JP6150207B2 - Monitoring system - Google Patents

Description

  The present invention relates to a system for detecting and monitoring human movement. In particular, the present invention relates to a monitoring system for capturing a person's movement on a bed and preventing the fall from the bed.

  In serious accidents at medical institutions such as hospitals, falling accidents around the bed account for a considerable percentage. Medical institutions have introduced various sensors to prevent accidents and are focusing on accident prevention, but the number of serious accidents has not decreased significantly.

  Many accident prevention sensors have been put into practical use. For example, a pressure sensor is provided on the bed to detect that the patient has woken up, a pressure sensor is provided on the floor beside the bed to detect that the patient has fallen, or a patient on the edge of the bed There is something like an infrared line sensor that detects when you are.

  A system for watching a patient using a multifocal CCD camera or the like as a detection means has also been proposed. In this system, when comparing the detection means capable of detecting the state of the subject in the room as information on the plane position and the height direction, and comparing the height direction information obtained by the detection means, the room is a bed, The area is divided into bedside and floor areas, and compared with the reference height set for each area. Based on the comparison result, the other person's situation is judged in each area. The status of the subject obtained from the determination result is reported to a nurse or the like via reporting means (for example, Patent Document 1).

JP2011-86286A

  However, in a system in which a pressure sensor is provided on the bed to detect that the patient has woken up, erroneous detection such as the sensor reacting even when turning over or changing the position has occurred. In the first place, it was unclear whether just raising the body from the bed would lead to dangerous behaviors such as the patient dropping from the bed afterwards, and it was not possible to detect only behaviors that would directly lead to dangerous behavior.

A device that detects that a patient has fallen by providing a pressure sensor on the floor beside the bed could not prevent the fall even if the fall of the patient could be known.
Further, contact type sensors such as pressure sensors tend to cause deterioration or failure of the patient's body weight or a heavy object such as a wheelchair due to the load. Furthermore, once detected, it could not be detected until it was reset, and there was a danger that the sensor would not work due to forgetting the reset and a serious accident would occur.

  An infrared line sensor that detects the presence of a patient on the edge of a bed can be detected in a non-contact manner, and has the advantage that there are few failures. However, since a futon is also detected, there are many false detections. In addition, the patient may learn the infrared position and try to get out of the bed so as to avoid the position. Not only the patient's dangerous behavior cannot be detected, but also the patient's dangerous behavior may be promoted.

  In the system disclosed in Patent Document 1, for example, when an object exceeding a predetermined ratio with respect to the bed area is detected on a plane at a predetermined height from the bed, an alarm is sounded. Therefore, although there is an advantage that false detection is reduced, false detection still occurs when the patient flips up the futon. In addition, there is a problem in that erroneous detection also occurs when light and darkness occurs at a measurement location due to the incidence of sunlight or the like.

  If false alarms occur frequently in this way, even if an alarm sounds, the nurse may question the authenticity of the alarm. is there.

  The present invention has been made to solve the above-described problems, and provides a monitoring system that makes it possible to minimize the number of false detections as much as possible and to reliably capture a patient's dangerous behavior.

  The monitoring system according to the present invention is a monitoring system mainly used in a hospital room provided with three-dimensional detection means and dangerous behavior determination means, and the three-dimensional detection means determines the distance to the detected object for each pixel. Infrared pulse irradiating means for intermittently irradiating infrared rays, a camera having a detection unit in which pixels for detecting infrared rays are two-dimensionally arranged, and infrared rays irradiated by the infrared pulse irradiating means A distance deriving unit capable of deriving the distance to the object to be detected for each pixel by measuring the time until the pulse returns to the detection unit of the camera, A dangerous area setting means for setting a dangerous area as a three-dimensional area and setting which part of the body part of the human body is determined to be dangerous in the dangerous area; Detecting hand The object is more likely to be a part of the human body than the two-dimensional shape or the three-dimensional shape of the object detected in the dangerous area derived from the distance for each pixel detected by the And an object determination means for determining which part of the human body the object is likely to be, and determined that the object determination means is highly likely to correspond. It is characterized in that it is determined that the action is dangerous when the part of the person's body matches the part of the person's body set by the dangerous area setting means.

  In particular, the object determination means is such that the size of the object detected in the dangerous area derived from the distance for each pixel detected by the three-dimensional detection means and the dangerous area is greater than or equal to a predetermined size. Determining whether or not the object is likely to be a part of a person's body and determining which part of the person's body is likely to be the object. .

  Further, it is determined whether or not an object detected in the dangerous area derived from the distance for each pixel detected by the three-dimensional detection means and the dangerous area is a part of a human body, and When determining which part of the human body the object is, the determination is made by comparing with a list of a three-dimensional shape of a part of the human body provided in advance.

  Furthermore, it is further characterized in that it is determined whether or not the action is dangerous in consideration of from which direction the object detected in the dangerous area enters the dangerous area.

  In addition, it is characterized in that it is determined whether or not the action is dangerous in consideration of the moving speed of the object detected in the dangerous area.

  Further, the alarm is activated when the dangerous behavior determining means determines that the behavior is dangerous.

  Further, when the dangerous behavior determining means determines that the behavior is dangerous behavior, an alarm that is different depending on a part of the human body detected in the dangerous area is activated.

Further, an image forming unit for deriving the position of the detection object for each pixel from the distance for each pixel detected by the three-dimensional detection unit and forming an image;
Communication means for transmitting the formed image as electronic data;
Display means for receiving the electronic data transmitted by the communication means and displaying it as an image;
Is further provided.

  The monitoring system according to the present invention is configured as described above, and the three-dimensional shape of the object can be obtained accurately and in real time. Therefore, by comparing the features of the two-dimensional shape or the three-dimensional shape, It is possible to determine with high accuracy whether or not there is a high possibility. Therefore, it is possible not only to greatly reduce false detection but also to determine in more detail whether or not the behavior is a sign that leads to a serious accident such as a fall. Furthermore, it is possible to improve the accuracy in consideration of the risk behavior pattern of hospitalized patients.

Even at night, monitoring can be performed without lighting.
Furthermore, it is not affected by the background of futons.
Moreover, since it is non-contact detection, there is no problem in durability.

Furthermore, an optimal monitoring situation can be obtained by a simple setting change in accordance with the arrangement of rooms and beds or the characteristics of the monitored person.
Having such many excellent features not only protects inpatients but also produces effects such as greatly reducing the mental and physical burdens of nurses.

It is an example of arrangement | positioning of the three-dimensional detection means which comprises the monitoring system which concerns on this invention. It is principle explanatory drawing of the three-dimensional distance image camera used with the monitoring system which concerns on this invention. It is an example of the danger area | region of the monitoring system which concerns on this invention. It is a block diagram of the monitoring system which concerns on this invention. It is a flowchart of the danger determination in the monitoring system which concerns on this invention. It is a flowchart of the danger determination in the monitoring system which concerns on this invention. 5 is a flowchart for determining whether or not a detected object is highly likely to be a predetermined part of the body in the monitoring system according to the present invention. It is an example of the danger area | region of the monitoring system which concerns on Embodiment 2 of this invention.

Embodiment 1 FIG.
A configuration of a monitoring system according to the present invention and a method of using the same will be described below with reference to the drawings. The following description discloses a good example relating to the present invention, and the present invention is not limited to the embodiment.

First, the configuration of the monitoring system will be described with reference to the drawings.
The monitoring system mainly includes a three-dimensional detection unit 1, a dangerous behavior determination unit 5, and the like. The dangerous behavior determination means 5 will be described later. First, the arrangement example of the three-dimensional detection unit 1 will be described with reference to FIG.

  FIG. 1 shows a hospital room or the like in which a bed 2 is placed at a corner of the room and a monitored person such as a patient is sleeping on the bed. Not only hospital hospital rooms, but also nursing homes and homes with home caregivers. The three-dimensional detection means 1 is provided at a position higher than the bed 2 so that the action of the monitored person on and near the bed 2 can be imaged. For example, it is installed on the top of a wall or furniture.

As shown in FIG. 2, the three-dimensional detection unit 1 includes an infrared pulse irradiation unit 11, a camera 12, and a distance deriving unit (not shown).
The infrared pulse irradiation unit 11 is an infrared pulse light source that irradiates infrared light intermittently at an extremely short time interval.

The camera 12 is an imaging device in which pixels are arranged two-dimensionally, such as a CMOS image sensor, and has sufficient sensitivity to infrared rays emitted by the infrared pulse irradiation unit. Here, a CMOS image sensor having about 20,000 pixels was used.
As shown in FIG. 2, the distance deriving unit measures the time until the pulsed light irradiated by the infrared pulse irradiation unit 11 is reflected from the subject and returns to the camera 12 for each pixel. Deriving the distance. As a measurement method, for example, the difference between the phase at the time of emission of the pulsed light and the phase at the time of feedback is measured, and the time and distance are derived from the phase difference and the light velocity.

By using such a measurement method, it is possible to obtain distance information of about 20,000 pixels in real time. Thereby, a captured image including distance information can be output at a frame rate of 10 to 15 frames per second. That is, it is possible to output a three-dimensional captured image at high speed.
Further, the distance to the subject can be measured with high accuracy of about 1 cm.

Note that a stereoscopic image can be obtained by an imaging method called stereo vision, that is, a method of capturing images from different angles with a plurality of cameras, but a lot of computation is required to construct a stereoscopic image from a plurality of planar images. Real-time processing is impossible. Also, the distance accuracy is quite bad. It is also susceptible to background effects.
In addition, it is possible to obtain images at a plurality of distances simultaneously using a multifocal CCD camera, but it is not possible to obtain a stereoscopic image only by obtaining images on a plurality of different surfaces. Further, since an image of the defocus plane is also taken, the distance to the subject includes a large error and at the same time is strongly influenced by the background.

  As described above, the major reason for using the three-dimensional detection means 1 is that a three-dimensional image can be obtained by real-time processing and that the measurement distance accuracy is excellent. However, there is an advantage that it can be used in the dark and is not affected by sunlight or shadow even in the daytime, and is suitable for imaging in a hospital room. The influence of the background of futons can be completely ignored.

  Using this three-dimensional detection means 1, the behavior of the monitored person on and near the bed 2 can be imaged and constantly monitored. At that time, a plurality of dangerous areas are provided in order to detect the dangerous behavior of the monitored person. The dangerous area is, for example, a three-dimensional area such as 4A, 4B, and 4C indicated by dotted lines in FIG. 4A and 4B are predetermined areas on the edge of the bed. If a person to be monitored enters this area, there is a risk of falling off the bed. Reference numeral 4C denotes an area of the floor right next to the bed. When the body of the monitored person is present in this area, the area can be determined that the monitored person has fallen from the bed. Further, when only the foot of the monitored person is in the 4C area, it can be determined that the monitored person has got out of the bed.

In addition, the accuracy of predicting the dangerous behavior of the monitored person can be improved by collecting behavior data of inpatients in advance and analyzing the behavior that is a precursor to causing dangerous behavior such as falling.
For example, if there is a monitored person's hand in the 4A area, it is not judged as dangerous behavior, but if there is a head or foot, it is judged as dangerous behavior, so only when there is a possibility of falling etc. It can be judged as dangerous behavior.

Next, the overall configuration of the monitoring system will be described with reference to FIG.
In addition to the three-dimensional detection unit 1, the monitoring system includes a dangerous behavior determination unit 5. Moreover, the report means 6 and the monitor means 7 are provided as needed.

The dangerous behavior determination unit 5 is a unit that determines whether or not the monitored person is performing a dangerous behavior based on the imaging information of the three-dimensional detection unit 1, and mainly by the cooperation between the computer and the processing software. It is a means for data processing and various judgments. A dedicated processor can be used instead of a computer and processing software.
The dangerous behavior determination unit 5 includes processing units such as a dangerous area setting unit 51, a coordinate conversion unit 52, an object determination unit 53, and a final determination unit 54.

  The dangerous area setting means 51 defines a dangerous area that is a three-dimensional area as shown in FIG. 3 using XYZ coordinates or the like. Although there may be only one dangerous area, as shown in FIG. 3, the plurality of areas improves the accuracy of predicting an action that leads to the danger of the monitored person. The shape of the dangerous area may be a rectangular parallelepiped or a shape including a curved surface. What is necessary is just to determine an optimal shape and position suitably according to the structure of a bed, the characteristic of a to-be-monitored person, etc.

  Further, the dangerous area setting unit 51 sets not only the dangerous area which is a three-dimensional area but also which part of the human body part is determined to be dangerous when it enters the dangerous area. For example, in the 4A area of FIG. 3, as described above, when the head or the foot enters, it is set as dangerous action.

  The coordinate conversion means 52 obtains actual coordinates (X, Y, Z) by coordinate conversion from the distance l between the pixel position (x, y) of the camera 12 and the subject measured at the pixel, and the partial position of the subject. Express. By converting into real coordinates in this way, the position and three-dimensional shape of the subject on the three-dimensional coordinate system can be obtained. Further, the actual size of the object can be derived by a simple calculation process.

The object determination means 53 derives the three-dimensional shape of the object in the dangerous area from the distance for each pixel detected by the three-dimensional detection means 1 and the dangerous area, and this object is a part of the human body. It has a function of determining whether or not the possibility is high and determining which part of the human body the object is likely to be.
The object determination unit 53 may have any configuration as long as it satisfies the above-described functions. In the present embodiment, as shown in FIG. 4, the size determination unit 531, the body part assumption unit 532, and It consists of each means of body part recognition means 533. Specific operations of these means will be described later.

  When the object determining unit 53 determines that there is a high possibility that the object detected in the dangerous area is a part of the human body, the final determining unit 54 determines that the human body part is It is determined whether or not it matches the set body part of the person. If it is determined that they match, a report is made to inform the nurse or caregiver of the abnormality.

This notification is made by the reporting means 6.
In the present embodiment, the reporting unit 6 includes an alarm activation unit 61, an alarm unit 62, a status confirmation unit 63, and an alarm stop unit 64.
The alarm activation means 61 has a role of activating the alarm 62 when an electrical signal indicating that an abnormality should be notified is obtained from the final determination means 54. For example, if the alarm 62 is an alarm sounding at the nurse's station, the alarm activation means 61 is a means for switching on the alarm. If the alarm 62 is an e-mail transmission device that transmits an e-mail notifying that an abnormality has occurred in the nurse's mobile phone, the alarm activation means 61 is software of a personal computer that controls the transmission.

If the alarm 62 is an alarm sounding at the nurse's station, an alarm stop means 64 is required to stop it, for example, an alarm stop switch provided on the wall.
Further, the alarm activation means 61 may activate the status confirmation means 63 simultaneously with the activation of the alarm means 62. The situation confirmation unit 63 is, for example, a display unit 73 such as a monitor that displays a captured image of the three-dimensional detection unit 1 described later. In hospitals, many patients are often monitored simultaneously by multiple monitoring systems. Therefore, when a dangerous action is performed, the image of the patient may be immediately displayed on the display means 73 such as a monitor placed in a nurse's office or the like. The nurse can immediately confirm the patient's situation by the display means 73 and can quickly determine what to do. If the alarm stop means 64 is provided in the display means 73, the nurse can stop the alarm by clicking the mouse after confirming the situation.

Next, the monitoring means 7 will be described.
The monitor unit 7 is a device that displays a moving image captured by the three-dimensional detection unit 1.
The imaging data of the three-dimensional detection unit 1 is converted by the image forming unit 71 into data that can be displayed on the display unit 73 such as a display monitor. This displayable data is transmitted from the image forming unit 71 to the display unit 73 by wired or wireless method.

Since the moving image picked up by the three-dimensional detection unit 1 is a three-dimensional image, it can be displayed on the display unit 73 in the same manner as the 3D television.
Or, more simply, the distance information may be converted into a color to display a pseudo three-dimensional image. For example, a portion close to the distance may be represented by a warm color such as red, and the color may be changed in order to a cold color such as blue as the distance increases. In this way, it is also possible to express the depth by color.

If the depth is expressed by color, a display such as a smartphone, a mobile phone, and a tablet PC can also be used as the display unit 73.
Nurses are not always in the stalls, and they spend a lot of time looking around hospital rooms. Therefore, if an image of the person being monitored is displayed on a portable device, a quick action can be performed even during a patrol of the hospital room.

Next, the operation of the monitoring system will be described using the flowcharts of FIGS.
First, the monitoring system is activated, and the three-dimensional detection unit 1 starts imaging. The captured three-dimensional video data is transmitted to the coordinate conversion means 52 every 1/10 second, and converted into real space coordinates (X, Y, Z) for each pixel. This is the imaging data processing (step A) in FIG. The conversion data for each pixel is sent to the object determination means 53.

  In the object determination unit 53, first, the size determination unit 531 determines whether or not an object exists in each danger area (step B). For details of step B, as shown in FIG. 6, it is first determined whether or not there is an object of a predetermined size or larger in the dangerous area (step B1). This determination is not always necessary, but the efficiency of the process can be improved by performing the subsequent determination process only when necessary. For example, even if an insect comes into the danger area, it is not determined as an object, and it is determined that an object exists only when an object of about 5 cm or more is detected.

This determination is performed by determining whether or not the real space coordinates (X, Y, Z) converted for each pixel are within the respective dangerous areas, and the length of the real space coordinates of the consecutive pixels determined as such. What is necessary is just to perform the process of calculating | requiring thickness.
Then, only when it is determined that there is an object of a predetermined size or more in the dangerous area, it is determined that the object has been detected, and the process proceeds to the next determination step C. As described above, this step B is an optional screening step that is not necessarily required. Therefore, if this step B is not performed, the process may proceed directly from step A to step C.

Next, it is determined whether or not this object is likely to correspond to a human body part (step C).
For details of step C, as shown in FIG. 7, first, a part that may correspond to the size of the object is specified (step C1). For example, the size of each part of the human body is listed in advance, and the determination can be performed by referring to the list. Here, the size may be a two-dimensional area or a three-dimensional surface area. For example, it is determined in advance that the head is 200 square centimeters to 500 square centimeters, and if an object within the range is detected, it is determined that an object that may be the head is detected.
Alternatively, the initial screening may be performed from a simple one-dimensional length. For example, the region and its length range may be listed, such as 3 to 10 cm for fingers of a hand, 10 to 20 cm for a hand, and 15 to 25 cm for a head. In this way, candidate body parts that can correspond to the detected object are determined.
Alternatively, in order to perform screening more precisely, the determination may be made based on the volume of the object.

  Then, it is determined whether the two-dimensional shape or the three-dimensional shape features of the body part and the object that are candidates match (step C2). Also in this determination, it is possible to make a determination by listing a two-dimensional shape or a three-dimensional shape of each part of a human body and referring to the list. For example, in the case of a hand, the three-dimensional shape of the hand viewed from various angles is registered in a list, and the features of the three-dimensional shape of the hand viewed from all angles are compared. Or not. Alternatively, if it is determined whether or not the head is a two-dimensional shape, it can be distinguished from arms and legs if the aspect ratio is set to 1/3 or more and 3.3 or less. It can also be distinguished from a rod-like object that is not a part of.

  As described above, when it is determined that the object is likely to be a hand, the final determination unit 54 determines whether it is a sign that the monitored person performs a dangerous action (step E). This determination is made as a sign of dangerous action when a hand is included in the body part set by the dangerous area recognition means, and is transmitted to the alarm activation means 61 by an electric signal, and finally. The alarm means 62 performs a reporting action such as sounding an alarm (step F).

The above is the configuration example and operation example of the monitoring system according to the present invention. The above are configuration examples and operation examples, and the present invention is not necessarily limited thereto.
For example, in step B1, whether or not a part of an object detected by a pixel is in a dangerous area is compared in real coordinates. Alternatively, the real coordinate range of the dangerous area may be converted back to the range of the pixel and the measurement distance (x, y, l) and compared.

  Furthermore, in steps C1 to C3, the coincidence with all the parts of the body corresponding to the object is examined, but it may be examined only whether or not the danger region 51 is a specific part set. In this way, further efficiency can be achieved, and the final determination means 54 is not necessary.

  The dangerous area 51 is set to be a dangerous action when one part such as a foot or a head enters the dangerous area. However, when a plurality of parts, for example, a hand and a head enter the dangerous area at the same time. You may set it as dangerous action.

  Moreover, it is possible to rank the dangerous behavior and change the type of alarm accordingly. For example, when a foot enters the danger area, if the foot height is at a low position, it is judged that the danger is small, an alarm with a long cycle is sounded, and the foot height position is at a high position. In some cases, it may be judged that the danger is great, and a vigorous alarm with a short cycle may be sounded. Of course, since the risk varies depending on the combination of the dangerous area and the body part, the type of alarm may be changed according to the risk.

  Furthermore, since this monitoring system is constantly monitoring, it can also detect the moving direction and moving speed of each part of the body. Therefore, it is possible to add a judgment condition that each part of the body may be dangerous when it moves from inside the bed, and conversely, it is not dangerous when it moves from outside the bed to inside the bed. It is. By doing in this way, when a caregiver puts a hand in a bed, it can prevent starting an alarm. Further, when the movement speed of the person or foot of the monitored person is abnormally fast, there is a possibility that convulsion or the like may occur, and an alarm may be activated even if it is not in the dangerous area.

The features of the monitoring system according to the present invention will be summarized below.
First, the major difference from the conventional monitoring system is that the three-dimensional shape of the object can be obtained accurately and in real time by using infrared pulsed light and an infrared camera. This provides the following specific advantages.

First, not only the shape of the object but also the distance from the camera can be measured, so in addition to the 2D shape and 3D shape of the object you want to measure, the actual length, area, surface area, or volume can be derived with a simple calculation. . Therefore, the screening accuracy of whether or not the object is likely to correspond to a body part is first improved by the area or the like.
Further, as a next step, it is possible to determine with high accuracy whether or not there is a high possibility of being a part of a human body by comparing features of two-dimensional or three-dimensional shapes. As a result, even if an object other than a body part, for example, a futon, a pillow, a person to be monitored, or the like enters the dangerous area, the possibility of erroneously determining that the action is dangerous is greatly reduced. For example, when the monitored person greatly kicks up the futon, even if the dangerous area is set to a predetermined height from the bed, it may be erroneously detected by a system other than the monitoring system. On the other hand, in this system, since the characteristics of the three-dimensional shape of the futon are different from the characteristics of the body part, there is basically no false detection.

  In addition, since this monitoring system can recognize which part of a person's body is likely to be, it is possible to determine whether the action is a precursor to a fall or the like in consideration of the risk behavior pattern of an inpatient. Judge in detail. That is, an alarm can be activated only when it is really necessary. Depending on the behavior of the monitored person, it is possible to rank the degree of risk.

Moreover, since infrared rays are used, illumination is unnecessary for monitoring even at night.
Furthermore, when an ordinary two-dimensional imaging camera is used, a defocused object also affects imaging. For example, since the color and pattern of the futon used as a background are reflected, it causes a false detection. On the other hand, in this monitoring system, since the three-dimensional position can be specified completely, it is not affected at all by the background futon.

  Since monitoring is always performed and processing can be performed in real time, the direction and speed of movement of the body part of the monitored person can be recognized, and these can be added to the condition for determining dangerous behavior. In this way, erroneous detection can be further prevented.

Moreover, since it is a non-contact system, durability is excellent.
Furthermore, since the setting of the dangerous area can be freely changed, it is possible to obtain an optimum monitoring situation according to the arrangement of the room, the bed, etc., the characteristics of the monitored person, or the like.

  As described above, this monitoring system has many advantages. What is particularly important is that it is possible to detect in advance a behavior that is a precursor of a monitored person leading to a serious accident such as a fall accident and to notify a nurse or the like. Even with conventional systems, it may have been possible to predict behavior to some extent, but over-detection that detects safe behavior as dangerous behavior frequently occurs, which puts a burden on nurses etc. There was a problem that proper actions of nurses were delayed when there were dangerous actions.

On the other hand, with this monitoring system, detailed monitoring and judgment based on analysis of risk behaviors of patients can be performed, so that over-detection can be minimized and signs that lead to real danger can be captured quickly. be able to.
This brings about the following merits for patients, nurses, hospitals and the like.

First, the risk of falling or the like is reduced for the patient.
In addition, since the risk of falling or the like is reduced, it is not necessary to take measures such as being restrained by the bed.

For nurses, they can work with peace of mind, reducing their mental burden. Moreover, since the excessive detection is reduced, a false alarm is not activated, and the burden on the body, such as frequent check-ups, is reduced.
Furthermore, since the 3D image can be confirmed on the monitor, the necessary measures can be performed accurately and quickly.

  For medical institutions such as hospitals, the risk of liability issues is reduced by preventing serious accidents. In addition, since the mental and physical burden on the nurse is reduced, nurses and doctors can work without stress, enabling efficient treatment and medical care as a whole hospital.

Embodiment 2. FIG.
A monitoring system according to Embodiment 2 of the present invention will be described with reference to FIG.
The difference from the first embodiment is that the setting of the dangerous area is complicated in order to improve the detection accuracy of the dangerous action.
In FIG. 8, six danger areas from 4D to 4I are set. These risk areas were set based on the experience of nurses who actually performed nursing on site.

First, the dangerous area 4D is an area outside the fall prevention fence provided beside the bed. If a limb or head is detected in this area, it is determined that the action is dangerous.
The dangerous area 4E is an area above the partition on the foot side of the bed, and if a limb or head is detected in this area, it is determined as dangerous action.

  The dangerous area 4F is an area at a high position on the bed and is an area close to the ceiling. If the head or shoulder is detected in this area, the person to be monitored stands on the bed and is determined to have the highest degree of risk. This is because it is an action with high risk such as jumping off the bed.

  The dangerous area 4G is an area near the partition on the head side of the bed. When the head or shoulder is detected in this area, the person to be monitored is determined to have raised his / her upper body, and is determined to be a dangerous preparatory action. Just raising your upper body is not likely to take a dangerous action right away, but you should be careful from the standpoint of a nurse. However, if a foot is detected in this area, it is determined as a dangerous act.

  The dangerous area 4H is an area inside the bed and having a position slightly higher than the normal height when the monitored person is sleeping. In this area, when a limb or head is detected, the person being monitored is not necessarily likely to take a dangerous action. This is because it is considered that the monitored person is awake from sleep and is moving. However, in this case as well, from the standpoint of a nurse or the like, it is a state to be careful of, and is judged as a dangerous preliminary act.

  The dangerous area 4I is an area of the floor right next to the bed. If the body of the monitored person is present in this area, or if the head or hand is detected, it is considered that the monitored person has fallen from the bed. So it is judged as an emergency. Further, when there is only the feet of the monitored person in the 4C area, it is highly likely that the monitored person takes actions such as getting out of the bed and going outside, so it is determined as a dangerous act.

  In this way, by setting multiple risk areas in the plane direction and height direction based on the opinions of nurses who actually have nursing experience in the field, the situation of the monitored person can be detected reliably and in detail. Therefore, it is possible to respond appropriately according to the situation.

Also, according to the body part detected in each dangerous area, rank the degree of danger, and according to the rank, change the reporting method such as how to sound an alarm and how to contact the nurse, Nurses can take prompt and appropriate actions and actions.
As a report method, for example, the volume of the alarm is increased in the order of dangerous preliminary action, dangerous action, highest dangerous action, and emergency. In any case, the state of the monitored person can be known more quickly by displaying the image of the monitored person on the monitor screen or portable information terminal installed in the nurse's office.

If the risk ranking is low, a simple primary alarm is issued on the monitor screen or personal digital assistant installed in the station. After that, if there is no change in the behavior of the monitored person, the primary alarm will be issued after a certain time. May be released. When an action with a high risk ranking is detected after the primary alarm, various usage methods such as displaying a full-scale danger alarm on a monitor screen or a portable information terminal are possible.
That is, in the present invention, since a plurality of risk areas can be set freely, an optimal setting based on the experience of the nurse can be performed.

  Note that the monitoring system according to the present invention disclosed in Embodiments 1 and 2 can be used not only in a hospital but also in a home where a care facility or home caregiver is present, and an alarm is also required for a caregiver or family member. By notifying people, the same features as those used in hospitals can be exhibited.

1. Three-dimensional detection means Bed 3. Monitored person 4A, 4B, 4C, 4D, 4E, 4F, 4G, 4H, 4I Hazardous area 5. 6. Dangerous behavior determination means Reporting means 7. Monitor means 11 Infrared pulse irradiation means 12 Camera 51 Hazardous area setting means 53 Object determination means

Claims (7)

A monitoring system mainly used in a hospital room, comprising a three-dimensional detection means and a dangerous behavior determination means,
The three-dimensional detection means is a detection means capable of measuring the distance to the detected object for each pixel,
An infrared pulse irradiation means for irradiating infrared rays intermittently;
A camera having a detection unit in which pixels for detecting infrared rays are two-dimensionally arranged;
A distance deriving unit capable of deriving the distance to the detected object for each pixel by measuring the time until the infrared pulse irradiated by the infrared pulse irradiation means returns to the detection unit of the camera;
With
The dangerous behavior determination means includes
A dangerous area setting means for setting a dangerous area as a three-dimensional area in the monitoring area, and setting which part of the body part of the human body is determined to be dangerous when the dangerous area is entered;
From the two-dimensional shape or three-dimensional shape of the object detected in the dangerous area derived from the distance for each pixel detected by the three-dimensional detection means and the dangerous area, the object is a part of the human body. Object determining means for determining whether or not there is a high possibility, and determining which part of the human body the object is likely to be;
With
The object determination unit is configured such that the size of an object detected in the dangerous area derived from the distance for each pixel detected by the three-dimensional detection unit and the dangerous area is greater than or equal to a predetermined size. Determining whether or not the object is likely to be a part of a person's body, and determining which part of the person's body is likely to be the object,
The dangerous behavior determination means is dangerous behavior when the body part of the person determined to be highly likely to correspond to the object determination means matches the body part of the person set by the dangerous area setting means. A monitoring system characterized by   Determining whether or not the object detected in the dangerous area derived from the distance for each pixel detected by the three-dimensional detection means and the dangerous area is a human body part; In determining which part of the human body the object is likely to be, the determination is made by comparing with a list of a two-dimensional shape or a three-dimensional shape of a part of the human body provided in advance.
  The monitoring system according to claim 1.   Judgment is made on whether or not the action is dangerous by further considering from which direction the object detected in the dangerous area has entered the dangerous area.
  The monitoring system according to claim 1 or 2, characterized by the above-mentioned.   Judgment is made on whether or not the action is dangerous in consideration of the moving speed of the object detected in the dangerous area.
  The monitoring system according to any one of claims 1 to 3, wherein:   An alarm is activated when the dangerous behavior determination means determines that the behavior is dangerous
  The monitoring system according to any one of claims 1 to 4, wherein:   When the dangerous behavior determination means determines that the behavior is dangerous behavior, an alarm is activated depending on a part of the human body detected in the dangerous area.
  The monitoring system according to claim 5.   From the distance for each pixel detected by the three-dimensional detection means, an image forming means for deriving the position of the detection object for each pixel and forming an image;
  Communication means for transmitting the formed image as electronic data;
  Display means for receiving the electronic data transmitted by the communication means and displaying it as an image;
  Further equipped
  The monitoring system according to any one of claims 1 to 6, wherein: