JP2020190889A - Monitoring system for care-needing person - Google Patents

Abstract

To provide a monitoring system for a care-needing person capable of reliably and quickly detecting any abnormality of the care-needing person.SOLUTION: A monitoring system for a care-needing person performs: high resolution processing S2 that converts low-resolution image data based on temperature distribution data from an infrared array sensor unit placed in a room of the care-needing person into high-resolution image data; person detection processing S3 that detects the care-needing person and identifies a person area based on the high-pixel image data; posture/behavior detection processing S5 that detects a posture/behavior of the care-needing person based on the identified person area; posture/behavior abnormality determination processing S6 that determines whether or not the posture/behavior of the care-needing person is abnormal based on posture/behavior resulting data of the posture/behavior detection processing S5; and report processing S7 that transmits an abnormal signal to a caregiver when the posture/behavior abnormality determination processing S6 determines that the posture/behavior is abnormal.SELECTED DRAWING: Figure 3

Description

The present invention relates to a care-requiring person watching system.

Conventionally, there is a known system for watching over a care-requiring person, such as an elderly person or a sick person, who reports to the caregiver as an abnormal state if the toilet is not used for a predetermined time or longer (see, for example, Patent Document 1).

Japanese Unexamined Patent Publication No. 2001-67576

However, if the care recipient falls or has difficulty breathing, or is in an abnormal state such as rampage or wandering, it may take time to report and the caregiver's discovery may be delayed, resulting in a serious accident. It was.

Therefore, an object of the present invention is to provide a care-requiring person watching system that can reliably and quickly detect an abnormality of a care-requiring person. Furthermore, the recent labor shortage of long-term care facilities has become a serious social problem, and the purpose is to solve such a problem and reduce the burden on long-term care staff.

The care-requiring person watching system of the present invention includes high-resolution processing for converting low-pixel image data based on temperature distribution data from an infrared array sensor unit arranged in a care-requiring person's room into high-pixel image data. A person detection process that detects a person requiring care and identifies a person area based on the high pixel image data, and a posture / behavior detection process that detects the posture and behavior of a person requiring care based on the specified person area. The posture / behavior abnormality judgment processing for determining whether the posture / behavior of the care recipient is an abnormal posture / behavior state based on the posture / behavior result data of the posture / behavior detection processing, and the posture / behavior abnormality judgment processing are abnormal. It performs a notification process that sends an abnormal signal to the caregiver when it is determined to be in a posture / behavioral state.
In addition, in a situation where a plurality of care-requiring persons are occupying the same room, a database of gait characteristics is created for each of the plurality of care-requiring persons in the same room, and the gait characteristics used by looking up from the database. Authentication processing is added.

In addition, when the caregiver inputs erroneous judgment information that means that the abnormality judgment is erroneous, the judgment standard correction process for correcting the judgment standard database for judging whether or not the patient is in an abnormal posture / behavioral state is performed. is there.
Further, in the detected person area, a body temperature abnormality determination process for determining whether or not the body temperature of the person requiring nursing care is abnormal, and an abnormality signal are transmitted to the caregiver side when the body temperature abnormality determination process determines that the body temperature is abnormal. It is the one that performs the report processing and.
Further, a sensor device having the infrared array sensor unit and performing the high resolution processing, the person detection processing, the posture / behavior detection processing, and the posture / behavior result transmission processing for transmitting the posture / behavior result data. A server device that receives the posture / action result data from the sensor device and performs the posture / behavior abnormality determination process and the report process, and a caregiver information terminal that receives the abnormality signal from the server device. , Is equipped.

According to the present invention, it is possible to quickly (in real time) detect an abnormality of a care recipient and prevent a delay in reporting. Posture / behavior detection enables quick notification. In addition, since dynamic changes in posture ─── that is, behavior ─── can be detected, falling / rampaging behavior, wandering, falling, etc. from the bed can be detected reliably and quickly. Moreover, visual privacy is protected. In addition, false alarms to caregivers are reduced. In addition, the financial burden at the time of system introduction and after introduction can be reduced. Furthermore, it is possible to solve the recent problem of labor shortage in long-term care facilities and reduce the burden on long-term care staff.

It is an overall block diagram which shows one Embodiment of this invention. It is a whole block diagram which shows one Embodiment of this invention. It is a flowchart for demonstrating the watching process. It is a flowchart for demonstrating the watching process. It is a simplified image data diagram which shows an example of low pixel image data. It is a simplified image data diagram which shows an example of high pixel image data. It is explanatory drawing for demonstrating an example of a person detection process. It is explanatory drawing for demonstrating an example of a posture detection process. The figure which shows the skeleton model concretely, (A) is the figure of the skeleton model which hits the head with both hands while swinging the head left and right, (B) is the figure of the skeleton model which fell from the bed. , (C) is a diagram of a skeletal model in which a person sits on a bed and tries to get out of bed.

Hereinafter, the present invention will be described in detail based on the illustrated embodiments.
As shown in FIGS. 1 and 2, the care-requiring person watching system of the present invention is connected to a plurality of monitoring sensor devices 1 and a plurality of sensor devices 1 by wire or wirelessly so as to be able to transmit and receive data (information). The server device 2 and the portable caregiver information terminal 3 wirelessly connected to the server device 2 so as to be able to send and receive data (information) are provided, and an elderly person, a sick person, etc. existing in the vicinity of the sensor device 1 are provided. This is a monitoring process that detects anomalies in the care recipient and reports it to caregivers such as caregivers, nurses, and parents. Although not shown, there are a plurality of caregiver information terminals 3.

The sensor device 1 is arranged in a room R for a person requiring nursing care, which is provided in a nursing care welfare facility, a hospital, an elderly house, or the like.
Although not shown, the server device 2 is arranged in a management room such as a staff room, a manager's room, or a nurse center.

As shown in FIG. 2, the sensor device 1 includes a low-resolution infrared array sensor unit 10 capable of acquiring temperature distribution data, a sensor-side arithmetic processing unit 11 such as a microcomputer or a CPU for performing predetermined processing, and data. A sensor-side storage unit 12 such as a RAM or a flash memory for storing data, a sensor-side transmission unit 18 capable of transmitting data stored in the sensor-side storage unit 12 to the server device 2, and information for the server device 2 and a caregiver. It includes a sensor-side receiver 19 capable of receiving command signals and data from an external device such as a terminal 3. In FIG. 2, the sensor-side transmitting unit 18 and the sensor-side receiving unit 19 are integrally shown as a sensor-side transmitting / receiving unit.

The server device 2 includes a server-side receiving unit 29 capable of receiving data from the sensor device 1 and command signals and data from the caregiver information terminal 3, and a server-side such as a microcomputer or a CPU for performing predetermined processing. A server-side transmission capable of transmitting the data stored in the arithmetic processing unit 21, the server-side storage unit 22 such as RAM, ROM, HD (hard disk) for storing data, and the server-side storage unit 22 to the caregiver information terminal 3. It has a part 28 and. In FIG. 2, the server-side transmitting unit 28 and the server-side receiving unit 29 are integrally illustrated as a server-side transmitting / receiving unit.

The caregiver information terminal 3 receives data and signals from the server device 2 directly or via the information communication network J or the like, and a terminal-side receiving unit 39 and a server for command signals and data. It is a portable type (portable device) including a terminal-side transmitter 38 capable of transmitting data to an external device such as the device 2 or the sensor device 1.
The caregiver information terminal 3 includes a terminal-side arithmetic processing unit 31 such as a microcomputer or a CPU for performing predetermined processing, a terminal-side storage unit 32 such as a RAM, ROM, or HD (hard disk) for storing data, and a terminal. Terminals such as liquid crystal displays and organic EL displays that can display data from the side storage unit 32 and information (reports) from the server device 2. Terminals such as touch panels and buttons for inputting command signals and information. It is equipped with a side input unit 34, such as a smartphone, a tablet PC, a portable game device, and a PDA. In FIG. 2, the terminal-side transmitting unit 38 and the terminal-side receiving unit 39 are integrally shown as a terminal-side transmitting / receiving unit.

First, the flow of monitoring processing of the entire system will be described.
In FIGS. 3 and 1 and 2, the temperature distribution measurement process S1 for acquiring temperature distribution data is performed by the low-resolution infrared array sensor unit 10. Low pixel image data 51 (see FIG. 5) based on the temperature distribution data is output from the infrared array sensor unit 10.

Next, the high resolution processing S2 for converting the low pixel image data 51 into the high pixel image data 52 (see FIG. 6) is performed.
For example, the low pixel image data (low resolution image data) 51 of 80 × 60 pixels is converted into the high pixel image data (high resolution image data) 52 of 320 × 240 pixels (QVGA). In the present invention, low pixel (low resolution) means less than 160 × 120 pixels (QQVGA). High pixel (high resolution) means 160 × 120 pixels or more.

Then, the person requiring care (person) is detected using the high pixel image data 52, and the person detection process S3 for identifying the person area E (see FIG. 7) in which the person requiring care exists is performed.
When the person detection process S3 detects the humanoid silhouette, it determines that a person exists, and specifies the range of the rectangular frame in the vicinity of the humanoid silhouette as the person area E.

Then, if the person area E can be identified in the next area identification confirmation step S4, the process proceeds to the next posture / action detection process S5.
If the person area E cannot be specified, the person detection process S3 is repeated one after another until the person area E is specified for the image to be processed for high resolution.

Then, as shown in FIGS. 8 and 9, the posture / behavior detection process S5 first estimates the joint points 6 of the person requiring care (person) in the specified person area E, and each estimated joint. A skeletal model 5 is created by connecting points 6 with a straight line, and a posture (form) is detected based on the skeletal model 5, and posture / action result data is created.

Posture / action result data is, for example, "posture / action to start walking from bed", "posture / action to trip over", "posture / action to crawl on all fours", "posture / action to lie down", "on the chest". There are various types of postures such as "postures and actions of leaning against each other", "postures of falling and poking hands", "postures of lying down", "postures of sitting on the floor", and "postures of standing up and walking". FIG. 9 (A) shows the behavior of hitting the head with both hands while swinging the head from side to side in a standing posture, FIG. 9 (B) shows the posture at the moment of falling from the bed, and FIG. 9 (C) shows the posture of sitting on the bed. Show attitude and behavior.

Here, the posture / action result data is associated with the person identification information.
The person identification information may be any information as long as it is possible to identify (determine) which care-requiring person the posture / action result data belongs to, and the sensor identification information of the sensor device 1 itself and the room R of the living room R in which the sensor device 1 is installed. This is information that is a combination of one or more of the room identification information such as a number and the personal identification information such as a caregiver list number and a chart number given to each person requiring long-term care in advance. The identification information is free, such as numbers, letters, and symbols.
Also, in the case of a multi-bed room, it is desirable to use gait authentication to identify which care recipient belongs to the room.

That is, in a situation where a plurality of care-requiring persons are occupying the same room, a database of gait characteristics is created for each of the plurality of care-requiring persons in the same room, and the gait characteristics retrieved from the database are used. It is most desirable to perform the authentication process from the viewpoint of privacy protection.
Here, the "gait feature" refers to the temporal change of each part based on the skeletal model 5 illustrated in FIGS. 8 and 9 (A), that is, the swing of the shoulder, the swing of the arm, and the movement of the foot. , The cycle, etc. are tracked, the habit of walking is obtained, and the characteristics of the movement (habit) are pointed out.

Then, as shown in FIG. 3, after the posture / behavior detection process S5, whether or not the posture or behavior of the person requiring care is an abnormal posture / behavior state based on the posture / behavior result data associated with the person identification information. The posture / behavior abnormality determination process S6 is performed.

Here, the posture / behavior abnormality determination process S6 obtains the abnormal posture / behavior standard data da to which the person identification information is associated from the judgment criterion database D for determining whether or not it is abnormal (abnormal posture / behavior state). ,read out.
That is, the abnormal posture / behavior standard data da in which the person identification information matching the person identification information associated with the posture / behavior result data is associated in advance is read out.

The abnormal posture / behavioral standard data da is not limited to the posture (abnormal posture) data in which an abnormality such as "prone posture" or "crawl on all fours" is occurring, and is in the process of collapsing. The movement (abnormal behavior) and the movement of falling from the bed shown in FIG. 9 (B) (abnormal behavior) are also important.

Then, the posture / behavior abnormality determination process S6 compares the posture / behavior result data with the abnormal posture / behavior standard data da, and determines whether or not the posture / behavior abnormality is in the abnormal posture / behavior state. That is, when the posture / behavior result data and the abnormal posture / behavior standard data da match or approximate, it is determined to be abnormal.

Then, when it is determined that the posture / behavioral state is abnormal, the notification process S7 for transmitting the abnormality signal T (see FIGS. 1 and 2) to the caregiver side is performed. In other words, the abnormal posture signal T is an abnormal posture / action signal Ta, and the process of transmitting (transmitting) the abnormal posture / action signal Ta may be referred to as an abnormal posture / action report process S71. That is, the caregiver knows that the care recipient is in an abnormal posture or an abnormal behavior.

Here, the determination standard database D has, in advance, a plurality (many) persons of abnormal posture / behavior standard data da associated with the person authentication information based on the above-mentioned gait. In addition, one or a plurality of types of abnormal posture / behavior standard data da are associated with one (one person) person authentication information.
Moreover, the type (content) of the associated abnormal posture / behavioral standard data da is different for each person authentication information (person requiring nursing care).

In this way, by creating the determination criterion database D in advance for each person requiring nursing care, it is possible to make an appropriate determination (according to individuality).
For example, the person authentication information of a person requiring nursing care (for example, "Mr. A") who cannot recognize that it is difficult to walk alone due to dementia due to weak legs is described as "posture to start walking from bed".・ Abnormal posture of "behavior" ・ Behavioral standard data da is associated. Further, the person authentication information of the caregiver who can walk (for example, "Mr. B") is not associated with the abnormal posture / behavior standard data da of "posture / behavior to start walking from the bed". Therefore, when Mr. A tries to start walking from the bed, it is determined that the posture / behavioral state is abnormal and the notification process S7 is performed. However, even if Mr. B tries to start walking from the bed, it is determined that the posture / behavioral condition is abnormal. The report processing S7 is not performed. It is possible to reduce unnecessary rushes of caregivers.

Further, when the caregiver confirms the situation of the care recipient after the report and finds that the behavior or posture is not problematic (no danger), the misjudgment information is input as shown in FIG. In the confirmation step S8, when it is confirmed that the caregiver has input the erroneous judgment information indicating that the abnormality judgment (determined as an abnormal posture / behavioral state) is erroneous, the judgment criterion database D is corrected. Proceed to the determination criterion correction process S9.

The determination standard correction process S9 corrects the abnormal posture / behavior standard data da that is the determination material (reference) in the posture / behavior abnormality determination process S6.
Here, the abnormal posture / behavior standard data da has a plurality (many) patterns of the posture even if the posture and the behavior to be detected are the same type.
For example, even if the behavior (posture) to be detected is one type of "posture trying to start walking from the bed", as a pattern of the posture predicted to be that posture, "hold the handrail of the bed". There are patterns such as "posture with the upper body raised", "posture with the upper body raised and the face and upper body facing the side of the bed with one knee upright", and "posture with the upper body sitting on the side edge of the bed and trying to stand up".

Then, in the posture / behavior abnormality determination process S6, for example, the abnormal posture / behavior standard data da of "the posture / behavior of holding the handrail of the bed and raising the upper body" is used as the judgment material (reference) as the abnormal posture / behavior. Although it was judged, the care recipient confirmed that the care recipient had a habit of repeating only this action (posture) and was not actually trying to start walking from the bed (abnormality judgment was an erroneous judgment). When it is confirmed that the erroneous judgment information has been input in the erroneous judgment information input confirmation step S8, the process proceeds to the judgment standard correction process S9, and the judgment material is "grasping the handrail of the bed". The abnormal posture / behavior standard data da of "the posture / behavior that raised the upper body" is deleted, the weighting of the abnormality is changed to be low, or the normal posture / behavior standard data is changed. Do. Then, next time, even if the caregiver becomes "a posture / action in which the upper body is raised by grasping the handrail of the bed", the report processing S7 is not performed.
That is, in the next posture / behavior abnormality determination process S6, the posture / behavior abnormality determination process S6 is performed based on the corrected (updated) determination criterion database D, so that the detection accuracy is improved and an excessive report is made. It can be reduced (reliability can be improved) and the burden on the caregiver can be reduced.

Further, the arithmetic processing unit (server-side arithmetic processing unit 21) that performs the posture / behavior abnormality determination processing S6 and the determination standard correction processing S9 is configured to be processed by the AI (artificial intelligence) program. Then, machine learning is performed based on the posture / behavior result data and the judgment standard database D, and the posture / behavior abnormality judgment processing S6 and the judgment standard correction processing S9 are performed. Furthermore, deep learning (deep learning) is performed based on the attitude / behavior result data and the judgment criterion database D to infer (estimate) the abnormal precursor attitude / behavior and the abnormal occurrence posture / behavior, and to infer (estimate) the new abnormal posture / behavior. It is configured to create the reference data da by itself and perform the self-correction process of correcting the judgment standard database D by itself (adding / deleting / changing the abnormal posture / behavior standard data da). By this deep learning, false alarms (and omissions of reports) can be efficiently eliminated, the accuracy (reliability) of judgment due to the difference in each care recipient (individuality) is improved, and it is safe and secure, and the burden on the caregiver is small. Can be a system.

Further, in FIGS. 3 and 4, when the person area E can be identified, the body temperature of the person requiring care (person) is detected in the person area E, and the body temperature result data associated with the person identification information is created. The detection process S11 is performed.
Then, after the body temperature detection process S11, a body temperature abnormality determination process S12 for determining whether or not an abnormality (body temperature is in an abnormal body temperature state) is performed based on the body temperature result data associated with the person identification information.

Here, the body temperature abnormality determination process S12 reads out the abnormal body temperature reference data db associated with the person identification information from the determination reference database D for determining whether or not the body temperature is abnormal.
That is, the abnormal body temperature reference data db associated with the person identification information that matches the person identification information associated with the body temperature result data is read out.
In addition, the abnormal body temperature reference data db differs for each person identification information (person requiring nursing care).

Then, the body temperature abnormality determination process S12 compares the body temperature result data with the abnormal body temperature reference data db to determine whether or not it is abnormal (abnormal body temperature state).
When the body temperature result data and the abnormal body temperature reference data db match, it is determined to be abnormal.

Then, when it is determined that the abnormal body temperature state is determined, the notification process S7 for transmitting the abnormal signal T to the caregiver side is performed. In other words, the abnormal body temperature signal T is an abnormal body temperature signal Tb, and the process of transmitting (transmitting) the abnormal body temperature signal Tb may be referred to as an abnormal body temperature reporting process S72.
If the abnormal body temperature of the person requiring nursing care continues due to a cold, influenza, etc., the judgment standard correction process S9 is executed by inputting the erroneous judgment information, and the abnormal body temperature standard data of the judgment standard database D is executed. Modify db. In this case as well, it is desirable to correct it by deep learning of the AI (artificial intelligence) program.
In the present invention, in the human detection process S3 of the posture / behavior monitoring system described with reference to FIG. 3, the temperature distribution data from the infrared array sensor unit 10 is always utilized in the human detection process S3. There is an advantage that the detection of abnormal body temperature of a caregiver can be carried out extremely easily and reliably.

Next, the processing (operation) of each of the devices 1, 2 and 3 will be described.
In the sensor device 1, the sensor-side arithmetic processing unit 11 performs temperature distribution measurement processing S1, high-resolution processing S2, human detection processing S3, area identification confirmation step S4, and posture / behavior detection processing S5, and further, posture / behavior detection processing S5. The posture / action result transmission process for transmitting the action result data from the sensor side transmission unit 18 to the server device 2 is performed. The sensor-side storage unit 12 stores person identification information such as sensor identification information.

The server device 2 receives the posture / action result data at the server-side receiving unit 29. The server-side arithmetic processing unit 21 calls the abnormal posture / behavior standard data da from the judgment standard database D stored in the server-side storage unit 22 in advance, and performs the posture / behavior abnormality determination processing S6 to be compared with the posture / behavior result data. If it is determined to be in an abnormal posture / behavior state, an abnormality signal T (abnormal posture / behavior signal Ta) is transmitted from the server-side transmission unit 28 to the caregiver information terminal 3 Report processing S7 (abnormal posture / behavior report). Process S71) is performed. By the way, it is desirable to add an AI function for performing machine learning (deep learning) to the arithmetic processing unit 21 and the storage unit 22 of the server device 2. As the number of days of use elapses, posture / behavior analysis becomes more accurate, false alarms decrease, and abnormalities can be predicted.

The caregiver information terminal 3 receives the abnormal posture / action signal Ta on the terminal-side receiving unit 39, and the terminal-side arithmetic processing unit 31 displays the abnormal posture / behavior signal Ta on the terminal-side display unit 33. Display behavior warning information.
When the caregiver inputs erroneous determination information to the terminal-side input unit 34, the terminal-side arithmetic processing unit 31 transmits the erroneous determination information from the terminal-side transmission unit 38 to the server device 2.
When the server device 2 receives the erroneous determination information in the server-side receiving unit 29, the server-side arithmetic processing unit 21 performs the determination standard correction process S9 regarding the abnormal posture / behavior standard data da.

Further, the sensor device 1 performs the body temperature detection process S11 and the body temperature result transmission process of transmitting the body temperature result data from the sensor side transmission unit 18 to the server device 2.
The server device 2 receives the body temperature result data at the server-side receiving unit 29. The server-side arithmetic processing unit 21 calls the abnormal body temperature reference data db from the judgment standard database D stored in advance in the server-side storage unit 22 and performs the body temperature abnormality judgment process S12 to compare with the body temperature result data, and determines the abnormal body temperature state. If it is determined, the notification process S7 (abnormal body temperature notification process S72) for transmitting the abnormal signal T (abnormal body temperature signal Tb) from the server-side transmission unit 28 to the caregiver information terminal 3 is performed.

The caregiver information terminal 3 receives the abnormal body temperature signal Tb at the terminal side receiving unit 39, and the terminal side arithmetic processing unit 31 causes the terminal side display unit 33 to display the body temperature alarm information based on the abnormal body temperature signal Tb. ..
When the caregiver inputs erroneous determination information to the terminal-side input unit 34, the terminal-side arithmetic processing unit 31 transmits the erroneous determination information from the terminal-side transmission unit 38 to the server device 2.
When the server device 2 receives the erroneous determination information in the server-side receiving unit 29, the server-side arithmetic processing unit 21 performs the determination standard correction process S9 regarding the abnormal body temperature reference data db.

Further, it is desirable that the plurality of sensor devices 1 and 1 and the server device 2 are connected by power line carrier communication, which is also called PLC or PLT, using a power line. No new wiring work such as LAN wiring work is required, installation costs (installation work costs) can be suppressed, and communication is stable.
Further, since the infrared array sensor unit 10 has a low resolution (to output low pixel image data 51), it monitors (monitors) the entire room such as the living room, kitchen, corridor, bathroom, and toilet in the room of the living room (room) R.・ Even if it is detected), privacy is protected and it is easy to understand the introduction (installation).

In the present invention, the design can be changed, the sensor device 1 performs the temperature distribution measurement process S1, and the server device 2 (server-side arithmetic processing unit 21) performs the high-resolution processing S2 and the human detection process S3. The area identification confirmation step S4, the posture / behavior detection process S5, and the body temperature detection process S11 may be performed.
Further, the sensor device 1 may perform posture / behavior abnormality determination processing S6, body temperature abnormality determination processing S12, notification processing S7, determination criterion correction processing S9, and retention (memory) of the determination criterion database D (server device). 2 may be omitted). The connection method between the sensor device 1 and the server device 2 is free, and may be a wired connection via a LAN line, a wireless connection via a wireless router, a wired connection via an information communication network J such as the Internet or a host management network, or a wired connection. It may be a wireless connection. In the abnormal signal T, the abnormal posture / behavior signal Ta and the abnormal body temperature signal Tb may be data having the same contents (simply data for transmitting an abnormality). Further, in the process of displaying the alarm information (attitude / behavior alarm information and body temperature alarm information) on the terminal side display unit 33, the alarm information transmitted from the server device 2 in association with the abnormality signal T is displayed or stored on the terminal side. The alarm information stored in the unit 32 can be freely read out according to the abnormal posture / action signal Ta.

The server device 2 has a server-side display unit such as a liquid crystal monitor capable of displaying data, a server-side input unit such as a keyboard or mouse for instructing data search and display commands, and stored data (information). ) And a server-side output unit such as a printer for outputting the display contents on paper or the like, and an information processing device such as a server-type personal computer (server PC) or a server-type workstation may be used. Then, the erroneous determination information may be input in the input unit on the server side. It is also possible to display posture / behavior warning information and body temperature warning information on the server side display unit.

In addition, the server-side arithmetic processing unit 21 of the server device 2 creates the life behavior record table data of the care recipient based on the detection result data such as the posture / action result data and the body temperature result data from the sensor device 1. good. The date and time data is associated with the detection result data produced by the sensor device 1.
The life behavior record table data is a table (including a graph) in which the life behavior record based on the posture / behavior result data and the temperature measurement record based on the body temperature result data are associated with the date and time data. For example, the number of turns, the number of steps or walking distance, the wake-up time, the bedtime, the number of times the toilet is used, the body temperature at predetermined time intervals, and the like. It is possible to reduce the labor and time required for the caregiver to create a care (nursing) record, and also to reduce the caregiver's patrol work such as hourly body temperature measurement. Furthermore, it is possible to obtain more accurate records than the self-report of the care recipient.
Further, the created life behavior record table data is transmitted from the server device 2 to the caregiver information terminal 3 and displayed on the terminal side display unit 33. Alternatively, it is displayed on the server side display unit and output (printed) on paper from the server side output unit.

The person requiring nursing care is not limited to the elderly and the sick, but may be a person who has difficulty in daily life alone.
The caregiver may be a care worker such as a caregiver, a medical worker such as a nurse or a doctor, a guard, a family member, or any other person who can assist the daily life of the care recipient. Living room R is not limited to nursing care welfare facilities, hospitals, housing for the elderly, etc., but may be a general apartment building, or even one room in a building where multiple people requiring nursing care live. Just do it.

As described above, the care-requiring person watching system of the present invention converts the low-pixel image data 51 based on the temperature distribution data from the infrared array sensor unit 10 arranged in the care-requiring person's room into the high-pixel image data 52. The high resolution processing S2 to be converted, the person detection process S3 that detects the person requiring care based on the high pixel image data 52 and identifies the person area E, and the person requiring care based on the specified person area E. Posture / behavior detection process S5 that detects posture / behavior and posture / behavior that determines whether the posture / behavior of the person requiring care is an abnormal posture / behavior state based on the posture / behavior result data of the posture / behavior detection process S5. Since the behavior abnormality determination processing S6 and the notification processing S7 for transmitting an abnormality signal T to the caregiver side when the posture / behavior abnormality determination processing S6 determines that the posture / behavior abnormality is determined, the abnormality of the person requiring care is performed. Can be detected quickly (in real time) and delays in reporting can be prevented. Posture / behavior detection enables predictive reporting (forecasting) and prompt reporting immediately after an abnormality occurs by detecting immediately before an abnormal state occurs. Excellent privacy protection for caregivers. The financial burden at the time of system introduction and after introduction can be reduced. Further, since the posture / behavior abnormality determination process S6 is performed based on the specified person area E, the processing capacity (processing range) is small, and the processing speed can be improved and the processing load can be reduced.

In addition, in a situation where a plurality of long-term care recipients are occupying the same room, a database of gait characteristics is created for each of the plurality of long-term care recipients in the same room, and the gait characteristics retrieved from the database are used. Since the authentication process is added, it is possible to identify each person (personal authentication) while protecting the visual privacy of each person requiring long-term care, which can be said to be a desirable monitoring system for each person requiring long-term care. In particular, as shown in FIGS. 9 (A), (B) and (C), the risk of falls, wandering, etc. differs for each person requiring care, or there is a risk of violence or (pathological) seizures. Even those who are at risk of having a seizure can be reliably protected without omission.

Further, when the caregiver inputs erroneous judgment information meaning that the abnormality judgment is erroneous, the judgment standard correction process S9 for correcting the judgment standard database D for determining whether or not the patient is in an abnormal posture / behavioral state is performed. Therefore, it is possible to reduce erroneous judgment (misinformation) and reduce the burden on the caregiver. It is possible to issue highly reliable (highly accurate) reports.

Further, in the detected person area E, the body temperature abnormality determination process S12 for determining whether or not the body temperature of the person requiring long-term care is abnormal, and the abnormality signal T for nursing care when the body temperature abnormality determination process S12 determines that the body temperature is abnormal. Since the notification process S7 to be transmitted to the person is performed, the caregiver can know the fever that the care recipient is not aware of, and can realize a quick response. In addition, the body temperature result can be obtained without visiting the living room R on a regular basis, the burden on the caregiver can be greatly reduced, and the care facility or the like can be operated with a small number of personnel.

In addition to having the infrared array sensor unit 10, the high-resolution processing S2, the person detection processing S3, the posture / behavior detection processing S5, and the posture / behavior result transmission processing for transmitting the posture / behavior result data are performed. The sensor device 1, the server device 2 that receives the posture / action result data from the sensor device 1 and performs the posture / behavior abnormality determination process S6 and the report process S7, and the abnormality from the server device 2. Since the information terminal 3 for a caregiver that receives the signal T is provided, a plurality (many) required sensor devices 1 can be easily manufactured (manufactured with inexpensive parts), and the processing load of the server device 2 can be reduced. Therefore, the processing speed of the entire system can be increased.

1 Sensor device 2 Server device 3 Information terminal for caregivers
10 Infrared array sensor
51 Low pixel image data
52 High pixel image data D Judgment standard database E Person area S2 High resolution processing S3 Person detection processing S5 Posture / behavior detection processing S6 Posture / behavior abnormality judgment processing S7 Notification processing S9 Judgment standard correction processing S12 Body temperature abnormality judgment processing T Abnormal signal

Claims (5)

High resolution processing (S2) that converts low pixel image data (51) based on temperature distribution data from the infrared array sensor unit (10) arranged in the room of the care recipient into high pixel image data (52). A person detection process (S3) that detects a person requiring long-term care based on the high-pixel image data (52) and identifies a person area (E), and a person requiring long-term care based on the specified person area (E). Whether or not the posture / behavior of the care recipient is an abnormal posture / behavior state based on the posture / behavior detection process (S5) that detects the posture / behavior of the person and the posture / behavior result data of the posture / behavior detection process (S5). Posture / behavior abnormality determination processing (S6) for determining whether or not, and notification processing (T) for transmitting an abnormality signal (T) to the caregiver side when the posture / behavior abnormality determination processing (S6) determines that the posture / behavior abnormality is abnormal (S6). S7) and a care-requiring person watching system characterized by performing. In a situation where multiple long-term care recipients are occupying the same room, a database of gait characteristics is created for each of the multiple long-term care recipients in the same room, and a gait authentication process using the gait features looked up from the database. The care-requiring person watching system according to claim 1, to which is added. Judgment standard correction processing (S9) for modifying the judgment standard database (D) for determining whether or not the patient is in an abnormal posture / behavior state when erroneous judgment information indicating that the abnormality judgment is incorrect is input by the caregiver. The care-requiring person watching system according to claim 1 or 2. In the detected person area (E), an abnormality occurs when the body temperature abnormality determination process (S12) for determining whether or not the body temperature of the person requiring long-term care is abnormal and the body temperature abnormality determination process (S12) determines that the body temperature is abnormal. The care-requiring person watching system according to claim 1, 2 or 3, which performs a notification process (S7) for transmitting a signal (T) to the caregiver side. It has the infrared array sensor unit (10), and also transmits the high resolution process (S2), the person detection process (S3), the attitude / action detection process (S5), and the attitude / action result data. A sensor device (1) that performs result transmission processing and
A server device (2) that receives the posture / action result data from the sensor device (1) and performs the posture / behavior abnormality determination process (S6) and the report process (S7).
The caregiver information terminal (3) that receives the abnormal signal (T) from the server device (2), and
The care-requiring person watching system according to claim 1, 2, 3 or 4.