CN113274584B

CN113274584B - Injection speed management method based on image processing

Info

Publication number: CN113274584B
Application number: CN202110581729.2A
Authority: CN
Inventors: 金姬延; 李葆华
Original assignee: Peking University Third Hospital Peking University Third Clinical Medical College
Current assignee: Peking University Third Hospital Peking University Third Clinical Medical College
Priority date: 2021-05-26
Filing date: 2021-05-26
Publication date: 2022-05-03
Anticipated expiration: 2041-05-26
Also published as: CN113274584A

Abstract

An injection speed management method based on image processing utilizes infusion detection equipment which is used for detecting the drop rate of liquid drops in a Murphy's dropper when a patient is infused according to an image detection mode, the infusion detection equipment comprises an image acquisition part and a data processing part, the image acquisition part acquires images of the Murphy's dropper and sends the images to the data processing part for processing, the data processing part detects the images from the image acquisition part according to a graph matching method, the Murphy's dropper images are searched according to a preset Murphy's dropper model, the images or the image changes of liquid level lines and the liquid drops in the Murphy's dropper image range corresponding to the preset Murphy's dropper model are continuously analyzed, and the drop drops are determined based on the liquid drops and/or the liquid level line change images caused by the liquid drops.

Description

Injection speed management method based on image processing

Technical Field

The invention relates to the field of medical infusion, in particular to an injection speed management method based on image processing.

Background

In the existing medical system, infusion is widely used in medical services of various large medical facilities as one of the common nursing and treatment contents. In infusion management, a nurse needs to regularly patrol the infusion condition of a patient, complete the work of adjusting the infusion dropping speed, predicting the residual infusion time and the like, the process is complicated, and the mode of checking parameters such as the dropping speed and the like by manually and visually observing the Murphy's dropper has large errors and is easy to generate misreading. Infusion pumps that are heavily used are not only constant in use but also costly. In recent years, some devices for assisting medical staff in checking the infusion condition appear on the market, so that the nursing work of the medical staff is relieved to a certain extent.

CN201110955 provides a medical infusion speed monitoring and controlling device based on machine vision, which comprises an image acquisition device, a parameter input device, a data processor, a data display device, a dripping speed control device and an alarm prompting device; the data processor comprises an image processing module, a data display module, a data analysis module, an alarm prompt module and a dripping speed control module. The utility model discloses a speed of infusion is calculated to the method that adopts image processing, and is reliable and stable, has solved the detection inaccuracy that easily receives external radiation source interference and utilizes the principle of weighing to lead to the fact when the infrared light detection method who uses at present generally exists and has detected because infusion bottle/bag weight differs, the problem of operation complicacy. The device can automatically adjust the infusion speed to ensure that the infusion speed is the same as or similar to the input expected dropping speed, and can also give an alarm to abnormal conditions without liquid medicine or/and in other infusion processes, thereby relieving the pain of patients, relieving the workload of medical staff and reducing the occurrence of medical accidents.

CN105664297A provides a method, a system and equipment for monitoring transfusion, which sets the transfusion monitoring equipment; shooting a video of the drip bag with the infusion monitoring device; and displaying the processing result of the infusion monitoring equipment according to the setting result of the infusion monitoring equipment. Video data of the drip bag is obtained according to the set time length, and the video data is processed to obtain the drip speed of liquid drops in the drip bag, so that medical staff or patients can control the drip speed, and the controllability of infusion is improved. Further, according to the set volume and the set dropping speed of the liquid to be input, the time required by the completion of the dropping of the liquid to be input is obtained, countdown is performed, and when the countdown is completed, an alarm is given out, so that the safety of infusion is improved.

CN103127583A is an infusion monitoring device based on video images. The device comprises an infusion bottle monitoring probe, a drip chamber monitoring probe, an infusion monitoring terminal, a signal wire, a network interface and a direct current converter; the infusion bottle monitoring probe and the drip chamber monitoring probe are both connected with the embedded system processor through signal lines, the power input end of the infusion monitoring terminal is connected with a power supply through a direct current converter, and the infusion monitoring terminal is connected with a nurse console through a network interface. The monitoring probe is used for acquiring video images of the infusion bottle and the drip chamber, the embedded system is used for carrying out real-time signal processing on the images of the infusion bottle and the drip chamber to acquire infusion dripping speed and residual liquid medicine information, and an alarm for over-fast infusion, over-slow infusion and infusion completion is sent out according to a preset threshold. The device adopts a non-intrusive scheme, can monitor for 24 hours, has no special requirements on the infusion device, does not interfere with the traditional manual monitoring, is convenient and flexible to install, and can greatly reduce the labor intensity.

In the prior art, some methods for detecting liquid drops in the Murphy's dropper by using an image testing method are proposed, but in the methods, a single color plate is used or liquid drops are searched in a mode of comparing pixels of front and back video frames or other image parameters, wherein the problem of liquid drop identification under the condition that more than two liquid drops simultaneously appear in the Murphy's dropper at the same time is not considered, and the problem of how to eliminate the liquid drops which are hung on the wall of the transparent Murphy's dropper is not involved, and a scheme for optimizing clipping aiming at an image to improve the efficiency of image data processing is not proposed.

Furthermore, on the one hand, due to the differences in understanding to the person skilled in the art; on the other hand, as the inventor studies a lot of documents and patents while making the present invention, but the space is not detailed to list all the details and contents, however, this invention doesn't have these prior art features, but this invention has all the features of the prior art, and the applicant reserves the right to add related prior art in the background art.

Disclosure of Invention

In order to solve at least part of the defects in the prior art, the invention provides an injection speed management method based on image processing, which utilizes an infusion detection device, which is used for detecting the dropping speed of liquid drops in a Murphy's dropper when a patient is transfused in an image detection mode, the transfusion detection device comprises an image acquisition part and a data processing part, the image acquisition part acquires images of the Murphy's dropper and sends the images to the data processing part for processing, the data processing part detects the images from the image acquisition part according to a pattern matching method, searching Murphy's dropper images according to a preset Murphy's dropper model, continuously analyzing the liquid level line and the images or the image changes of the liquid drops in the Murphy's dropper image range corresponding to the preset Murphy's dropper model, and determining the dropping of the liquid drops based on the liquid drops and/or the liquid level line change images caused by the liquid drops.

The method for continuously detecting the liquid level line image fluctuation in the Murphy's dropper image and simultaneously detecting the liquid drop image according to the image matching mode can prevent the situation that the liquid drop is not counted when the liquid drop is not identified by the image due to low transparency or irregular shape of the liquid drop, and can also prevent the situation that the liquid level line is mistakenly judged to have the situation that the liquid drop is generated for multiple times due to the continuous shaking caused by the interference of external factors. The two judgment modes have causal relationship and can be used as mutually complementary evidences to judge whether real liquid drops drop or not, the accuracy of liquid drop judgment of the invention is improved, and errors caused by some external factors are prevented

Preferably, when the data processing portion divides the image data in the contour line of the murphy's dropper with the liquid level line image as a reference, the liquid drop image is matched and identified with a preset liquid drop model as a reference in the image range in the murphy's dropper above the liquid level line, wherein, within a preset interval time, the liquid drop matched and identified is repeatedly detected, the top position of the liquid drop is marked, and the dropping speed is calculated according to the change of the top position along with time, so as to determine effective dropping and/or eliminate the interfering wall-sticking liquid drop.

The top position of the droplet image is set as a reference base line to calculate the dropping speed of the liquid, and the real droplet dropping and the wall-attached droplet dropping are distinguished according to the dropping speed which is slower than the dropping speed according to the rule that the wall-attached droplet dropping speed is slower than the dropping speed, so that the interference of the wall-attached droplet is favorably eliminated, and the identification precision of the invention is improved.

Preferably, the data processing part identifies the liquid level line image by taking the liquid level line fluctuation model as a reference, wherein effective dripping is comprehensively determined and/or interference liquid level line jitter is eliminated according to a method of judging whether the fluctuation amplitude is larger than a threshold value and inquiring whether a record of the liquid drop image is identified before the fluctuation.

Preferably, the infusion detection device further comprises an illumination part, under the condition that the infusion medicine is not transparent and/or has high viscosity, the data processing part controls the illumination part to perform stroboflash according to a certain frequency, and the data processing part performs image recognition on the droplets according to the pixel difference of the droplets in the Murphy's dropper images in the light state and the dark state so as to obtain higher recognition accuracy.

Under the condition that the liquid medicine is not transparent, a pixel region with the maximum change intensity of the reflected light in the images in the light state and the dark state is inspected according to the judgment that the opaque object reflects more light compared with the transparent object, and the region is used as a liquid drop identification basis, so that the identification precision of the liquid drop is improved, and the condition that the liquid drop image identification is invalid is avoided.

Preferably, the data processing section cuts the image from the image capturing section on the basis of the contour line of the murphy dropper so that only the murphy dropper image within the murphy dropper contour line is finally retained as an analysis range, and subjects the murphy dropper image to monochromatization or high-contrast processing so that an irrelevant background at the periphery of the murphy dropper image is excluded.

The design is beneficial to improving the data processing efficiency of the data processing chip adopted by the invention, and image differences caused by the differences of the cameras are eliminated by operations such as cutting and monochromization, so that the equipment adopted by the invention can adopt the cameras of different brands and different models as image acquisition parts. In addition, the personnel who hold the equipment for shooting do not need to be restricted by the rigid requirement of placing the Murphy's dropper in the center of the picture, and the invention ensures that the user can more relax use the equipment for collecting the image by automatically searching the image range of the Murphy's dropper and cutting and deducting irrelevant background around the Murphy's dropper.

Preferably, the data processing section judges whether the image from the image pickup section is clear on the basis of the contour line image of the Murphy's dropper, sends an instruction to the image pickup section to refocus the same when judged to be unclear, and repeats the judgment process until the image is judged to be clear.

Preferably, when the data processing part obtains the infusion dosage, the infusion device model and the liquid drop volume data, the data processing part can calculate the residual dosage and the residual infusion time parameter according to the data and the dropping speed.

Preferably, the method further comprises the following steps: the central server at least comprises a database and is used for managing patient files and communicating data with the infusion detection equipment, and the infusion detection equipment further comprises a display part which is connected to the data processing part to obtain and display image data, dropping speed, residual medicine amount, residual infusion time and alarm information. The infusion detection device further comprises a data transmission part which transmits the infusion data and the patient identification information to the central server, and the central server detects the patient file in the database according to the patient identification information and updates the patient file according to the infusion data.

Preferably, in the case that the data processing part obtains the designated dropping speed range data, the data processing part can judge that the dropping speed value is within the dropping speed range, and if the dropping speed value is not within the dropping speed range, the data processing part sends out an alarm instruction, wherein the designated dropping speed range data can be provided by a scheme preset in the central server database.

Preferably, when the data base of the central server stores the infusion dosage, the model of the infusion device and the volume data of the liquid drops in advance, the infusion detection device asks the central server for the parameters through the data transmission part and calculates the residual dosage and the residual infusion time parameters.

Drawings

FIG. 1 is a schematic circuit diagram of an infusion detection device of the present invention;

FIG. 2 is a diagram of the steps for identifying a fluid image by the infusion detection device of the present invention;

fig. 3 is a substep diagram of the step S1 in the step of recognizing a droplet image according to the present invention;

FIG. 4 is a substep diagram of step S2 in the step of identifying a drop image according to the present invention;

fig. 5 is a substep diagram of the step S3 in the step of recognizing a droplet image according to the present invention;

fig. 6 is a substep diagram of the step S4 in the step of recognizing a droplet image according to the present invention;

FIG. 7 is a schematic view of an image in the Murphy' S dropper of step S4 in the step of identifying a droplet image according to the present invention;

in the figure: 100. an infusion detection device; 110. an image acquisition unit; 120. a data processing unit; 130. a display unit; 140. a data transmission unit; 200. central server

Detailed Description

The following detailed description is made with reference to the accompanying drawings.

The invention provides an injection speed management method based on image processing, and in order to realize the purpose of injection speed management, the system adopted by the invention at least comprises two parts with different division of labor, namely an infusion detection device 100 and a central server 200. The infusion detection device 100 is applied to a medical worker for detecting infusion parameters such as infusion dropping speed, infusion duration and the like in the process of examining an infusion patient, namely, the device belongs to a use device on one side of a patient. The central server 200 is a central storage or management device for storing the medical facility or all patient-related cases or medical records in a certain range, and is a large device on the medical facility or regional medical center side because the information is usually transmitted by long-distance lines or wireless transmission. The invention at least has the technical effect that medical staff can randomly check the parameters such as the dropping speed, the dropping condition, the transfusion time, the color change of liquid medicine and the like in the general transfusion equipment adopted by the patient in the process of transfusion by utilizing the transfusion detection equipment 100 to perform graphic imaging on one side of the patient or check the parameters in real time. And under the condition that the parameters are displayed on the infusion detection device 100 for the medical staff to confirm on site, the parameters are communicated with the central server 200 of the medical facility or medical management center, so as to realize the real-time update of the patient file or the feedback adjustment of the infusion process according to the patient file.

Fig. 1 shows a schematic functional diagram of an infusion detection device 100, which at least includes an image acquisition unit 110, a data processing unit 120, a display unit 130, and a data transmission unit 140. Since the present infusion detection device 100 is a device for medical personnel and for checking the infusion parameters of a patient at any time, in most cases, the present infusion detection device 100 will be carried on the body of the medical personnel, and since the present infusion detection device 100 has a simple structure and does not require a complicated circuit design, the present infusion detection device 100 can be designed as a small and exquisite instrument that is portable. Preferably, the present infusion detection device 100 can be implemented by a portable personal mobile terminal, such as a smart device with a camera and a data processing chip, e.g., a mobile phone, a PAD, a handheld terminal, etc.

The image capturing part 110 for detecting the general infusion device by using a visual image manner may be configured as an instrument device with a camera part, and in order to measure the dropping speed of the liquid medicine in the general infusion device, it is generally selected to align the image capturing object of the image capturing part 110 on the murphy dropper in the general infusion device. The Murphy's dropper is usually designed into an olive-shaped or ellipsoidal transparent pipeline, when the liquid medicine is delivered into the body of a patient through the transparent pipeline, at least one part of the Murphy's dropper forms a certain liquid level due to the pressure in the body of the patient, and a hollow space is formed at the other parts of the Murphy's dropper, and the liquid medicine flowing into the upper medicine bottle drops into the liquid level at the lower half part of the Murphy's dropper in a dropping manner. Due to the basic working principle of the Murphy's dropper, the image acquisition part 110 can realize the function of acquiring the droplet image in the Murphy's dropper.

The process of the image acquisition part 110 acquiring the droplet image in the Murphy's dropper may be that the image acquisition part 110 adjusts focusing to ensure that the clearest view image is obtained on the Murphy's dropper. The image collecting part 110 photographs or recognizes the liquid droplets in the murphy's dropper in a manner of detecting a moving droplet-shaped object in the murphy's dropper in real time. The process of identifying the dropper by the image acquisition part 110 can be realized only by using the processing function of the data processing part 120 in the infusion detection device 100, and the specific process may be that the image acquisition part 110 continuously shoots a clear image or a monitoring video in the Murphy's dropper and transmits the image or the monitoring video to the data processing part 120, wherein the data processing part 120 adopts the following mode to distinguish the protrusion of the moving liquid drops.

According to different collecting modes of the image collecting part 110, the image data obtained by the data processing part 120 can be divided into two forms, one of which is that the image collecting part 110 automatically takes pictures of the Murphy's dropper according to a certain set time interval, and because the dropping time of the liquid drops is relatively short, the set time interval is required to be relatively short so as to ensure that the image collecting part 110 can collect at least one time of images of completely dropping liquid drops in the air in the process of dropping liquid drops at least one time; secondly, the image acquisition part 110 continuously performs the photographing operation on the Murphy's dropper, and the continuous video image in the Murphy's dropper can be obtained by adopting the photographing mode, so that the method has the advantage of detecting the liquid drops in motion.

In this embodiment, the process of the data processing unit 120 processing the image data acquired by the image acquisition unit 110 and finally obtaining effective infusion parameters such as dropping speed is roughly as follows (fig. 2):

s1, judging the definition of the image data collected by the image collecting part 110, and controlling the image collecting part 110 to refocus when the definition does not meet the requirement;

s2, processing the clear image data according to the mode of cutting the background and monochromating;

s3, searching a liquid level line in the processed image data by an image matching method;

s4, dividing the Murphy' S dropper image area by taking the searched liquid level line as a reference line, searching for liquid drops by an image matching method aiming at the area above the liquid level line, and recording the time when the liquid drops are searched;

and S5, calculating the drop speed according to the recording time of different drops.

In step S1, the data processing unit 120 determines the sharpness of the image of the murphy' S dropper transmitted by the image capturing unit 110, and the determining process includes at least the following sub-steps (fig. 3):

s1a, searching an ellipsoidal or olive-shaped contour line of the Murphy 'S dropper in image data by taking a preset Murphy' S dropper model as a reference;

s1b, judging whether the Murphy ' S dropper image is clear according to a method for comparing the contrast between adjacent pixels at the edge of the contour line of the Murphy ' S dropper or judging the frequency component of the Murphy ' S dropper image, if so, continuing the step after S2 by adopting the clear image by the data processing part 120, and if not, performing the step of S1 c;

s1c, the data processing unit 120 sends a refocusing instruction to the image capturing unit 110;

s1d, the image capturing unit 110 receives the refocusing command, re-adjusts the focal length, and transmits the image data after the focal length adjustment to the data processing unit 120, and the data processing unit 120 repeats steps S1a and later until the image is determined to be sharp.

The above-mentioned process of judging whether clear to the image of image acquisition portion 110 collection also can be accomplished by medical personnel's manual work, and the image of specifically being image acquisition portion 110 collection exports medical personnel through display part 130 simultaneously, and medical personnel judge through the naked eye whether accurate focusing, if medical personnel judge to focus inaccurately, then can send the instruction of focusing once more to image acquisition portion 110 through input device. Preferably, when the infusion detecting device 100 is an intelligent handheld device such as a mobile phone, the medical staff touches the fizeau dropper area on the display unit 130 to send a focusing instruction of a designated area to the image collecting unit 110, so as to achieve a faster focusing effect.

After the focusing process is completed, the data processing unit 120 starts to continuously receive the clear image of the murphy's dropper collected by the image collecting unit 110 and further processes the image to obtain a condition for more conveniently detecting the moving dropping liquid drops, the data processing unit 120 may adopt a mode of removing the blurred background according to the blurred background caused by the difference of the focal lengths, and the data processing unit 120 removes the blurred background according to the step of judging whether the image is clear. Or the area surrounded by the determined contour line of the Murphy's dropper is used as a detection area, and the image outside the area is cut off to remove interference. And the remaining Murphy's dropper image area deducts the interference of complex colors from the image by carrying out monochromatization on the image or high-contrast processing and other processing modes, and finally obtains image information with high contrast, and the image only presents or retains the visual information in the Murphy's dropper range. The mode of further refining and splitting the image acquisition area enables the data processing part 120 to reduce the attention area as much as possible, thereby reducing the data processing difficulty and improving the image identification precision. In addition, in a complicated medical facility, the influence of other complicated backgrounds in a patient, a medical staff or a ward is deducted as much as possible, so that the data processing part 120 can more accurately match the droplet images in the Murphy's dropper without being interfered by the complicated backgrounds.

Substeps (fig. 4) which convert the above steps to step S2:

s2a, cutting image data except the contour line by using the contour line of the Murphy 'S dropper determined in S1a as a reference, and finally only keeping the image data inside the contour line of the Murphy' S dropper;

and S2b, carrying out monochromatization or high-contrast processing on the cut image.

Then the data processing part 120 searches and fixes the position of the liquid level line of the discharged liquid medicine in the Murphy's dropper image area by using a preset liquid level line profile model as a reference. Preferably, the liquid medicine generally forms a liquid seal with a certain height in the Murphy's dropper due to the influence of gravity and pressure, and the liquid level basically occupies 1/2 to 2/3 of the overall height of the Murphy's dropper according to common medical experience, so that when the data processing unit 120 searches for the position of the liquid medicine liquid level line, it is possible to preferentially search for whether a liquid level line matching a preset model exists at 1/2 to 2/3 of the height of the contour line of the Murphy's dropper in the image, and if not, the search is continued at the rest part. The above-mentioned preferred method is a simplified condition according to common medical experience, and the operation process of the data processing unit 120 for finding the liquid level line is simplified, so that the whole system is more efficient and faster.

The above step is converted into a substep of S3 (fig. 5):

s3a, dividing the image processed in the S2 step into regions of heights 1/2 to 2/3, searching a liquid level line image in the region by taking a preset liquid level line model as a reference, and performing the steps S4 and later if the liquid level line image is matched, or performing the step S3 b;

s3b, the area outside the area divided in the step S3a is searched for the liquid level line image in the manner described in the step S3 a.

After the data processing unit 120 finds the liquid level line, the liquid level line is continuously tracked with the liquid level line as a key object of interest, which may be specifically the data processing unit 120 continuously finds and marks the liquid level line in the image collected and transmitted by the image collecting unit 110. When the image acquisition part 110 acquires images in the form of taking pictures at intervals, the data processing part 120 searches for a liquid level line in each picture received in time sequence; when the image pickup section 110 picks up an image in the form of a video, the data processing section 120 keeps tracking the liquid level line in a manner of finding the liquid level line on each frame.

After the liquid level line is found, the data processing unit 120 divides the image into two parts in the Murphy's dropper image according to the found liquid level line as a division basis, wherein the area below the liquid level line is a liquid area filled with a liquid chemical, and the area above the liquid level line is a hollow area. The data processing unit 120 may find an image of the droplet landing with the region above the liquid level line as an important attention region in the following manner.

The data processing unit 120 performs pattern recognition and matching on an upper area of the murphy's dropper divided by a liquid level line according to a preset water drop or liquid drop model, wherein the matching sequence is searched frame by frame according to image pictures sent by the image acquisition unit 110 one by one or videos. When the data processing section 120 matches the graphic shape of one droplet for the first time in a picture or video frame, the picture or video frame is highlighted and the next valid droplet record is recorded. The above process is called an effective matching process, and the effective droplet data may include information such as the time of the moment and the image data of the moment. The recorded picture information is preferably stored according to the original image captured by the image capturing part 110, so that the medical staff can conveniently display the original image to the relevant staff when necessary.

According to the above-mentioned manner of matching droplets, the images continuously acquired by the image acquisition unit 110 are continuously matched and monitored, so that the data processing unit 120 can store at least one time and image information each time a droplet is dropped. The time information is used to calculate the dropping speed of the liquid, and can be converted into drops of liquid per minute.

Preferably, in order to prevent the data processing unit 120 from recognizing the image of the same droplet at different times as two liquid drops, the data processing unit 120 marks the top position of the contour line of the droplet after recognizing the droplet pattern for the first time, compares the top position of the recognized liquid contour line in the subsequent picture or video frame with the top position recorded at the previous time, and if the top position is lower than the position recorded at the previous time, the recognized droplet and the droplet recognized at the previous time are the same droplet, and the data processing unit 120 does not record the time and the image data at this time.

Preferably, for a droplet that rapidly drops in a short time, the data processing unit 120 continuously updates the top position of the same droplet with time in the process of recording the droplet for the first time and continuously determining the same droplet subsequently, so that the top position of the droplet is effectively tracked in the continuous time. The data processing unit 120 divides the image into two parts, i.e., the upper part and the lower part, by dividing the image by parallel lines at the top position, and if the data processing unit 120 fits the droplet image again at the upper part of the position, the data processing unit 120 determines that a new droplet is dropped, and records the time and the image data at that time. In the case where 3 drops or more are simultaneously present in the hollow portion of the murphy's dropper at the same time or in a short time, the data processing unit 120 records and tracks the top position of the second drop while determining that the second drop is dripping, determines whether a third drop is present according to the above manner, and so on to recognize a plurality of simultaneously present drops. Preferably, after the data processing unit 120 determines a new droplet and marks the top position of the droplet, the data processing unit 120 may not track and record the top position of the last droplet, so as to achieve the purpose of optimizing image processing and improving efficiency.

As the judgment process illustrated in fig. 7, the image data processing section 120 marks its top position as line L1 for different times of the same drop, and the position of the line L1 should be moved gradually downward with time. When a second drop occurs, the data processing section 120 marks the top of the new drop image as line L2, and at this point the data processing section 120 will no longer track the position of line L1, and line L2 will iterate the function of line L1 and apply to subsequent recognition.

Preferably, for the monitoring video data sent by the image capturing part 110 capturing image data by using a shooting video method, the data processing part 120 can also determine moving liquid drops by using a method different from a method of performing pattern matching by using a liquid drop model. Specifically, the data processing unit 120 checks the video frame by frame, and if the image pixel change in a continuous video frame is not greater than a preset error threshold, it is determined that no droplet is dropped in the murphy's dropper at this time, and if the data processing unit 120 finds out that a pixel change greater than the error threshold has occurred in the comparison between the previous frame and the next frame of the video frame, and the pixel change continues to be greater than the error threshold in a subsequent period of time, it may be determined that a droplet has occurred at this time, and the data processing unit 120 records the time and the image data at this time. The error threshold is set to prevent the data processing unit 120 from making an erroneous determination due to a small range of pixel changes on the image caused by a slight shake of the Murphy's dropper, incomplete background removal, sliding of a residual droplet on the Murphy's dropper, or the like.

Preferably, the data processing unit 120 may further use the image change of the liquid level line to assist in determining whether or not a droplet is generated (as shown in the rightmost small diagram in fig. 7), since the liquid level line fluctuates when a droplet lands on the liquid level line. The specific process is that after the data processing unit 120 finds the liquid level line and continuously tracks the liquid level line as an important object of interest, the data processing unit 120 determines and analyzes the shape of the liquid level line within a duration time, if the liquid level line has an arc shape that is approximately a straight line or a little concave and convex within a period of time, the data processing unit 120 determines that no liquid drop is dropped into the liquid level line at the time, and if the liquid level line has an irregular wavy shape at a certain moment, the data processing unit 120 determines that a liquid drop is dropped into the liquid level line at the time, and it is described that a liquid drop appears in the hollow portion before the time. The testing method for detecting the liquid level line type and the method for searching the liquid drops through image matching can be used in a combined mode, so that the problem that the liquid drop image identification fails due to the fact that the shapes of the liquid drops are irregular at some time can be solved, the problem that the liquid level line type detection method fails due to the fact that the liquid drops are continuously in a fluctuation state due to rapid dropping of the liquid drops can be solved, and the identification accuracy of the infusion detection device 100 is improved.

The above-described step is converted into a substep of S4 (fig. 6):

s4a, dividing the image data inside the contour line of the Murphy' S dropper according to the liquid level line image found in the step S3 as a reference;

s4b, matching and recognizing a droplet-shaped or spherical droplet image in the image area above the liquid level line by taking a preset droplet model as a reference, and recording the moment and image data when finding a matched droplet image for one time;

s4c, after the droplet image is detected for the first time, marking the top position of the droplet image contour line, and dividing the image data in the Murphy' S dropper contour line again by taking the bottom position as the reference;

s4d, when the liquid drop images are matched again in the subsequent time, judging whether the position of the liquid drop image is positioned below the top position marked in the step S4c, if so, executing the step S4e, and if not, executing the step S4 f;

s4e, updating the top position marked in the previous step to be the top position of the current liquid drop searching image;

s4f, recording the time and the image data, and updating the top position marked in the previous step to the top position of the current seek droplet image.

In an optional case, the step of S4 may further include the following substeps:

s4g, identifying the liquid level line image searched in the step S3 by taking the liquid level line fluctuation model as a reference, judging whether the fluctuation amplitude of the fluctuation image is larger than a preset threshold value or not when a matched liquid level line fluctuation image is searched for once, if the fluctuation amplitude is smaller than the threshold value, not executing the subsequent steps, and if the fluctuation amplitude is larger than or equal to the threshold value, executing the following step S4 h;

s4h, judging whether a liquid drop image is found in the step S4b before the fluctuation image is found, if a record of the liquid drop image is found, recording the time and the image data, setting that the record of the liquid drop found in the time cannot be applied to the judgment process of the step S4h next time, and if no record of the liquid drop image is found, not executing the operation of recording the time and the image data;

the S4g step will change to the following S4i step and not perform the subsequent S4h step after the healthcare worker chooses to use level only identification:

s4i, identifying the liquid level line image searched in the step S3 by taking the liquid level line fluctuation model as a reference, judging whether the fluctuation amplitude of the fluctuation image is larger than a preset threshold value or not when a matched liquid level line fluctuation image is searched for once, if the fluctuation amplitude is smaller than the threshold value, not executing the subsequent steps, and if the fluctuation amplitude is larger than or equal to the threshold value, recording the moment and image data;

the method for judging whether effective liquid drops drop or not by adopting the liquid drop identification method and the liquid level line identification method can judge one-time effective liquid drop by manually selecting a method of only adopting one or both of the methods to trigger.

Preferably, the data processing part 120 calculates and processes the moving distance and the process time of the top position as the dropping speed of the droplet in the process of tracking the top position of the same droplet, and determines whether the speed is greater than a preset value, determines that the droplet is a valid droplet drop if the speed is greater than the preset value, and determines that the droplet image is not a valid droplet drop if the speed is less than the preset value. This design prevents a droplet that sticks to the wall of the murphy's dropper and slowly slides down from being mistaken by the data processing portion 120 for a valid drop. Preferably, the data processing unit 120 records only the droplet dropped at the fast speed as valid data once when detecting the droplets dropped at the fast speed a plurality of times and the droplets dropped at the slow speed a few times at a sporadic time over a period of time. It should be noted that the dropping speed is different from the dropping speed, and the dropping speed is the speed of a droplet during the period from the hollow portion to the liquid level line, and the unit is the distance/time, and the dropping speed is the parameter set by the medical facility for checking the liquid medicine input speed, and the unit is usually the drop/minute.

The steps of S1-S4 are only operation steps of some embodiments of the present invention, and other persons in the art should understand that new technical solutions formed by any of the steps of the present invention and the combination and arrangement of the sub-steps should be included in the disclosure of the present invention.

The above methods for detecting and identifying droplets all refer to the requirement of recording time, so the data processing unit 120 at least comprises a module for recording time, and the module records time according to the time when the data processing unit 120 detects each droplet, and the time may be local standard real time or cumulative time. If the time module records the accumulated time, the time point for starting recording is determined according to the start time manually selected by the medical staff, for example, the medical staff clicks to start measurement through an operation device or an operation interface, and the time point for starting clicking is taken as the time for starting recording of the time module. Preferably, the time module may use the time when the data processing unit 120 first detects the droplet as the recording start time.

If the infusion detecting device 100 is used for a handheld ward patrol of medical staff, after the medical staff aims the infusion detecting device 100 at the Murphy's dropper, the instant when a drop of liquid drops is observed is input through the operation device, a detection starting instruction is input, the image acquisition part 110 and the data processing part 120 start working, and the time module starts to record accumulated time. Alternatively, the time module starts recording the accumulated time at the time when the data processing unit 120 first detects the droplet after receiving the start detection command. The time module then marks the current time each time a new drop is detected by the data processing portion 120. When the medical care personnel inputs an instruction for finishing the detection by using the operation device, the data processing part 120 calls all the time point data marked by the time module, calculates the single dropping speed according to the two time point data before and after and calculates the average dropping speed according to the total number of the drops and the total accumulated time. When the time module records the local standard real time, the accumulated time or the process time is the difference value of the two real times. Different nursing requirements caused by different liquid medicine input or body fat of different patients and the like are met, and different choices are provided for the dropping speed testing precision. The common precision mode is suitable for a conventional infusion mode with low precision requirement, and the detection time of medical staff is short; the high-precision mode is suitable for occasions requiring strict control over the dropping speed. The different data processing unit 120 may collect only 3 continuous droplet images for droplet velocity calculation in the normal precision mode, the data processing unit 120 may collect 6 continuous droplet images for droplet velocity calculation in the high precision mode, and the number of detected droplets in the two modes may be changed in the program. The different precision modes can be switched and selected by the medical staff by using the operating device to send instructions to the data processing part 120, and preferably, at this time, the image acquisition part 110 and the data processing part 120 automatically stop acquiring and recording new liquid drops after the data processing part 120 records 3 or 6 data and calculates the dropping speed.

Preferably, if the medical staff performs fast collection of infusion data on the infusion device when the medical staff starts infusion, the infusion detection device 100 continues to record the accumulated time in the program background after obtaining the average infusion drop speed, calculates the number of dropped drops according to the product of the measured average drop speed and the accumulated time, calculates the volume of the dropped liquid medicine according to the preset volume of one drop, and calculates the predicted remaining infusion time according to the preset weight of the liquid medicine in cooperation with the average drop speed. Wherein, the average dropping speed is updated iteratively by a new average dropping speed obtained after the next medical staff carries out quick detection on the same patient again.

In order to realize the collection of the Murphy's dropper image by the image collection part 110 under the environment of weak light or no light, an illumination part can be arranged on the image collection part 110. The illumination portion can automatically detect the intensity of the light in the current environment and automatically or manually open the illumination portion under the condition of weak light or no light, and the illumination portion emits light to meet the requirement that the image acquisition portion 110 acquires the image.

Preferably, in the case where the input medical fluid is opaque, low in transparency or high in viscosity, the shape of the fluid may not be well recognized by only the above-described pattern of the droplet image matching. Therefore, in this case, the user can manually select or the device can automatically select to turn on the illumination portion, and the illumination portion is set to perform auxiliary illumination, so that the data processing portion 120 can obtain the image of the murphy's dropper in a bright state and a dark state within a certain time. For non-transparent droplets, they generally do not have good light transmission, so when light impinges on them, the droplet reflects more light than the transparent Murphy's dropper wall around it. Therefore, when a region with large pixel change caused by light reflection in light and shade change in an Murphy's dropper image with contrast of light and shade states is obtained, and the pixel composition of the region basically accords with a preset liquid drop model, the region can be judged to be a liquid drop image. Preferably, the illumination unit may be controlled to perform a stroboscopic operation according to a certain set frequency, so that the data processing unit 120 may obtain a plurality of sets of bright and dark images, and the accuracy of the data processing unit 120 in finding the type of liquid drop is enhanced through the above-mentioned analysis and image matching processes.

The display unit 130 disposed on the infusion detecting device 100 is used to display the dropping speed data processed and calculated by the data processing unit 120 to the medical staff, and preferably, other relevant parameters may also be displayed according to different use situations and different use purposes of the infusion detecting device 100, and specific parameters will be mentioned in the following embodiments.

When the present infusion detecting device 100 is connected to the vicinity of the Murphy's dropper in a relatively fixed manner with respect to the dropper, the present infusion detecting device 100 can implement a function of automatically monitoring the dropping speed or other parameters for a long time. Specifically, the display unit 130 displays the single dropping rate calculated by the data processing unit 120 in real time, and may also display the average dropping rate from the start of detection to the present, and display the current accumulated time of infusion. Preferably, the data processing unit 120 calculates parameters such as the amount of the liquid to be infused, the remaining amount to be infused, and the estimated remaining infusion time according to the parameters such as the number of already-dropped drops of the liquid, the average drop speed, the preset drop volume, the preset total amount of the liquid, and the model of the infusion device, and sends the parameters to the display unit 130 for display.

Preferably, the medical staff can input a standard drop number range to the present infusion detecting device 100 before starting long-time detection, and if the data processing unit 120 determines that the current single drop number is higher or lower than the preset standard range, an over-high or over-low warning word or mark will be displayed on the display unit 130. Preferably, the infusion detecting device 100 may further include an alarm portion, and when the data processing portion 120 determines that the current dropping speed is higher or lower than a predetermined standard range or the average dropping speed in a predetermined time period is continuously higher or lower than the predetermined standard range, the data processing portion sends an instruction to the alarm portion and controls the alarm portion to alarm a person near the infusion detecting device 100 by using an audio alarm such as a buzzer or a visual alarm such as a flashing light.

The central server 200 located at one side of the hospital is used for storing, updating in real time, and providing data for medical staff, and has at least functions of storing patient information in a classified manner and certain computing and processing capability. After the infusion detection device 100 detects infusion data of a patient infusion, the infusion data is transmitted to the central server 200 through the data transmission unit 140, and the central server 200 combines and updates the infusion data and the patient file information to obtain an updated patient file containing the patient infusion data, thereby facilitating the management of a large amount of patient infusion histories in the hospital. The infusion data comprises but is not limited to single dropping speed, average dropping speed, accumulated time, starting time, number of dropped drops, amount of dropped medicine, remaining amount to be infused, remaining infusion time and the like, and the infusion data is divided into single data for ward inspection and dynamic change data for continuously detecting the infusion process of a certain patient according to different specific using modes of medical staff. In the process of storing or monitoring the infusion data in different forms by the central server 200, for single-time data, the generation time sequence of the data can be used for sequencing and marking storage, for dynamically changing data, all changed data can be continuously recorded, and statistical information such as a recording table, a trend block diagram and the like is formed for storage processing.

In order to achieve the above-mentioned classification management of the patient infusion data, the central server 200 needs to perform a distinguishing and identifying process on the infusion data transmitted by the infusion detecting device 100, and the distinguishing and identifying process is to identify the patient to which the infusion data transmitted this time belongs, so that the data transmitted from the infusion detecting device 100 to the central server 200 should include identification information in addition to the infusion data, and the identification information can be manually input by the medical staff at one side of the infusion detecting device 100 or preferably obtained by the medical staff through operations such as scanning a two-dimensional code or a bar code containing patient information or a number by using the image collecting part 110. The identification information can be the unique information or the combination of the information corresponding to the real individual of the patient, such as the name, the number, the ID card number, the medical insurance number and the like of the patient. After the identification information is sent to the central server 200 along with the infusion data, the central server 200 searches out a unique corresponding patient file in the database according to the identification information as a mark, adds the infusion data transmitted this time to the file as an update, and then records the updated file in the database.

Preferably, since the central server 200 stores the profile of all information related to the patient and the infusion detecting device 100 and the central server 200 have a data transceiving path, the infusion detecting device 100 can implement more display or recommendation functions according to the profile stored in the central server 200. The specific function may be a function shown in some embodiments as follows.

After the medical staff obtains the identification information of a certain patient by manual input or code scanning before testing the infusion parameters of the patient and transmits the information to the central server 200 through the data transmission part 140, the central server 200 retrieves the patient file from the database, selects file data related to the infusion from the database, and transmits the selected data to the display part 130 of the infusion detection device 100 for display to the medical staff. The infusion-related profile data may include, but is not limited to, the following: the information of the previous medical history of the patient, the allergen of the patient, the previous treatment history of the patient, the information of the liquid medicine used by the patient for the infusion, the advice of the infusion of the patient and the like can comprise sub-information such as a name of a used medicine, a contraindication item of the medicine, main physicochemical properties of the medicine, a medicine specification and the like, and the information of the infusion advice of the patient can comprise a full text of the advice about the infusion or a sub-information such as an infusion dripping speed control range, an appointed infusion amount, an appointed infusion time and an appointed infusion equipment signal which are prescribed by a doctor.

After reading the above-mentioned archive data related to infusion, the medical staff can know the basic situation related to infusion of the patient and use it as the guide standard of the next infusion nursing work. For example, medical care personnel check whether the medicine taking of the patient is correct through the past medical history of the patient, the allergen of the patient, the contraindication item of the medicine and other information, check whether the medicine taking amount is correct through the information of the medical advice information and the information of the transfusion amount specified by the doctor, and check whether the transfusion device used accords with the medical advice through the information of the medical advice and the model of the transfusion device specified by the doctor.

Preferably, when the medical staff obtains the patient infusion dropping speed value in the subsequent infusion parameter test, the patient infusion dropping speed value is compared with the infusion dropping speed control range specified by the doctor in the order information, so as to know whether the dropping speed exceeds the control range specified by the order, and the dropping speed adjustment control is correspondingly made. In other embodiments, the dropping speed control range specified by the medical order is directly input as an instruction into the present infusion detecting apparatus 100, and after the data processing unit 120 determines that the dropping speed value obtained by processing exceeds the dropping speed control range, a warning slogan or icon that is too high or too low is displayed on the display unit 130, and the warning unit may be used to perform a corresponding warning operation.

Preferably, when the infusion detection device 100 is applied to continuously detecting infusion information of a patient for a long time, infusion data containing single infusion speed or average infusion speed in a period of time is monitored by the central server 200 in real time, if the central server 200 judges that the single infusion speed or the average infusion speed in a period of time exceeds the drop number range specified in the medical advice information, an alarm operation is performed at the central server 200, or an alarm notification is sent to a nurse station or related medical staff on a floor to which a ward belongs by the central server 200, so that medical staff near the patient can go to the patient immediately to adjust the infusion speed. Preferably, after the end of the infusion detection device 100 judges that the dropping speed exceeds the specified range, the alarm function is performed on the infusion detection device itself, the data transmission part 140 sends alarm information to the central server 200, and the central server 200 receives the alarm information and then performs self-alarm or alarm notification to medical staff in the above manner, so that the design reduces the two judgment processes of the central server 200 and the infusion detection device 100 into one judgment, optimizes the program design and improves the working efficiency of the system.

Preferably, for a medical facility where big data management has been implemented, after a patient looks at a doctor, an infusion order is ordered by the doctor and entered into a patient profile of the central server 200. The dispensing room is connected to the central server 200 and prepares the medication and the corresponding equipment according to the parameters of the medicine, the dosage, the model of the infusion device and the like specified by the medical order in the patient file, and the identification information of the patient is pasted on the medicine bottle in a mode of printing a two-dimensional code or a bar code. After the materials are received by the dispensing room, the responsible nurse responsible for the infusion work of the patient goes to the infusion position arranged for the patient, the two-dimensional code is scanned by the image acquisition part 110 of the infusion detection device 100, and then the infusion detection device 100 automatically acquires the patient file information from the central server 200 and displays the information to the responsible nurse. Preferably, the responsible nurse also enters information such as the patient's bed number into the infusion detecting device 100 by scanning or manual entry, and the patient profile is updated as it is transmitted from the infusion detecting device 100 to the central server 200. The above mode omits the step of manually inputting the drug information by medical staff, and the patient file information is supported and can be updated in real time from the doctor, the medicine dispensing to the responsibility nurse circulation procedure, so that the medical system is simplified, and the working efficiency of the medical staff is improved.

Preferably, when the medical staff performs rapid detection by using the infusion detection device 100, in order to implement multi-thread calculation management of remaining infusion time of a plurality of patients, the work of tracking accumulated time, calculating the number of dropped drops, calculating the amount of remaining drug, and calculating the remaining infusion time is performed on the central server 200, and the infusion detection device 100 can implement centralized management of infusion states of a plurality of patients at the same time only by communicating the rapidly detected infusion data with identification information for identifying the identities of the patients and sending the identification information to the central server 200, thereby improving the integration level and the operating efficiency of the system, and reducing the calculation pressure of the infusion detection device 100.

Preferably, the above method of the central server 200 for calculating the remaining infusion time and other parameters is also applicable to the case of using the infusion check device for monitoring the infusion data of a certain patient for a long time. Alternatively, the parameters such as the remaining infusion time calculated by the infusion detecting device 100 are transmitted to the center server 200 for the same management.

The central server 200 can push information to a nurse station or a related medical staff on the floor where the patient is located after obtaining the parameters such as the remaining infusion time, or send a prompt that the infusion is to be completed to the place or the person when judging that the remaining infusion time is to be reset to zero, and the prompt is accompanied with the personal file of the patient so that the medical staff can conveniently know the identity of the object completing the infusion.

Preferably, the central server 200 stores a plurality of sets of commonly used infusion protocols, and the infusion protocols at least include parameters such as infusion drug type, drug quantity, infusion device model, infusion dropping speed range and the like. The protocol may be selected by the physician from the central server 200 at the time of ordering. The method has the advantages of saving the time for medical staff to set infusion parameters and improving the working efficiency of the medical staff for the infusion treatment process of a large number of common diseases.

It should be noted that the above-mentioned embodiments are exemplary, and that those skilled in the art, having benefit of the present disclosure, may devise various arrangements that are within the scope of the present disclosure and that fall within the scope of the invention. It should be understood by those skilled in the art that the present specification and figures are illustrative only and are not limiting upon the claims. The scope of the invention is defined by the claims and their equivalents.

Claims

1. An injection speed management method based on image processing,

the method utilizes an infusion detection device (100) which is used for detecting the dropping speed of liquid drops in a Murphy's dropper when a patient is infused in an image detection mode, the infusion detection device (100) comprises an image acquisition part (110) and a data processing part (120), the image acquisition part (110) acquires images of the Murphy's dropper and sends the images to the data processing part (120) for processing,

it is characterized in that the preparation method is characterized in that,

the data processing part (120) detects the image from the image acquisition part (110) according to a pattern matching method, searches Murphy's dropper images according to a preset Murphy's dropper model, continuously analyzes the liquid level line and the image or the image change of the liquid drop in the Murphy's dropper image range corresponding to the preset Murphy's dropper model, determines the dropping of the liquid drop based on the liquid drop and/or the liquid level line change image caused by the liquid drop,

under the condition that the data processing part (120) divides image data in the contour line of the Murphy's dropper by taking a liquid level line image as a reference, matching and identifying a droplet image by taking a preset droplet model as the reference in an image range in the Murphy's dropper above a liquid level line, wherein the droplets matched and identified are repeatedly detected within a preset interval time, the top positions of the droplets are marked, and the dropping speed is calculated according to the change of the top positions of the droplets along with time so as to determine effective dropping and/or eliminate interfering wall-sticking droplets.

2. The injection rate management method according to claim 1, wherein the data processing section (120) identifies the level line image with reference to a level line fluctuation model, and wherein effective dripping and/or interfering level line jitter are comprehensively determined in accordance with a method of determining whether a fluctuation amplitude is larger than a threshold value and inquiring whether a record of the droplet image identified precedes the fluctuation.

3. The injection speed management method according to claim 2, the infusion detecting device (100) further comprises an illumination part, and the data processing part (120) controls the illumination part to perform auxiliary illumination when the infusion medicine is opaque and/or has high viscosity, and the data processing part (120) performs image recognition on the droplets according to the pixel difference of the droplets in the Murphy's dropper images in the light and dark states so as to obtain higher recognition accuracy.

4. The injection speed management method according to claim 3, wherein the data processing section (120) cuts the image from the image capturing section (110) on the basis of the contour line of the Murphy's dropper so that only the Murphy's dropper image within the contour line of the Murphy's dropper is finally retained as an analysis range, and performs monochromatization or high-contrast processing on the Murphy's dropper image so that an irrelevant background at the periphery of the Murphy's dropper image is excluded.

5. The injection speed management method according to claim 4, wherein the data processing section (120) determines whether the image from the image capturing section (110) is clear on the basis of the contour line image of the Murphy's dropper, sends an instruction to the image capturing section (110) to refocus when it is determined to be unclear, and repeats the determination process until the image is determined to be clear.

6. The injection speed management method according to claim 5, wherein when the data processing unit (120) obtains the amount of infusion solution, the type of infusion device, and the volume data of the liquid drop, the data processing unit (120) can calculate the remaining amount of infusion solution and the remaining infusion time parameter by combining the data with the drop rate.

7. The injection speed management method of claim 6, further comprising:

a central server (200) comprising at least a database and for managing patient profiles and communicating data with the infusion detection device (100),

it is characterized in that the preparation method is characterized in that,

the infusion detection device (100) further comprises a display part (130) which is connected to the data processing part (120) to obtain and display image data, dropping speed, residual medicine amount, residual infusion time and alarm information, wherein the infusion detection device (100) further comprises a data transmission part (140) which transmits infusion data and patient identification information to the central server (200), and the central server (200) detects a patient file in the database according to the patient identification information and updates the patient file according to the infusion data.

8. The injection speed management method according to claim 7, wherein when the data processing unit (120) obtains data of a designated dropping speed range, which can be provided by a scheme preset in the database of the center server (200), the data processing unit (120) can determine that the value of the dropping speed is within the dropping speed range, and if the value of the dropping speed is not within the dropping speed range, the data processing unit (120) issues an alarm instruction.

9. The injection speed management method according to claim 8, wherein when the data on the amount of infusion solution, the type of infusion device, and the volume of liquid drops are stored in the database of the center server (200) in advance, the infusion detection device (100) requests the center server (200) for the parameters through the data transmission unit (140) and calculates the parameters on the amount of remaining solution and the time of remaining infusion.