A video monitoring method for smart medical patient care

Abstract

The present application relates to the field of video monitoring, and more particularly to a video monitoring method for intelligent medical patient care, comprising: determining personnel state based on personnel proportion value and non-illness influence representation value of monitoring area, and determining whether to perform frame optimization according to personnel state; determining data uploading mode as uploading preset video frame and corresponding compensation video frame or uploading preset video frame and corresponding video paragraph according to comparison result of associated density corresponding to preset video frame and preset associated density; constructing area sequence based on stable number of nursing personnel in reference range corresponding to each monitoring area, and determining whether to perform sequence adjustment according to comparison result of stable number difference value of area sequence and preset number difference value; uploading data of finally obtained area sequence in ascending order of sequence number. The present application improves the video monitoring data processing efficiency of patient care.

Description

Technical Field

[0001] This invention relates to the field of video surveillance, and more particularly to a video surveillance method for smart healthcare patient care. Background Technology

[0002] In the field of smart healthcare, the refinement and intelligence of patient care have become the core direction for improving the quality of medical services. As a key means to monitor patients' status in real time and ensure medical safety, video surveillance technology is constantly expanding its application scenarios. From real-time vital sign monitoring in intensive care units (ICUs) to fall warnings in geriatric wards, and behavioral trajectory analysis of rehabilitation patients, video surveillance provides medical staff with objective and continuous nursing evidence by capturing visual information such as patients' physiological movements and facial expressions. This effectively reduces the pressure of manual inspections and improves the speed of emergency response.

[0003] Chinese Patent Publication No. CN108172306A discloses a cross-regional hospital monitoring system and method. The system includes: multiple hospital monitoring information modules, each including medical resource information and ward information; a monitoring module including multiple monitoring points, each equipped with a monitoring camera, used to monitor the behavior of the ward and trigger an alarm when abnormal behavior occurs; a cloud platform used to acquire information from the hospital monitoring information modules and the monitoring module, and to establish a distributed database based on the medical resource information, ward information, and ward behavior of each hospital monitoring information module; and an allocation module, which communicates with the cloud platform and is used to comprehensively analyze the information in the distributed database and allocate medical resources to the ward. It is evident that while the above technical solution utilizes centralized cloud-based data processing, which is beneficial for data management, video data in medical scenarios is characterized by strong real-time variability and large data volume. This technical solution cannot effectively preprocess the data according to the actual video data conditions, leading to the cloud platform's processing efficiency potentially failing to meet the needs of real-world scenarios. Summary of the Invention

[0004] To address this issue, the present invention provides a video monitoring method for smart medical patient care, which overcomes the problem that existing technologies cannot effectively preprocess data based on the actual video data in medical scenarios, resulting in cloud platforms having processing efficiency that is difficult to meet the needs of actual scenarios.

[0005] To achieve the above objectives, the present invention provides a video monitoring method for smart healthcare patient care, comprising:

[0006] The monitoring video corresponding to the monitoring area is extracted evenly to obtain a number of preset video frames.

[0007] The status of personnel is determined based on the proportion of personnel in the monitored area and the non-affective factor values, and whether frame optimization should be performed is determined based on the status of personnel.

[0008] The data upload method is determined based on the comparison result between the association density corresponding to the preset video frame and the preset association density. It is to upload the preset video frame and the corresponding compensation video frame or the preset video frame and the corresponding video segment.

[0009] A regional sequence is constructed based on the stable number of nursing staff within the reference range corresponding to each monitoring area, and whether to adjust the sequence is determined based on the comparison between the stable number difference value of nursing staff in the regional sequence and the preset number difference value.

[0010] The final region sequence is uploaded in ascending order of its serial number.

[0011] Furthermore, the personnel status is determined based on the personnel ratio value and the non-impact characteristic value of the monitored area. If the personnel status is that the personnel ratio value is less than the preset personnel ratio value and the non-impact characteristic value is less than the preset non-impact characteristic value, then frame optimization is not required.

[0012] If the personnel status is such that the personnel ratio is greater than or equal to the preset personnel ratio or the non-affected status is greater than or equal to the preset non-affected status, then frame optimization is performed.

[0013] Furthermore, the non-affected impact characterization value is determined based on the distribution dispersion value of non-affected personnel in the monitoring area and the difference in collection distance;

[0014] The non-affective influence characterization value, the distribution dispersion value, and the difference in collection distance are all positively correlated.

[0015] Furthermore, frame optimization includes: determining the keypoint state based on the keypoint quantity fluctuation and the average keypoint quantity;

[0016] If the key point status is that the key point quantity fluctuation is greater than the preset key point quantity fluctuation or the key point quantity average is less than or equal to the preset key point quantity average, then the corresponding compensation quantity is determined based on the key area distance corresponding to each preset video frame.

[0017] If the keypoint status is that the keypoint quantity fluctuation is less than or equal to the preset keypoint quantity fluctuation and the average keypoint quantity is greater than the preset keypoint quantity, then the corresponding compensation quantity is determined based on the keypoint quantity corresponding to each preset video frame.

[0018] Furthermore, the data upload method is determined based on the comparison result between the association density corresponding to the preset video frame and the preset association density;

[0019] If the correlation density is less than the preset correlation density, the data upload method is to upload the preset video frame and the corresponding compensation video frame.

[0020] If the association density is greater than or equal to the preset association density, the data upload method is to upload the preset video frame and the corresponding video segment.

[0021] The correlation density is determined based on the compensation video frames within a preset time period corresponding to the preset video frames.

[0022] Furthermore, the regional sequence is obtained by sorting the nursing staff in ascending order based on the stable number of nursing staff within the reference range corresponding to each monitoring area;

[0023] The area of ​​the reference range is positively correlated with the number of patients in the monitored area.

[0024] Furthermore, the comparison results of the stable number difference value of nursing staff in the detection area sequence with the preset number difference value are analyzed. When the stable number difference value of nursing staff is less than the preset number difference value, sequence adjustment is performed on the area sequence.

[0025] Furthermore, in sequence regulation, the regional sequence is divided into several subsequences, and for a single subsequence, they are sorted in descending order according to the patient monitoring demand coefficients corresponding to each monitoring area within the subsequence.

[0026] Furthermore, for the monitoring areas of the regional sequence, data is uploaded in ascending order of the serial number. When uploading data for a single monitoring area, the number of valid cloud processing nodes is detected. If the number of valid cloud processing nodes is less than the preset number of nodes, the monitoring data corresponding to the monitoring area is transmitted to the valid cloud processing node with the smallest load characterization value.

[0027] If the number of valid cloud processing nodes is greater than or equal to the preset number of nodes, the monitoring data corresponding to the monitoring area will be transmitted to the valid cloud processing node with the largest status compensation value.

[0028] An effective cloud processing node is a cloud processing node whose load characterization value is less than the preset load characterization value.

[0029] Compared with the prior art, the beneficial effects of the present invention are that the technical solution of the present invention determines the status of personnel based on the personnel ratio value and the non-patient influence characterization value in the monitoring area. The personnel status reflects the degree of possible influence of non-patients in the monitoring area on patient monitoring and analysis to determine whether frame optimization should be performed. Furthermore, frame optimization can compensate for the extracted video frames to improve the monitoring value of the video frames subsequently transmitted to the cloud processing node, thereby improving the efficiency of monitoring and nursing care.

[0030] Furthermore, in the frame optimization of the technical solution of the present invention, the key point state is determined based on the key point quantity fluctuation and the key point quantity average. The key point state reflects the stability and quantity state of the key points of the preset video frame, and different confirmation methods for the compensation quantity are selected accordingly, thereby improving the accuracy of the compensation quantity and thus improving the data processing efficiency.

[0031] Furthermore, in the technical solution of the present invention, the data upload method is determined based on the comparison result between the association density corresponding to the preset video frame and the preset association density. The selection of the data upload method is more in line with the actual application scenario, avoiding the problem that the cloud platform may have difficulty meeting the actual scenario requirements due to a single upload method.

[0032] Furthermore, in the technical solution of the present invention, the nursing staff are arranged in ascending order based on the stable number of nursing staff within the reference range corresponding to each monitoring area to obtain the regional sequence. The order is confirmed based on the distribution of nursing staff in the scene, which ensures that nursing needs are met. Moreover, when adjusting the sequence, the patient monitoring demand coefficients corresponding to each monitoring area in the sub-sequence are arranged in descending order, which further improves nursing efficiency.

[0033] Furthermore, in the technical solution of the present invention, when uploading data for a single monitoring area, the monitoring data corresponding to the monitoring area is transmitted to the effective cloud processing node with the smallest load characterization value or to the effective cloud processing node with the largest state compensation value, depending on the number of effective cloud processing nodes. This improves the stability of data processing while ensuring load stability and avoids the problem of poor monitoring efficiency on a large scale when monitoring areas that are distributed close to each other are assigned to the same cloud processing node, which would cause the cloud processing node to malfunction. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the video monitoring method for smart medical patient care according to the present invention;

[0035] Figure 2 This is a flowchart illustrating how the present invention determines whether to perform frame optimization based on personnel status;

[0036] Figure 3 This is a flowchart illustrating how the present invention determines the data upload method based on a comparison between the association density corresponding to a preset video frame and a preset association density.

[0037] Figure 4 This is a flowchart illustrating the process of determining whether to perform sequence adjustment based on the comparison between the stable number difference value of nursing staff in the regional sequence and the preset number difference value. Detailed Implementation

[0038] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0039] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0040] Please see Figures 1 to 4 As shown, the present invention provides a video monitoring method for smart medical patient care, comprising:

[0041] The monitoring video corresponding to the monitoring area is extracted evenly to obtain a number of preset video frames.

[0042] The status of personnel is determined based on the proportion of personnel in the monitored area and the non-affective factor values, and whether frame optimization should be performed is determined based on the status of personnel.

[0043] The data upload method is determined based on the comparison result between the association density corresponding to the preset video frame and the preset association density. It is to upload the preset video frame and the corresponding compensation video frame or the preset video frame and the corresponding video segment.

[0044] A regional sequence is constructed based on the stable number of nursing staff within the reference range corresponding to each monitoring area, and whether to adjust the sequence is determined based on the comparison between the stable number difference value of nursing staff in the regional sequence and the preset number difference value.

[0045] The final region sequence is uploaded in ascending order of its serial number.

[0046] This invention is applied to the monitoring data processing of patient care videos. The invention utilizes several monitoring devices, the specific locations of which are set by the user. Each monitoring device corresponds to a monitoring area, and the scene area corresponding to the monitoring screen of the monitoring device is the monitoring area. Non-patients include nursing staff and other personnel. Other personnel are those not associated with nursing staff or patients. Nursing staff and patients wear wristbands with at least positioning functionality to confirm their location and thus determine the distribution of personnel within the monitoring area. The above is readily understood by those skilled in the art and will not be elaborated upon further.

[0047] The present invention utilizes several historical records. Each historical record contains at least one historical process, including the personnel ratio, non-disease impact characterization value, key point quantity fluctuation, key point quantity average, correlation density, stable quantity difference value, number of effective cloud processing nodes, and load characterization value. The historical records also contain corresponding qualification marks, which indicate whether the historical records meet the user's needs. Users can determine whether the historical records meet their needs based on the processing effect of the monitoring data. This is a conventional technical method for those skilled in the art and will not be elaborated here.

[0048] This invention employs a continuous cyclical monitoring cycle. At the end of each monitoring cycle, a preset video frame is extracted, and frame optimization, data upload method, and regional sequence analysis are performed. The duration of the monitoring cycle is set by the user; the higher the user's requirement for monitoring and analysis accuracy, the shorter the monitoring cycle. At the end of a single monitoring cycle, the preset video frame extraction method involves uniformly extracting video frames from the monitoring video collected in that monitoring cycle. The extracted video frames are recorded as preset video frames. The number of preset video frames is set by the user; the higher the user's requirement for data analysis accuracy, the greater the number of preset video frames extracted.

[0049] Specifically, the personnel status is determined based on the personnel ratio and non-impact characteristic value of the monitored area. If the personnel status is that the personnel ratio is less than the preset personnel ratio and the non-impact characteristic value is less than the preset non-impact characteristic value, then frame optimization is not required.

[0050] If the personnel status is such that the personnel ratio is greater than or equal to the preset personnel ratio or the non-affected status is greater than or equal to the preset non-affected status, then frame optimization is performed.

[0051] Specifically, the non-affected impact characterization value is determined based on the distribution dispersion of non-affected personnel in the monitoring area and the difference in collection distance;

[0052] The non-affective influence characterization value, the distribution dispersion value, and the difference in collection distance are all positively correlated.

[0053] The personnel ratio in the monitoring area = number of patients in the monitoring area / number of non-patients in the monitoring area. For a single monitoring area, the corresponding non-patient impact characterization value = distribution discrete value / preset distribution discrete value + acquisition distance difference / preset acquisition distance difference. Wherein, the distribution discrete value is equal to the average of the sub-discrete values, and the acquisition distance difference is equal to the average of the sub-distance difference. Each preset video frame has a corresponding sub-discrete value and sub-distance difference. The sub-discrete value is the area of ​​the smallest rectangle in the preset video frame that can include all non-patients, and the sub-distance difference is the absolute value of the difference between the maximum distance and the minimum distance between non-patients and the monitoring device in the preset video frame.

[0054] Users can set preset distribution discrete values ​​and preset acquisition distance difference values ​​according to their own experience and actual needs. It can be understood that the present invention reflects the distribution of non-patients in the monitoring screen corresponding to the monitoring area through distribution discrete values ​​and acquisition distance difference. Since non-patients can affect patient care monitoring, the smaller the user's acceptance of the impact of non-patients, the smaller the preset distribution discrete value and the smaller the preset acquisition distance difference. A value setting method is provided, where the preset distribution discrete value is 50% of the screen area of ​​the preset video frame, and the preset acquisition distance difference is 35% of the maximum distance between non-patients and the monitoring device in the preset video frame.

[0055] Users can set the preset personnel ratio and preset non-patient impact characterization values ​​according to their actual application scenarios. It is understood that this invention uses the personnel ratio and non-patient impact characterization values ​​to comprehensively represent the potential impact of non-patient personnel on patient care monitoring. When the corresponding personnel status is that the personnel ratio is greater than or equal to the preset personnel ratio or the non-patient impact characterization value is greater than or equal to the preset non-patient impact characterization value, frame optimization is performed to improve the accuracy of subsequent data upload methods. Therefore, the greater the user's demand for subsequent judgment accuracy, the smaller the preset personnel ratio and preset non-patient impact characterization values ​​will be. A value setting method is provided to extract the personnel ratio and non-patient impact characterization values ​​corresponding to the historical records that meet the user's needs, remove outliers from the personnel ratio and non-patient impact characterization values ​​respectively, and record the average values ​​of the personnel ratio and non-patient impact characterization values ​​after removing outliers as the preset personnel ratio and preset non-patient impact characterization values ​​respectively.

[0056] Specifically, frame optimization includes: determining the keypoint state based on the keypoint quantity fluctuation and the average keypoint quantity;

[0057] If the key point status is that the key point quantity fluctuation is greater than the preset key point quantity fluctuation or the key point quantity average is less than or equal to the preset key point quantity average, then the corresponding compensation quantity is determined based on the key area distance corresponding to each preset video frame.

[0058] If the keypoint status is that the keypoint quantity fluctuation is less than or equal to the preset keypoint quantity fluctuation and the average keypoint quantity is greater than the preset keypoint quantity, then the corresponding compensation quantity is determined based on the keypoint quantity corresponding to each preset video frame.

[0059] Specifically, the maximum number of identifiable key points of a single patient within each preset video frame is obtained and recorded as the number of key points corresponding to the preset video frame. Key points are the joint nodes of the patient. The identification of joint nodes is performed using an OpenPose pre-trained model. Joint nodes include, but are not limited to, shoulders, elbows, and knees. The training of the model is a subject already known to those skilled in the art and will not be elaborated here.

[0060] The average number of key points is equal to the average number of the maximum number of identifiable key points of patients within each preset video frame. The fluctuation of the number of key points is denoted as S, and the formula for calculating the fluctuation of the number of key points S is:

[0061]

[0062] Where Mi is the maximum number of identifiable key points of the patient within the frame of the i-th preset video frame. Where i = 1, 2, 3, ..., N, and N is the total number of preset video frames. The value of i corresponding to the preset video frame can be randomly selected and will not affect the calculation result. As a special case, if the calculation result of S is 0, the value of S is directly set to the preset key point quantity fluctuation degree + 1.

[0063] The user can set the preset keypoint quantity fluctuation and preset keypoint quantity average value according to the actual application scenario. It can be understood that the present invention reflects the stability and quantity status of keypoints in preset video frames through keypoint quantity fluctuation and keypoint quantity average value, and selects the corresponding compensation quantity based on the key area distance or keypoint quantity corresponding to each preset video frame, thereby improving the frame optimization effect. Therefore, the higher the user's requirements for the stability and quantity of keypoints, the smaller the preset keypoint quantity fluctuation and the larger the preset keypoint quantity average value. A value selection method is provided to extract the keypoint quantity fluctuation and keypoint quantity average value corresponding to the historical records that meet the user's requirements, and remove the outliers in the keypoint quantity fluctuation and keypoint quantity average value respectively. The average values ​​of the keypoint quantity fluctuation and keypoint quantity average value after removing outliers are recorded as the preset keypoint quantity fluctuation and preset keypoint quantity average value respectively.

[0064] When determining the corresponding compensation amount based on the distance to the key area corresponding to each preset video frame, taking a single preset video frame as an example, the key area within the preset video frame is detected, and the distance to the nearest non-patient in the key area is also detected. The compensation amount corresponding to the preset video frame = basic compensation amount + distance to the nearest non-patient in the key area within the preset video frame × k1, where k1 is the first conversion coefficient. When determining the corresponding compensation amount based on the number of key points corresponding to each preset video frame, taking a single preset video frame as an example, the compensation amount corresponding to the preset video frame = basic compensation amount + number of key points within the preset video frame × k2, where k2 is the second conversion coefficient. Since the monitoring device is fixed, users can pre-set the location of sensitive areas within the monitoring area corresponding to the monitoring device as key areas. Sensitive areas are areas with high security requirements that users are concerned about, including but not limited to... Not limited to the placement of medical equipment (ventilators), medical supplies used by patients (such as IV drips), and elevator exits, the value of the basic compensation quantity can be set by the user. It is understood that the greater the user's demand for frame optimization compensation, the greater the value of the basic compensation quantity. A specific implementation value is provided: basic compensation quantity = 5. The values ​​of k1 and k2 are understood to be: the lower the user's acceptance of the distance to the nearest non-patient in the key area within the preset video frame, the greater the value of k1; the greater the user's attention to key points, the greater the value of k2. The calculation results of the distance to the nearest non-patient in the key area within the preset video frame × k1 and the number of key points within the preset video frame × k2 are both fixed to be rounded up to the nearest integer. A specific implementation value is provided: k1 = 0.4, k2 = 0.6.

[0065] For a single preset video frame, after confirming the completion of the corresponding compensation quantity, frame compensation is performed on that preset video frame. The video frame with the smallest time interval that was not included in the compensation video frame count is extracted and recorded as the compensation video frame (if the number of video frames with the smallest time interval that was not included in the compensation video frame count is greater than 1, then one of these video frames is randomly selected and recorded as the compensation video frame), until the number of extracted compensation video frames equals the compensation quantity. It is worth noting that after confirming the completion of the compensation quantity and extracting all compensation video frames, the compensation video frames and the preset video frames are collectively recorded as the preset video frame.

[0066] Specifically, the data upload method is determined based on the comparison result between the association density corresponding to the preset video frame and the preset association density;

[0067] If the correlation density is less than the preset correlation density, the data upload method is to upload the preset video frame and the corresponding compensation video frame.

[0068] If the association density is greater than or equal to the preset association density, the data upload method is to upload the preset video frame and the corresponding video segment.

[0069] The correlation density is determined based on the compensation video frames within a preset time period corresponding to the preset video frames.

[0070] The correlation density is determined by performing correlation analysis on the preset video frames corresponding to the monitored video in chronological order. For the correlation analysis of a single preset video frame, a video segment with the longest duration that meets the segment requirements is constructed starting from the preset video frame. The total number of preset video frames in the video segment is recorded as the correlation density. The segment requirement is that the time interval between any preset video frame in the video segment and its two adjacent video frames before and after it in the time sequence is less than a preset time interval. The value of the preset time interval is set by the user. The higher the user's requirement for the accuracy of the correlation density determination, the smaller the value of the preset time interval. One preset time interval is provided, which is 1 minute.

[0071] Users can set the preset correlation density value according to the actual application scenario. It can be understood that the higher the correlation density, the greater the monitoring value of the video segment. Therefore, the higher the user's demand for judging the monitoring value of the video segment, the higher the preset correlation density. One method is to extract the correlation density from the historical records that meet the user's needs, and record the average correlation density after removing outliers as the preset correlation density.

[0072] Specifically, the regional sequence is obtained by sorting the nursing staff in ascending order based on the stable number of nursing staff within the reference range corresponding to each monitoring area;

[0073] The area of ​​the reference range is positively correlated with the number of patients in the monitored area.

[0074] The area of ​​the reference range = the area of ​​the baseline range × the number of patients in the monitoring area / the preset number of patients. The reference range is a circular area constructed with the center point of the monitoring area as the center. The values ​​of the baseline range area and the preset number of patients can be set by the user. It can be understood that the higher the user's needs for patient care, the larger the baseline range area and the smaller the preset number of patients. A specific implementation value is provided: the baseline range area is 40m², and the preset number of patients = 6. The area of ​​the reference range is an integer rounded up.

[0075] The method for confirming the stable number of nursing staff within the reference range is to extract the average number of nursing staff within the reference range for several moments corresponding to the monitoring area within the most recently ended monitoring period. The number of moments is set by the user. The greater the user's attention to patient care, the greater the number of moments. A specific implementation value is provided, in which the number of moments corresponding to the monitoring area within the most recently ended monitoring period is 10.

[0076] Specifically, the comparison results of the stable number difference value of nursing staff in the detection area sequence with the preset number difference value are used. When the stable number difference value of nursing staff is less than the preset number difference value, the sequence is adjusted for the area sequence.

[0077] The stable quantity difference value of nursing staff is the absolute value of the difference between the maximum and minimum stable quantity of nursing staff in the monitoring area within the regional sequence. The preset quantity difference value is set by the user. It can be understood that the stable quantity difference value of nursing staff reflects the degree of difference in the stable quantity of nursing staff in each monitoring area within the regional sequence. Therefore, the higher the user's acceptance of this degree of difference, the larger the preset quantity difference value will be. A value selection method is provided to extract the stable quantity difference value corresponding to the historical records that meet the user's needs, and the average value of the stable quantity difference value after removing outliers is recorded as the preset quantity difference value.

[0078] Specifically, in sequence regulation, the regional sequence is divided into several subsequences, and for a single subsequence, they are sorted in descending order according to the patient monitoring demand coefficients corresponding to each monitoring area within the subsequence.

[0079] Among them, several subsequences are constructed based on the ascending order of the stable number of nursing staff within the reference range corresponding to each monitoring area. The number of subsequences is j, and the value of j is set by the user. It can be understood that the greater the user's requirement for sequence adjustment accuracy, the larger the value of j. If the number of monitoring areas in the regional sequence cannot be divided by j, the number of the last subsequence in the ascending order of the stable number of nursing staff may not be the same as the previous subsequence.

[0080] When sorting the monitoring areas within a subsequence in descending order according to the patient monitoring demand coefficients, the monitoring areas within the subsequence are reordered in descending order according to the patient monitoring demand coefficients, but the order between the subsequences remains unchanged. For a single monitoring area, the corresponding patient monitoring demand coefficient is the number of high-risk patients within the reference range of each monitoring area. High-risk patients are those with high nursing needs. Whether a patient is considered a high-risk patient is set by the user. It is understood that the user can determine whether a patient is a high-risk person based on the patient's condition and recovery level. This is content that is already known to those skilled in the art and will not be elaborated further.

[0081] Specifically, for the monitoring areas of the regional sequence, data is uploaded in ascending order of the serial number. When uploading data for a single monitoring area, the number of valid cloud processing nodes is detected. If the number of valid cloud processing nodes is less than the preset number of nodes, the monitoring data corresponding to the monitoring area is transmitted to the valid cloud processing node with the smallest load characterization value.

[0082] If the number of valid cloud processing nodes is greater than or equal to the preset number of nodes, the monitoring data corresponding to the monitoring area will be transmitted to the valid cloud processing node with the largest status compensation value.

[0083] An effective cloud processing node is a cloud processing node whose load characterization value is less than the preset load characterization value.

[0084] When uploading data, if sequence adjustment is not performed, the smaller the sequence number of the monitoring area in the regional sequence, the smaller the stable number of nursing staff. Similarly, if sequence adjustment has been performed, the sequence number of the monitoring area will be as follows. The regional sequence is the regional sequence after sequence adjustment.

[0085] The present invention utilizes several cloud processing nodes, which are cloud servers with data processing capabilities. The load characterization value is the current CPU usage percentage of the corresponding cloud processing node.

[0086] Users can set the preset number of nodes and preset load characterization values ​​according to their actual application scenarios. It can be understood that the smaller the load characterization value, the larger the number of effective cloud processing nodes, and the stronger the current data processing capability. Therefore, the greater the user's demand for data processing capability, the smaller the preset load characterization value and the larger the preset number of nodes. One method is to extract the number of effective cloud processing nodes and load characterization values ​​corresponding to the historical records that meet the user's needs, remove outliers from the number of effective cloud processing nodes and load characterization values ​​respectively, and record the average value of the number of effective cloud processing nodes and load characterization values ​​after removing outliers as the preset number of nodes and preset load characterization values ​​respectively.

[0087] For a single monitoring area, the monitoring area is designated as the target monitoring area. The method for confirming its status compensation value is to detect the average of the minimum distances between the monitoring areas assigned to the current cloud processing node and the target monitoring area, and record this as the status compensation value between the cloud processing node and the target monitoring area.

[0088] If the data upload method is to upload the preset video frames and the corresponding compensation video frames, then the monitoring data corresponding to the monitoring area is the preset video frames and the corresponding compensation video frames of the monitoring video of the most recently ended monitoring cycle.

[0089] If the data upload method is to upload preset video frames and corresponding video segments, then the monitoring data for the monitoring area is the preset video frames and corresponding video segments of the monitoring video from the most recently completed monitoring cycle.

[0090] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0091] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A video monitoring method for smart medical patient care, characterized in that, The method comprises the following steps: Uniformly extracting the monitoring video corresponding to the monitoring area to obtain a plurality of preset video frames; Determining the personnel state based on the personnel proportion value and the non-patient influence representation value of the monitoring area, and determining whether to perform frame optimization according to the personnel state; Determining the data uploading mode according to the comparison result of the correlation density corresponding to the preset video frame and the preset correlation density, wherein the data uploading mode is uploading the preset video frame and the corresponding compensation video frame or uploading the preset video frame and the corresponding video segment; Constructing a region sequence based on the stable number of nursing staff in the reference range corresponding to each monitoring area, and determining whether to perform sequence adjustment according to the comparison result of the stable number difference value of the nursing staff in the region sequence and the preset number difference value; Uploading the finally obtained region sequence in ascending order of sequence number. 2.The video monitoring method for smart medical patient care of claim 1, wherein, If the personnel state is that the personnel proportion value is less than the preset personnel proportion value and the non-patient influence representation value is less than the preset non-patient influence representation value, frame optimization is not needed. If the personnel state is that the personnel proportion value is greater than or equal to the preset personnel proportion value or the non-patient influence representation value is greater than or equal to the preset non-patient influence representation value, frame optimization is performed. 3.The video monitoring method for smart medical patient care of claim 2, wherein, The non-patient influence representation value is determined according to the distribution dispersion value of the non-patient personnel corresponding to the monitoring area and the collection distance difference degree. The non-patient influence representation value, the distribution dispersion value and the collection distance difference degree are all positively correlated. 4.The video monitoring method for smart medical patient care of claim 2, wherein, Frame optimization includes: determining the key point state based on the key point number fluctuation degree and the key point number average value; If the key point state is that the key point number fluctuation degree is greater than the preset key point number fluctuation degree or the key point number average value is less than or equal to the preset key point number average value, the compensation number corresponding to each preset video frame is determined based on the distance of the key area corresponding to the preset video frame. If the key point state is that the key point number fluctuation degree is less than or equal to the preset key point number fluctuation degree and the key point number average value is greater than the preset key point number average value, the compensation number corresponding to each preset video frame is determined based on the key point number corresponding to the preset video frame. 5.The video monitoring method for smart medical patient care of claim 4, wherein, Determining the data uploading mode according to the comparison result of the correlation density corresponding to the preset video frame and the preset correlation density; If the correlation density is less than the preset correlation density, the data uploading mode is uploading the preset video frame and the corresponding compensation video frame. If the correlation density is greater than or equal to the preset correlation density, the data uploading mode is uploading the preset video frame and the corresponding video segment. The correlation density is determined according to the compensation video frame in the preset time period corresponding to the preset video frame. 6.The video monitoring method for smart medical patient care of claim 5, wherein, The stable number of nursing staff in the reference range corresponding to each monitoring area is arranged in ascending order to obtain a region sequence. The area of the reference range and the number of patient personnel corresponding to the monitoring area are positively correlated. 7.The video monitoring method for smart medical patient care of claim 6, wherein, When the comparison result of the stable number difference value of the nursing staff in the region sequence and the preset number difference value is that the stable number difference value is less than the preset number difference value, sequence adjustment is performed on the region sequence. 8.The video monitoring method for smart medical patient care of claim 7, wherein, In sequence adjustment, the region sequence is divided to obtain a plurality of subsequences, and for a single subsequence, the patient monitoring demand coefficient corresponding to each monitoring area in the subsequence is arranged in descending order. 9.The video monitoring method for smart medical patient care of claim 8, wherein, The monitoring areas for the region sequences perform data uploading in ascending order of sequence numbers, wherein when data uploading is performed for a single monitoring area, valid cloud processing nodes are detected, if the number of the valid cloud processing nodes is less than a preset node number, monitoring data corresponding to the monitoring area is transmitted to a valid cloud processing node with a minimum load representation value; if the number of the valid cloud processing nodes is greater than or equal to the preset node number, the monitoring data corresponding to the monitoring area is transmitted to a valid cloud processing node with a maximum state compensation value; the valid cloud processing node is a cloud processing node with a load representation value less than a preset load representation value.

Images