of cavitation or infusion termination. The judgment result serves as the trigger condition for the warning output module, which outputs the corresponding level of warning information.

[0078] The tubing alarm can also employ an optical detection structure, mounted on the infusion rod and forming a detachable clamp with the upper end of the infusion tubing. The alarm internally houses a light-emitting unit and a light-receiving unit, arranged opposite each other to form a transmissive light path, ensuring the infusion tubing is centered within the light path after clamping. The light-emitting unit preferably uses a near-infrared LED with a center wavelength of 940nm (850nm is also an option) to reduce visible light exposure to common medications. The light-emitting unit is driven by a low-power pulse method, such as a peak current of 5–15mA, a duty cycle of 1–5%, and a sampling frequency of 50–200Hz. A light-shielding cavity is formed within the alarm housing to prevent light leakage or prolonged continuous irradiation of the infusion tubing.

[0079] Under normal infusion conditions, the infusion tubing is filled with liquid, and the transmission or reflection intensity of near-infrared light on the tubing cross-section remains relatively stable, resulting in minimal fluctuations in the output signal of the light-receiving unit. When tubing clamping, bending, or blockage causes abnormal liquid flow, the flow pattern within the infusion tubing changes, and the light-receiving signal exhibits a continuous deviation. When cavitation, continuous passage of air bubbles, or near-cavitation occurs, the gas-liquid interface causes significant changes in transmission / scattering characteristics, resulting in increased pulsation or transient jumps in the light-receiving signal. The data processing and judgment module performs threshold and duration judgments on the light-receiving signal. For example, a change in the light-receiving signal relative to the baseline exceeding 15–30% and lasting for 3–10 seconds is considered a clamping / blockage abnormality; a high-amplitude pulsation in the light-receiving signal that repeats a preset number of times within 30–120 seconds is considered a risk of cavitation or infusion termination, triggering an alarm output accordingly.

[0080] To better suit clinical use, the optical tubing alarm is preferably designed with a quick-release snap-on structure that allows for "plug and play" operation. The position of the infusion tubing is defined within the clamping channel, ensuring that the relative position of the optical path and the infusion tubing is consistently accurate. Simultaneously, a baseline signal (e.g., 2–5 seconds) is acquired during the first clamping as a reference baseline for that infusion, to accommodate differences in transparency between different brands of infusion tubing and ambient light interference.

[0081] Specifically, the back panel module is equipped with a luminous display structure or a self-illuminating display structure to form a recognizable background in low-light or nighttime environments, so that the image acquisition and recognition module can identify the liquid level of the infusion container under different lighting conditions.

[0082] In this embodiment, the luminescent or self-emissive display structure of the backplane module provides a stable visual background for liquid level identification in low-light or nighttime environments. The luminescent display structure is preferably made of a long-afterglow luminescent material, such as a phosphorescent luminescent material based on a strontium aluminate system. After absorbing light energy under ambient light or ward lighting conditions, it can continuously emit light for several hours without external light. The emission wavelength is concentrated in the visible light range easily identifiable by the human eye and cameras, such as the green or blue-green band. This type of material does not require an external power source and is suitable for long-term deployment in ward environments.

[0083] Luminescent materials can be applied to the surface of the backplate by coating, attaching, or embedding to form structures such as linear scales, grid markings, or outlines, creating a uniform and stable luminescent background at night. When an infusion container is suspended in front of the backplate, the liquid area inside the container, due to its absorption, refraction, and blocking of light, creates a clear contrast in brightness and shape with the luminescent area of ​​the backplate, resulting in a distinct boundary between the liquid and non-liquid areas in the image.

[0084] During nighttime inspections, the image acquisition and recognition module uses imaging parameters adapted to low-light environments to acquire images of the backplate and infusion container. For example, it increases the photosensitivity or extends the exposure time to maintain a stable brightness of the luminous area of ​​the backplate in the image. By identifying the spatial distribution of the luminous markings, using the backplate as a background reference, and analyzing locations of abrupt changes in brightness or contour within the infusion container, the module determines the liquid level within the container.

[0085] In another embodiment, the back panel module can use a self-emissive display structure instead of photoluminescent materials. For example, a low-power light-emitting unit can be installed inside the back panel. The light-emitting unit preferably uses visible light LEDs or electroluminescent sheets, and the brightness is controlled at the minimum level required for image recognition, for example, only turning on at night or under low light conditions. To avoid affecting the infusion solution, a certain distance is maintained between the light-emitting unit and the infusion container, and the light path is only used to illuminate the marking area on the back panel, without directly illuminating the liquid inside the infusion container.

[0086] In a self-emissive display structure, a brightness detection sensor can be installed within the backplane module to sense ambient light intensity. When the ambient light is detected to be below a preset threshold, the light-emitting unit is automatically turned on; when the ambient light recovers, the light-emitting unit is turned off or the brightness is reduced to reduce energy consumption and extend the device's lifespan. Through this method, the backplane can always provide a recognizable visual background under different lighting conditions, enabling the image acquisition and recognition module to stably complete infusion level recognition even at night, when the blackout curtain is drawn, or when local lighting is insufficient.

[0087] Specifically, the data processing and judgment module is configured to judge the liquid level status and pipeline status during the infusion process, and the liquid level status judgment is based on at least one or more of the following detection methods: Detection methods based on changes in the weight of infusion containers; Liquid level detection method based on image acquisition and recognition; Infrared image liquid level detection method based on temperature field formed by backplane module; Liquid level detection method based on the temperature dissipation characteristics of the backplane module; The warning output module is configured to output warning information based on the liquid level or pipeline status; The communication and notification module is configured to send alert messages to workstations or mobile terminals.

[0088] Based on the above solutions, the following are specific implementation scenarios and examples: Example 1 Seating area in public infusion hall 1. Procedures performed by the nurse before starting the IV infusion and system initialization 1) The nurse takes the patient to the designated infusion position in the seating area of ​​the infusion hall. This infusion position corresponds to a fixed infusion stand (or fixed hook position). The hook position is pre-numbered in the system, for example, "seat area - position K".

[0089] 2) The nurse hangs the IV bottle / IV bag on the hook. The back panel of the IV container has been fixedly installed (the back panel has square / gradient markings, and the markings are clearly visible to the public surveillance camera).

[0090] 3) The nurse scans the label on the infusion container (which may contain the drug batch / specification / prescription number / patient ID or wristband code) using the barcode scanner on the infusion pole, and the system creates a record for this infusion task: Infusion level number K Patient identification (de-sensitized ID) Drug type / specification Start time T0 Initial state: Pending 4) The nurse opens the infusion tubing and completes the routine air venting and drip rate adjustment. The system enters the "monitoring start" state and records the first valid image as the reference frame (the reference frame includes: hook position, back panel markings, infusion container outline, and initial position of the liquid surface boundary).

[0091] 2. Public surveillance inspection 5) The lobby's public surveillance cameras are either pan-tilt-zoom (PTZ) or fixed multi-camera coverage; the system pre-calibrates the "target area parameters" for each hook position in the monitoring screen, including: Gimbal angle (horizontal angle / tilt angle / zoom angle) or ROI coordinate frame of fixed lens Reference positions of the back panel marking area (grid corners / start and end points of scale lines) 6) The system generates an inspection queue based on the "In-Service Infusion List" for the day: K1, K2, ..., Kn (arranged according to start time or nurse entry order), and sets the inspection cycle, for example, collecting data once every 5-10 seconds for each infusion station, and cycling through the entire inspection every 1-3 minutes (the cycle can be extended when there are many people in the lobby and shortened when there are few people).

[0092] 3. Image acquisition, recognition, and liquid level calculation during inspections. 7) When the system switches to the infusion position Ki: the pan-tilt unit rotates to the preset angle and stabilizes for 0.3–1 seconds (anti-shake), and captures 1–3 images (or takes 1 second of video) as the inspection data for this inspection.

[0093] 8) Identification process (the key is "how to use the back panel label"): First, identify the grid / scale marking area on the back panel (locate the grid border or scale lines) to obtain the coordinate system of the back panel in the image; Re-identify the outline of the infusion container (bottle / bag shape) and the internal liquid area; Identify liquid / non-liquid boundaries (edges caused by differences in brightness, refractive differences, and contour changes); Map the liquid surface boundary to the backplate coordinate system to obtain "liquid level height H(t)" or "remaining proportion P(t)" (e.g., the liquid surface falls between the m-th and m+1-th scales, or covers the number of grid rows r).

[0094] 9) To be more realistic, the system does not draw conclusions directly from a single frame result, but instead uses a "short window stability judgment": the liquid level value is confirmed to be valid only if the results of 2-3 consecutive inspections of the same infusion level are similar (with an error within 1-2 divisions or 1 scale line); if the difference is too large, it is marked as "image unstable".

[0095] 4. Remaining Time Estimation and Tiered Early Warning Triggering 10) The system maintains the time series of each infusion level: {T0,H0}, {T1,H1}, ..., and calculates the rate of drop in liquid level per unit time V=ΔH / Δt (or ΔP / Δt).

[0096] 11) Based on the rate and current liquid level, estimate the remaining time TR (e.g., TR≈(H(t)-H_end) / V), and set multiple threshold levels: Warning 1 (Reminder): 10–15 minutes remaining. Warning 2 (Ending Soon): 3-5 minutes remaining. Warning 3 (High Priority): Liquid level nearing the end or continuous occurrence of "vacuuming characteristics / pipeline abnormalities". 12) When the warning 2 / 3 is triggered, the system will automatically raise the priority of the infusion level: put the level at the front of the inspection queue and shorten the repeated collection interval of the level (for example, from once per round to twice per round).

[0097] 5. Auxiliary judgment of pipeline abnormalities 13) If pipeline status detection (such as a clamp-on pipeline alarm) is enabled for this infusion level, then the pipeline status is used as an "abnormality weighting factor": If the liquid level drops normally but the pipeline alarm indicates clamping / no dripping: report "pipeline abnormality" first; If the liquid level in the image is unstable but the pipeline shows continuous dripping: delay the alarm and continue inspection; If the image display is nearing its end and cavitation / fluctuation occurs in the pipeline: a high-priority alarm is triggered directly.

[0098] 6. Notify the nursing station and have the nurses arrive to complete the closed-loop process. 14) The system pushes alarms to the nurse station workstation and nurse mobile terminals through the communication and notification module. The push content should include at least: Infusion station number (seat area - Kth position) Patient desensitization information / seat number Current liquid level status (estimated remaining time, nearing end level) Triggering reasons (image liquid level threshold, pipeline abnormality, image instability requiring manual confirmation, etc.) The most recent snapshot (used by nurses for quick confirmation, avoiding wasted trips). 15) After receiving the order, the nurse station can "confirm the order with one click" on the system (record the order time). The system will change the alarm status from "pending" to "processing" and continuously track the infusion level.

[0099] 7. Termination of handling and monitoring after the nurse's arrival 16) After the nurse arrives, she will perform routine procedures: confirm the patient, check the drip rate, change the fluid or prepare to end the infusion if necessary; when it is confirmed that the infusion has ended and the needle needs to be removed, the nurse will clamp the tubing, remove the needle and apply a hemostatic dressing.

[0100] 17) The system's "termination conditions" adopt a feasible combination of judgments (in the lobby environment, it is recommended that at least one condition be met for termination to avoid accidental termination): The nurse clicks "Needle removed / Completed" on the mobile device (manual confirmation to terminate); or The system detected that the infusion container was removed (the container outline disappeared from the hook position screen, and the container was not detected in N consecutive inspections, where N can be set to 2-3 times); or The pipeline alarm continuously displays "No flow / Closed" and the image recognition liquid level no longer changes (for M minutes).

[0101] 18) After termination, the system execution log is archived: End time Tend Final liquid level / final image evidence Did any abnormalities occur (clamping, occlusion, repeated instability, premature termination, etc.)? Nurses' on-site response time (from alarm to order acceptance, from order acceptance to arrival) Then remove the infusion level from the inspection queue, freeing up monitoring resources for the next inspection.

[0102] Example 2 Intravenous infusion in ward I. Nurses begin intravenous infusions and system record creation at night. 1) During night rounds or the execution of nighttime medical orders, nurses enter the ward and administer intravenous infusions to patients without turning on the main lighting. The nurse hangs the infusion bottle or bag on the fixed hook of the infusion stand next to the bed. A backplate module is installed on the back of the hook, which integrates a low-power heating unit and an infrared recognition adapter structure.

[0103] 2) After the nurse completes the air removal and drip rate adjustment according to standard procedures, she uses the barcode scanner on the infusion pole to scan the infusion container label or patient wristband information. The system records and establishes the basic information for this infusion task, including: Ward number and bed number Patient identification information (de-identified ID) Drug types and specifications Infusion start time T0 Initial infusion weight W0 3) When the system detects that the current ambient light intensity is lower than the preset threshold (e.g., the bedside lamp is off or the brightness is lower than the set value), it automatically marks the infusion task as "night infusion mode", disables the visible light image recognition function, and only keeps the infrared image acquisition, weight detection and temperature-related detection modules in working state.

[0104] II. Backplate heating start-up and preparation for nighttime infrared monitoring 4) After entering the nighttime infusion mode, the system controls the heating unit of the backplate module to start. The heating unit slowly heats the backplate at low power, so that the surface temperature of the backplate has a small difference relative to the ambient temperature of the ward. For example, when the ambient temperature is about 20℃~22℃, the backplate temperature is stably controlled within the range of 22℃~26℃.

[0105] 5) This heating process is only used to create a stable temperature background in infrared imaging and is not intended to heat the infusion fluid. The heating power and upper temperature limit are both limited to a safe range that is imperceptible to the patient and does not affect the stability of the medication.

[0106] 6) After the backplate temperature stabilizes, the infrared image acquisition module starts to continuously acquire infrared images. The acquisition area covers the main body of the infusion container and its corresponding backplate area. The acquisition cycle is, for example, once every 5 to 10 seconds, which is used to form the basic data sequence for nighttime liquid level monitoring.

[0107] III. Specific Determination Process of Infrared Liquid Level Recognition 7) In each infrared image acquisition, the system first identifies the back panel area and uses the stable temperature distribution formed by heating the back panel as a reference background to segment the infrared image into regions.

[0108] 8) Subsequently, the system analyzes the temperature distribution in the corresponding area in front of the infusion container's backplate. Due to the large heat capacity and slow temperature change characteristics of the liquid inside the infusion container, under heating conditions, the liquid area appears as a region with slow temperature change and a gentle temperature gradient in the infrared image; while the gas area or empty area above the liquid has a temperature change more similar to that of the backplate, appearing as a region with a faster rate of heating or cooling.

[0109] 9) The system identifies the boundary locations of regions with significantly different rates of temperature change, uses these boundary locations as the boundary between the liquid region and the non-liquid region, and maps the height of this boundary in the image to the current infusion liquid level height H(t).

[0110] 10) To avoid momentary interference caused by patients turning over at night, being covered by blankets, or slight shaking of the IV stand, the system does not use a single infrared recognition result for direct judgment. Instead, it compares the results of multiple consecutive infrared image recognitions of the same IV level. Only when the IV level boundary shows a consistent trend of change in 2 to 3 consecutive acquisitions is the change of the IV level confirmed as valid.

[0111] IV. Verification of Redundancy Identification for Weight Changes at Night 11) While the infrared liquid level recognition is running continuously, the weight sensing unit at the infusion hook continuously collects the weight change data of the infusion container, forming a curve W(t) of weight change over time.

[0112] 12) The system compares and analyzes the infrared liquid level change trend with the weight change trend: When the infrared liquid level display shows a gradual decrease in liquid level, and the weight data shows a synchronous downward trend, the system confirms that the infusion status is normal. When the infrared liquid level recognition is unstable for a short period of time (e.g., increased image noise), but the weight change continues to decrease, the system maintains the "normal infusion" judgment and delays the alarm. When the infrared liquid level indicator is nearing the end and the rate of weight change slows down significantly or tends to stabilize, the system determines that the infusion is about to end.

[0113] V. Nighttime triage alarm and nurse station notification 13) The system classifies the infusion status based on the infrared liquid level height H(t), the rate of liquid level descent, and the weight change. For example: Nighttime alert level: Approximately 10 minutes remaining; Nighttime processing level: Estimated remaining time 3-5 minutes; Nighttime Emergency Level: Infrared indicator shows liquid is about to disappear or abnormal weight changes.

[0114] 14) When the "Nighttime Processing Level" or "Nighttime Emergency Level" is reached, the system pushes alarm information to the nurse station workstation and nurse mobile terminal through the communication and notification module. Nighttime alarms are output in a low-interference mode, without generating audio and visual prompts in the wards, and only reminding the patient at the nurse station or on the nurse terminal.

[0115] 15) The alarm information should include at least: bed number, patient ID, current liquid level status, estimated remaining time, current judgment basis (infrared liquid level + weight verification), and a summary of the most recent infrared identification result.

[0116] VI. Nurse's Arrival and Determination of End of Infusion 16) After receiving the alarm information, the nurse goes to the corresponding ward to handle the patient's infusion status, including preparing to change the infusion container or removing the needle.

[0117] 17) When the nurse clamps the infusion tubing and removes the needle, if the system detects that the weight of the infusion container no longer changes and the infusion container is gradually removed or no longer shows stable liquid area features in the infrared image, the system determines that the infusion has ended.

[0118] 18) The system records the infusion end time Tend and marks this nighttime infusion task as "completed". It automatically terminates the infrared acquisition, backplate heating and weight monitoring of the infusion position and releases system resources.

[0119] VII. Termination of Data Archiving and Monitoring 19) After the infusion is completed, the system archives the data for the entire nighttime infusion process, including: Total infusion time Infrared liquid level change recording Weight change curve Nighttime alarm trigger frequency and nurse response time 20) Once the archiving is complete, the infusion task is removed from the night monitoring list, and the system enters standby mode, waiting for the creation of the next night infusion task.

[0120] Example 3 Infusion monitoring under shielding / curtain conditions in the ward I. Infusion Start and System Record Building in Obstructed Scenario 1) In the ward, patients may draw their bed curtains for privacy or rest, partially or completely obscuring the IV stand and IV containers, making direct observation of the containers via visible light or infrared imaging impossible. Nurses perform the IV infusion procedure from outside the curtain or beside the bed, hanging the IV bottle or bag onto the fixed hook of the IV stand beside the bed.

[0121] 2) A backplate module is provided on the rear side of the hook. The backplate module integrates a low-power heating unit and has multiple temperature sensors arranged along the height direction. A weight sensing unit is also provided at the hook position to obtain real-time weight data of the infusion container.

[0122] 3) After the nurse completes the air removal and adjusts the drip rate, she scans the infusion container label or patient wristband information using a barcode scanner. The system records the basic information for this infusion task, including: Ward number, bed number Patient identification information (de-identified ID) Drug types and specifications Infusion start time T0 Initial infusion weight W0 4) When the system detects that the image acquisition conditions corresponding to the infusion position are limited (e.g., the bed curtain is closed, the image signal is missing, or it is marked as an obstruction), it automatically marks the infusion task as "non-direct infusion monitoring mode", disables the image recognition related modules, and only enables the weight detection, back panel heating and temperature sensor detection modules.

[0123] II. Establishment of an environment for back panel gentle heating and temperature detection 5) After entering the non-direct-view infusion monitoring mode, the system activates the heating unit of the backplate module. The heating unit slowly heats the backplate at low power to create a stable and controllable temperature difference between the backplate surface temperature and the ward ambient temperature. For example, when the ambient temperature is about 20℃ to 22℃, the backplate temperature is controlled within the range of 22℃ to 26℃.

[0124] 6) This temperature range is only used to create detectable temperature change conditions and is not intended to directly heat the infusion fluid. The heating process will not cause patients to feel obvious heat and will not affect the stability of the medication.

[0125] 7) Multiple temperature sensors arranged along the height of the back plate synchronously collect temperature change data at each height position during the heating process. Each sensor corresponds to a potential liquid level range of the infusion container, forming a three-dimensional data basis of temperature-height-time.

[0126] III. Indirect Liquid Level Determination Based on Temperature Change Characteristics 8) When there is liquid at the corresponding height position of the infusion container, due to the large heat capacity and strong heat absorption capacity of the liquid, the temperature rise rate detected by the temperature sensor at this height position is significantly lower than that in the area without liquid under the back plate heating condition; after the heating is stopped or reduced, the temperature drop rate at this position is also relatively slow.

[0127] 9) When the liquid has not yet dropped to a certain height, the temperature change curve detected by the temperature sensor at that height shows the characteristics of "slow change and slow response"; when the liquid drops below that height, the temperature change at that position begins to approach the temperature change characteristics of the back plate itself, which is manifested as a significant acceleration in the heating and cooling rates.

[0128] 10) The system compares and analyzes the temperature change rate collected by the temperature sensors at each height position. By identifying the height range where the "temperature change characteristics change abruptly", the boundary position between the liquid area and the non-liquid area is determined, thereby indirectly determining the trend of liquid level change in the infusion container.

[0129] IV. Cross-validation of weight change as the primary judgment channel and temperature judgment 11) In non-direct-view mode, the weight sensing unit serves as the main channel for judging the infusion status, continuously collecting data on the weight change of the infusion container to form a weight-time curve W(t).

[0130] 12) The system performs trend analysis on the weight change data. When the weight shows a continuous downward trend, it determines that the infusion is in progress. When the rate of weight change slows down significantly or tends to stabilize, the system marks that the infusion is nearing its end or that there may be an abnormality.

[0131] 13) The system cross-validates the weight change trend with the liquid level change trend determined by the temperature sensor: When the weight continues to decrease and the temperature reading shows that the fluid level is gradually decreasing, the infusion is confirmed to be normal. When the weight decreases normally but the temperature reading is unstable for a short period of time, the system will prioritize the weight reading and continue monitoring. When the weight change approaches the end and the temperature reading shows the liquid area has disappeared, the system confirms that the infusion is about to end.

[0132] V. Tiered Alarms and Nurse Station Notifications under Non-Direct Visual Conditions 14) The system classifies the infusion status based on weight changes, temperature changes, and consistency assessment results. For example: Advance warning level: The weight display indicates an estimated remaining time of approximately 10 minutes, and the temperature indicates that liquid is still present; Preparation stage: Weight changes are nearing their end, and the temperature-judging liquid zone is about to disappear; Immediate handling level: Weight change stops or becomes abnormal, temperature indicates the liquid is no longer present.

[0133] 15) When the "Preparation for Processing" or "Immediate Processing" level is reached, the system pushes an alarm message to the nursing station through the communication and notification module. The alarm message is clearly marked as "Non-Direct Visual Detection Mode" and includes: bed number, patient ID, current judgment result, estimated remaining time, and detection basis (weight + temperature).

[0134] VI. Nurse's Arrival and Determination of End of Infusion 16) After receiving the alarm information, the nurse goes to the corresponding ward, opens the bed curtain and handles the infusion situation, including changing the infusion container or removing the needle.

[0135] 17) When the nurse clamps the infusion tubing and removes the needle, the system detects that the weight of the infusion container no longer changes, and the temperature change characteristics detected by multiple temperature sensors all return to the temperature response characteristics of the backplate itself. The system then determines that the infusion has ended.

[0136] 18) The system records the infusion end time Tend and automatically terminates the weight monitoring, backplate heating and temperature sensor acquisition at that infusion level.

[0137] VII. Monitoring Termination and Data Archiving 19) After the infusion is completed, the system archives the data for the entire non-direct-view infusion process, including: Total infusion time Weight change curve Temperature change characteristics record Alarm trigger count and nurse response time in non-direct-view mode 20) Once the archiving is complete, the infusion task is removed from the system monitoring list, and the system enters standby mode, waiting for the next infusion task to be created.

[0138] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A liquid level detection and automatic alert system for medical infusion processes, characterized in that, The application relates to a kind of intelligent infusion monitoring system, comprising: Infusion hanging module, comprising a hook for hanging an infusion bottle or bag, and a weighing sensor unit arranged on the hook for obtaining real-time weight information of the infusion container; Information identification and binding module, arranged on the infusion rod, comprising a code scanning identification device for obtaining drug category, specification and patient association information when the infusion container is hung, and binding the initial weight information obtained by the weighing sensor unit; Liquid level detection module for detecting the state of the liquid in the infusion container, the liquid level detection module comprises at least one or more of the following detection methods: weight change-based detection method, image recognition-based liquid level detection method, temperature difference-based liquid level detection method and heat conduction characteristic-based detection method; Pipeline state detection module, arranged on the infusion rod, for detecting the flow state of the infusion pipeline; Data processing and judgment module for analyzing or fusing the data obtained by the liquid level detection module and the pipeline state detection module, and judging the infusion state according to a preset threshold or model; Warning output module for outputting warning information when detecting that the infusion liquid level is close to the end or an abnormal state occurs; Communication and notification module for sending the warning information to a workstation or a mobile terminal to realize remote prompting and disposal linkage.

2. A liquid level detection and automatic alarm system for medical infusion processes according to claim 1, characterized in that, The liquid level detection module comprises a weight change detection module, which continuously collects the weight change of the infusion container during the infusion process based on the weighing sensor unit arranged on the infusion hanging module, records the infusion liquid volume change in combination with the initial weight information obtained by the information identification and binding module, and transmits the information to the data processing and judgment module for comparison with the initial weight information.

3. The liquid level detection and automatic alarm system for medical infusion process as claimed in claim 1 wherein, The liquid level detection module comprises an image acquisition and identification module, which uses a high-definition camera to acquire video of the infusion container, analyzes the acquired images to identify the liquid level height in the infusion container; wherein the image acquisition and identification module continuously monitors the infusion process in the patrol state, and sends the analyzed images or liquid level identification results to the operation site through the communication and notification module for remote viewing and management by the staff.

4. A liquid level detection and automatic alarm system for medical infusion processes according to claim 3, characterized in that, The infusion hanging module is provided with a back plate module on the back of the hook for hanging the infusion container, the back plate module is located on the back side of the infusion container after the infusion container is hung, and forms a relatively fixed positional relationship with the infusion container.

5. A liquid level detection and automatic alarm system for medical infusion processes as claimed in claim 4 wherein, When the image acquisition and identification module identifies the liquid level of the infusion container, the back plate module is provided with color identification, scale identification or visual calibration structure, and the image acquisition and identification module analyzes the relative positional relationship between the liquid in the infusion container and the back plate by using the identification to obtain the liquid level height information in the infusion container.

6. A liquid level detection and automatic alarm system for medical infusion processes as claimed in claim 4 wherein, The back plate module is provided with a heating unit for forming a preset temperature field for the infusion container during the infusion process, and the image acquisition and identification module has infrared image acquisition function, which acquires infrared images of the infusion container, and analyzes the temperature distribution difference between the liquid area and the non-liquid area to obtain the liquid level height information in the infusion container.

7. A liquid level detection and automatic alarm system for medical infusion processes as claimed in claim 4 wherein, The back plate module is provided with a heating unit and a temperature sensor corresponding to the heating unit. Under a preset heating condition, whether there is liquid in the corresponding position of the infusion container is determined by detecting the temperature change characteristics or temperature dissipation rate difference of the back plate module in the contact or close area with the infusion container, so as to obtain the liquid level state information in the infusion container.

8. A liquid level detection and automatic alarm system for medical infusion processes as claimed in claim 4 wherein, The pipeline state detection module includes a pipeline alarm arranged on the infusion rod. The upper end pipe segment of the infusion pipeline is clamped in the pipeline alarm during use, for detecting the flow state or abnormal state of the infusion pipeline, and taking the detection result as the triggering condition of the warning output module.

9. A liquid level detection and automatic alarm system for medical infusion processes as claimed in claim 1, wherein, The back plate module is provided with a night light display structure or a self-luminous display structure, for forming an identifiable background in a low-illumination or night environment, so that the image acquisition and recognition module can recognize the liquid level of the infusion container under different illumination conditions.

10. A liquid level detection and automatic alarm system for medical infusion processes according to any one of claims 1 to 9, characterized in that, The data processing and judgment module is configured to judge the liquid level state and the pipeline state during infusion. The judgment of the liquid level state is based on one or more of the following detection methods: a detection method based on the weight change of the infusion container; a liquid level detection method based on image acquisition and recognition; an infrared image liquid level detection method based on the temperature field formed by the back plate module; a liquid level detection method based on the temperature dissipation characteristics of the back plate module; The warning output module is configured to output warning information according to the liquid level state or the pipeline state. The communication and notification module is configured to send the warning information to the workstation or the mobile terminal.