JP5782737B2 - Status detection device, status detection method, and status detection program - Google Patents

Description

The present disclosure relates to a state detection device , a state detection method, and a state detection program that detect the state of a care recipient.

  When a hospitalized patient or a resident of a care facility tries to get out of the bed, an accident may occur that falls accidentally from the bed. In order to prevent the occurrence of such an accident, various fall prevention techniques have been proposed.

  One of the fall prevention techniques is a technique for detecting the condition of a subject who receives nursing or nursing care. For example, a technique has been proposed in which a subject's state is detected by a plurality of sensors including a pressure-sensitive sensor and an optical sensor, and the detected state is notified (see Patent Document 1). In addition, a technique for detecting the state of the subject using an image obtained by photographing the subject has been proposed. For example, the state of the subject is detected when a change equal to or greater than a threshold is detected in a preset monitoring area (see Patent Documents 2 and 3).

  In addition, a technique has been proposed in which a video image of a subject is relayed to a place where a nurse or a caregiver waits (see Patent Document 4).

JP 2001-327549 A JP 2006-175082 A JP 2005-128967 A JP 2000-105885 A

  In the conventional detection technique as described above, when an abnormality is detected in the state of a subject on or near the bed by a sensor including a camera, a detection result is notified to a nurse or an assistant. However, in the conventional technology, there are cases where a natural motion such as the subject turning over on the bed is erroneously detected as an abnormal state of the subject. Such frequent occurrence of erroneous detection may bother nurses and caregivers and hinder care. For this reason, reducing the frequency with which an abnormal state of the subject is erroneously detected is a problem of the state detection device and the state detection method.

  Moreover, although it is for watching the state of the subject person, if a large number of equipment is brought into a hospital room or living room, there is a risk that nursing may be hindered. For this reason, it is desirable that an apparatus for detecting the state of the subject be realized with as few equipment as possible.

Apparatus of the present disclosure, methods and programs, with a simple configuration, capable of detecting state detection device the condition of the subject with high precision, and to provide a state detection method and a state detection program.

A state detection device according to one aspect analyzes a plurality of images captured by an imaging device including a bed in an imaging range, and a moving direction of a subject included in the image is a longitudinal direction of the bed or a lateral direction An image processing unit that detects the orientation of the subject's body, a moving direction detected by the image processing unit, the orientation of the subject's body, and the past state of the subject And a determination unit that determines the current state of the subject and an output unit that outputs information indicating the current state of the subject obtained by the determination unit.

According to another aspect of the state detection method, a plurality of images captured by an imaging device including a bed in an imaging range are analyzed, and a moving direction of a target person included in the image is a longitudinal direction of the bed or a horizontal direction. And the direction of the subject's body, and based on the detected moving direction, the orientation of the subject's body, and the past state of the subject, A current state is determined, and information indicating the current state of the subject obtained by the determination is output.
Further, a state detection program according to another aspect analyzes a plurality of images captured by an imaging device including a bed in an imaging range, and the moving direction of the subject included in the image is the longitudinal direction of the bed or the horizontal direction. An image processing step for detecting whether the direction of the subject, the orientation of the subject's body, the moving direction detected by the image processing step, the orientation of the subject's body, and the past state of the subject Based on the above, the computer is caused to execute a determination step of determining the current state of the subject and an output step of outputting information indicating the current state of the subject obtained by the determination step.

According to the apparatus , method, and program of the present disclosure , the state of the subject can be detected with high accuracy by a simple configuration.

It is a figure which shows one Embodiment of a state detection apparatus. It is a figure explaining the relationship between the state of the subject person P and the moving direction of the subject's P image. It is a figure which shows the example of the judgment reference | standard for every past state. It is a figure which shows the example of the state transition diagram regarding the person of a detection target. It is a flowchart showing operation | movement of a state detection apparatus. It is a figure which shows another embodiment of a state detection apparatus. It is a figure explaining the process which detects the image of a person. It is a flowchart showing the process which detects the image of a person. It is a figure explaining the process which detects the moving direction of a person. It is a figure explaining the process which detects direction of a body. It is a figure which shows the example of the transition conditions which considered the direction of the body. It is a figure which shows another embodiment of a state detection apparatus. It is a flowchart showing the example of the notification determination process based on setting corresponding to a subject.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 shows an embodiment of a state detection device. The state detection device 10 illustrated in FIG. 1 includes an image processing unit 12, a determination unit 13, and an output unit 14. The output unit 14 illustrated in FIG. 1 includes a notification processing unit 15 and a speaker 16.

  For example, the state detection device 10 may be disposed close to the bed 1 where the state detection target person P stands. An image captured by the camera 2 is input to the image processing unit 12 of the state detection device 10. Note that the state detection device 10 may be installed in a place away from the camera 2 and the bed 1.

  As shown in FIG. 1, the camera 2 is desirably arranged so that the optical axis of the optical system for imaging coincides with the longitudinal direction of the bed 1. In this case, the imaging direction of the camera 2 is as indicated by the symbol Dr in FIG. 1, and the camera 2 images the subject on the bed 1 mainly from behind. And the imaging range R1 of the camera 2 arrange | positioned in this way can be made to correspond substantially with the range which the subject P of the bed 1 moves by the operation | movement at the time of standing on the bed 1. FIG. The operation when the subject person P stands on the bed 1 includes an action of the subject person P getting up on the bed 1 and an action of moving to the end of the bed 1.

  For example, the camera 2 performs imaging for the imaging range R1 described above at a predetermined frame rate. The captured image of each frame is sequentially sent to the image processing unit 12. The image processing unit 12 detects the moving direction of the image of the subject P described above based on all or part of the image of each frame. For example, the image processing unit 12 detects the moving direction of the subject P as follows.

  First, the image processing unit 12 detects an image of the subject P from each image. For detecting the image of the subject P, various known techniques such as a contour extraction technique and a subject recognition technique are applied. Next, the image processing unit 12 detects the moving direction of the subject person P by comparing the positions of the images of the subject person P detected from the images.

  The moving direction of the subject P detected for each image by the image processing unit 12 is passed to the determination unit 13. Note that the image processing unit 12 notifies the determination unit 13 that, for example, the image of the subject P is not detected, the subject P is not captured in the imaging range.

  The determination unit 13 determines the current state of the target person P based on the movement direction notified corresponding to the new image and the past state of the target person P. Processing for determining the current state of the subject P by the determination unit 13 will be described later.

  Information indicating the current state of the subject P determined by the determination unit 13 is passed to the notification processing unit 15 included in the output unit 14. The notification processing unit 15 determines whether the state indicated by the received information is a notification target state. The notification target state can be set in the notification processing unit 15 in advance. When the current state of the target person P matches the state to be notified, the notification processing unit 15 outputs a voice message or an alarm sound through the speaker 16, for example, to the nurse or caregiver. The state of the target person P is notified.

  Next, a process in which the determination unit 13 determines the current state of the target person P will be described. In the example described below, the determination unit 13 determines and distinguishes the state of the target person P into a plurality of types based on the posture of the target person P on the bed 1 or the position of the target person P and the bed 1.

  The state of the subject P is, for example, any of a landing state, a wake-up state, a leaving state, and an absent state. The landing state is a state in which the subject P is lying on the bed 1. In addition, the wake-up state is a state in which the subject P is standing on the bed 1. The bed leaving state is a state in which the subject P is at the end of the bed 1. The bed leaving state includes a state in which the subject P is sitting on the end of the bed 1, a state in which a side rail is installed on the bed 1, and a state in which the subject P stands on the side of the bed 1. The absence state is a state in which the subject P is away from the bed 1. These four states correspond to the start point or branch point of the motion of the subject P on the bed 1. Therefore, these states are suitable for notifying the nurse or the caregiver when watching the subject P.

  In FIG. 2, the figure explaining the relationship between the state of the subject P and the moving direction of the subject P's image is shown. 2A, 2C, and 2E are examples of images obtained by imaging with the camera 2 described with reference to FIG. 2 (B), (D), and (F) are postures of the subject P when the images shown in FIGS. 2 (A), (C), and (E) are captured. It is an example. In FIGS. 2B, 2D, and 2F, the imaging direction of the camera 2 is indicated by a symbol Dr.

  In the example shown in FIGS. 2A, 2C, and 2E, the detection region R2 is provided inside the imaging range R1. In the examples of FIGS. 2A, 2C, and 2E, the detection region R2 is a trapezoidal region surrounded by a thick solid line. The detection region R2 is divided into three regions by boundary lines BL and BR indicated by dotted lines in FIG. In the following description, a rectangular area sandwiched between the boundary line BL and the boundary line BR in the detection area R2 is referred to as a central portion. In the detection region R2, a region on the left side of the boundary line BL is referred to as a left end portion, and a region on the right side of the boundary line BR is referred to as a right end portion.

  FIG. 2B shows an example in which the subject P is in the landing state. In the landing state, as shown in FIG. 2A, the image of the subject P does not appear in the imaging range R1. An image that does not include an image of the subject person P is acquired by the camera 2 even when the subject person P is absent. Therefore, as shown in FIG. 2 (A), the state of the subject P estimated when the image not including the image of the subject P is obtained is either the landing state or the absence state. is there.

  On the other hand, FIG. 2D shows a scene in which the subject P is about to get up on the bed 1. An arrow shown in FIG. 2D indicates the moving direction of the subject P on the bed 1. At this time, as shown in FIG. 2C, the image of the subject P is included in the center of the detection region R2. Then, the image processing unit 12 detects a moving direction from the bottom to the top of the screen, as indicated by reference numeral V1 in FIG. 2C, from an image captured while the subject person P is about to get up.

  FIG. 2F shows a scene in which the subject P is about to stand up from a state where he is sitting on the end of the bed 1. At this time, as shown in FIG. 2E, the image of the subject P is included in the left region of the detection region R2. Here, the image of the subject P corresponds to the end of the bed 1 from the central portion of the detection region R2 until the subject P who has been standing on the bed 1 shifts to the state of sitting on the end of the bed. Move to the left edge. Further, when the subject person P tries to stand up from the state of sitting on the bed 1, the subject person P often performs an operation of tilting the body forward. Therefore, from the images taken before and after the image shown in FIG. 2 (E), the image processing unit 12 detects the moving direction from the right to the left of the screen as indicated by reference numeral V2 in FIG. 2 (E). To do.

  As can be seen from the example shown in FIG. 2, the process of changing the posture of the subject P on the bed 1 and the process of changing the positional relationship between the subject P and the bed 1 are related to each other. In other words, the process in which the target person P in the landing state becomes absent, or conversely, the process in which the target person P who has been in the absence state becomes in the landing state is a state transition through some intermediate states. Can be caught.

  Then, for example, while the state of the subject person P transitions from the landing state to the wake-up state, the image processing unit 12 displays the image of the subject person P in the detection region R2 as shown in FIG. A moving direction V1 directed upward is detected. Conversely, while the state of the subject P transitions from the wake-up state to the landing state, the image processing unit 12 reverses the detection region R2 for the image of the subject P, contrary to the example of FIG. The moving direction from the inside to the outside of the detection region R2 across the lower boundary of the detection region R2 is detected. In addition, while the state of the subject person P transitions from the wake-up state to the leaving state, the image processing unit 12 performs the image of the subject person P from the center of the detection region R2 as shown in FIG. A moving direction V2 toward the left end or the right end is detected. Conversely, while the state of the subject person P transitions from the leaving state to the rising state, the image processing unit 12 reverses the left end of the detection region R2 for the image of the subject person P, contrary to the example of FIG. Direction of movement from the center or right end to the center. On the other hand, during the transition of the state of the subject P from the bed leaving state to the absent state, the image processing unit 12 performs the image of the subject P from the left end or the right end of the detection region R2 on the corresponding side of the detection region R2. A moving direction toward the outside of the detection region R2 beyond the boundary is detected. Then, while the state of the subject P transitions from the absence state to the bed leaving state, the image processing unit 12 moves the image of the subject P from the outside of the detection region R2 toward the left end or the right end of the detection region R2. Detect direction. In some cases, the subject P may lie near the position where the subject P sits on the end of the bed 1. As described above, while the state of the subject person P transitions from the bed leaving state to the landing state, the image processing unit 12 detects the image of the subject person P from the left end portion or the right end portion of the detection region R2 below the detection region R2. The moving direction toward the outside of the detection region R2 across the side boundary is detected.

  Thus, there is a correlation between the above-described state transition between the four states and the moving direction of the target person P detected by the image processing unit 12. Accordingly, the current state of the subject person P can be uniquely identified based on the combination of the past state already identified for the subject person P and the moving direction of the image of the subject person P.

  The determination unit 13 determines the current state of the target person P based on the combination of the past state of the target person P and the newly detected moving direction using the above-described correlation. For example, a set of determination criteria combining the past state and the moving direction of the subject person P can be set in advance in the determination unit 13 for each state.

  FIG. 3 shows an example of judgment criteria for each past state. The determination criteria shown in FIG. 3 summarize the correspondence between the above-described state transition and the detected moving direction in a table.

  In the example illustrated in FIG. 3, there are two criteria for determining the wake-up state as the current state of the subject P. One is a combination of “the past state is a landing state” and “the moving direction toward the detection region R2 upward beyond the lower boundary of the detection region R2 has been detected”. The other is a combination of “the past state is a bed leaving state” and “the movement direction from the left or right end of the detection region R2 toward the center of the detection region R2 is detected”.

  Similarly, in the example illustrated in FIG. 3, there are also two criteria for determining the landing state as the current state of the subject P. One is a combination of “the past state is a wake-up state” and “the movement direction toward the outside of the detection region R2 is detected downward beyond the lower boundary of the detection region R2”. The other is a combination of “the past state is a bed leaving state” and “the moving direction toward the outside of the detection region R2 is detected downward beyond the lower boundary of the detection region R2”. .

  In the example shown in FIG. 3, there are two criteria for determining the bed leaving state as the current state of the subject person P. One is a combination of “the past state is a wake-up state” and “the moving direction from the central portion of the detection region R2 toward the left end portion or the right end portion is detected”. The other is a combination of “the past state is absent” and “the movement direction from the outside of the detection region R2 toward the left end or the right end of the detection region R2 is detected”.

  On the other hand, in the example shown in FIG. 3, the determination criteria for determining the absence state as the current state of the subject P are “the past state is a bed leaving state” and “from the left end portion or the right end portion of the detection region R2. This is a combination that “the moving direction toward the outside of the detection region R2 has been detected”.

  By using a determination criterion that combines the past state and the moving direction of the image of the subject P in the determination process of the determination unit 13, it is possible to avoid outputting an erroneous determination result in the following example.

  For example, the image processing unit 12 may detect a moving direction from the central portion of the detection region R2 toward the left end portion or the right end portion in accordance with the movement of an object other than the target person P. Such a detection result is obtained from the image processing unit 12 when the curtain shakes or when another person passes around the bed 1. In this case, the detected moving direction coincides with the moving direction detected when the subject enters the bed leaving state. However, the fact that the past state of the subject person P is the landing state does not satisfy the condition for determining the leaving state. Therefore, according to the determination unit 13 that employs the determination criterion that combines the moving direction and the past state, the state of the subject P is not erroneously determined as the bed leaving state in the above-described example.

  Further, in the set of determination criteria illustrated in FIG. 3, the moving directions included in the determination criteria used when determining whether the subject P is in the wake-up state or the leaving state are clearly different for each past state. .

  For example, among the determination criteria used for determining the wake-up state, the moving direction combined when the past state is the landing state is the direction from the bottom to the top of the image. On the other hand, among the determination criteria used for determining the wake-up state, the moving direction combined when the past state is the leaving state is the left-right direction of the image. In this way, by using a determination criterion in which a clearly different moving direction is combined with a corresponding past state type, the determination unit 13 can determine the current state of the subject P more accurately. That is, according to the determination unit 13 to which the set of determination criteria illustrated in FIG. 3 is applied, the rising state and the leaving state can be determined more accurately.

  In this way, by applying a determination criterion combining the moving direction and the past state to the determination process of the determination unit 13, the probability that the current state of the subject P is erroneously determined can be reduced. That is, by using a determination criterion that combines the moving direction and the past state, the determination unit 13 can reliably determine the state of the subject P with high accuracy.

  Then, by outputting information indicating the current state of the subject P obtained by the discriminating unit 13 by the output unit 14, accurate information is provided to people in charge of the subject, such as nurses and assistants. You can be notified.

  Thus, according to the state detection device of the present disclosure, the state of the subject P can be reliably detected with high accuracy. Then, by enabling accurate determination of the state of the subject person P, the fact that the subject person P is erroneously notified of an abnormal state, that is, false detection can be reduced. That is, according to the state detection device of the present disclosure, it is possible to prevent the occurrence of a situation in which people who are in charge of care of the subject are bothered unnecessarily by reducing erroneous detection. Thereby, the work burden of the people in charge of the care of the subject can be reduced.

  Further, the state detection apparatus 10 illustrated in FIG. 1 can be realized using, for example, a small computer system including a processor and a memory. Moreover, since the imaging direction of the camera 2 that inputs an image to the state detection device 10 disclosed in FIG. 1 is the longitudinal direction of the bed 1, the camera 2 is attached to a headboard or a wall of the bed 1, for example. The camera 2 installed in such a place is unlikely to obstruct the operation of the subject P and the nurse or caregiver. In addition, according to the camera 2 installed in this way, it is possible to capture the range in which the subject P moves on the bed 1 without omission, so that it is not necessary to install another camera. Therefore, in order to introduce the state detection apparatus 10 of this indication, the apparatus newly arrange | positioned in a hospital room or a living room is a small scale. The state detection device 10 realized using a small-scale computer system can be miniaturized to the extent that the bed 1 can be embedded in the headboard together with the camera 2 described above, for example.

  Furthermore, in the image captured by the camera 2 with the longitudinal direction of the bed 1 as the imaging direction, for example, depending on the motion of the subject P getting up and the motion of the subject P moving to the end of the bed 1 The image of P moves in a clearly different direction. Based on such an image, the image processing unit 12 can detect clearly different moving directions corresponding to various operations of the subject P.

  In addition, a state transition diagram can be created for the state transitions between the above-described landing state, wake-up state, leaving bed state, and absent state. And the discrimination | determination part 13 may perform a discrimination | determination process based on this state transition diagram.

  In FIG. 4, the example of the state transition diagram regarding a subject is shown. In the state transition diagram illustrated in FIG. 4, transitions assumed among the four states of the landing state, the wake-up state, the leaving state, and the absent state are indicated by arrows with reference signs T <b> 1 to T <b> 7.

  A state transition T1 denoted by reference numeral T1 in FIG. 4 indicates a transition in which the state of the subject P moves from the landing state to the rising state. The condition of this state transition T1 is that the image processing unit 12 detects the moving direction of the image of the subject P, which goes upward in the detection region R2 beyond the lower boundary of the detection region R2. .

  Moreover, the state transition T2 shown by attaching | subjecting the code | symbol T2 in FIG. 4 has shown the transition from which the state of the subject P goes from a leaving state to a rising state. The condition of this state transition T2 is that the moving direction from the left end or the right end of the detection region R2 toward the center of the image of the subject P is detected by the image processing unit 12.

  On the other hand, the state transition T3 indicated by the reference numeral T3 in FIG. 4 indicates a transition in which the state of the subject P moves from the rising state to the landing state. The condition of the state transition T3 is that the image processing unit 12 detects a moving direction of the image of the subject P that goes downward beyond the lower boundary of the detection region R2 and goes outside the detection region R2. is there.

  Moreover, the state transition T4 shown by attaching | subjecting the code | symbol T4 in FIG. 4 has shown the transition which the state of the subject P heads from a leaving state to a landing state. The condition of this state transition T4 is that the image processing unit 12 detects the moving direction of the image of the subject P toward the outside of the detection region R2 over the lower boundary of the detection region R2. is there.

  In addition, a state transition T5 indicated by a reference numeral T5 in FIG. 4 indicates a transition in which the state of the subject P is from the wake-up state to the leaving state. The condition for this state transition T5 is that the moving direction from the center of the detection region R2 toward the left end or the right end of the image of the subject P is detected by the image processing unit 12. In addition, when the side to which the subject person P gets off is limited, detection of the moving direction toward the range of the detection region R2 corresponding to the side where the person can get out may be set as the condition of the state transition T5. it can. For example, if a side rail or the like is installed on the right side of the bed 1, detection of the moving direction toward the left end of the detection region R2 corresponding to the left side that is not installed is selectively set as a condition for the state transition T5. be able to.

  Moreover, the state transition T6 shown by attaching | subjecting the code | symbol T6 in FIG. 4 has shown the transition which the state of the subject P goes from an absent state to a bed leaving state. The condition of this state transition T6 is that the moving direction from the outside of the detection region R2 toward the left end or the right end of the detection region R2 is detected by the image processing unit 12 with respect to the image of the subject P.

  And state transition T7 which attached | subjected code | symbol T7 in FIG. 4 has shown the transition which the state of the subject P goes to a absent state from a bed leaving state. The condition for this state transition T7 is that the image processor 12 detects the moving direction of the image of the subject P from the left end or the right end of the detection region R2 toward the outside.

  The state transition diagram illustrated in FIG. 4 is equivalent to the set of criteria illustrated in FIG. The determination unit 13 determines the current state of the target person P based on a condition corresponding to the state transition from the past state according to the state transition diagram.

  Among the four states illustrated in FIG. 4, the wake-up state and the leaving state correspond to the branch points of the operation of the subject P on the bed 1, respectively. The moving directions indicated by the conditions corresponding to the state transitions T3 and T5 starting from the wake-up state are clearly different. Similarly, the moving directions indicated by the conditions corresponding to the state transitions T2 and T7 starting from the bed leaving state are also clearly different.

  Therefore, when the past state corresponds to a branch point of the motion of the subject P, the determination unit 13 determines the state starting from the past state depending on which of the clearly different movement directions is detected by the image processing unit 12. It is possible to determine which of the transitions has occurred.

  For example, when the past state of the target person P is a wake-up state, when the image processing unit 12 detects a moving direction from the center of the detection region R2 toward the left end or the right end, the determination unit 13 It is determined that the state transition T5 described above has occurred. In this case, the determination unit 13 determines the bed leaving state as the current state of the subject P. On the other hand, when the past state of the subject person P is a wake-up state, when the moving direction toward the outside of the detection region R2 is detected downward beyond the lower boundary of the detection region R2, the determination unit 13 It is determined that the state transition T3 described above has occurred. In this case, the determination unit 13 determines the landing state as the current state of the subject person P.

  In this way, the determination unit 13 determines the current state of the target person P by tracing the state transition diagram. Such a determination unit 13 is an example of a determination unit that uses whether or not the moving direction of the target person P corresponding to a state transition starting from the past state is detected as a different reference for each past state. is there. And such a discrimination | determination part 13 is based on whether the moving direction of the target person P corresponding to each of at least 1 state transition which makes a past state a starting point is detected, which of these state transitions is. It is possible to reliably determine whether it has occurred. Further, based on the determination result, the determination unit 13 can reliably determine the current state of the subject P.

  Next, how the state detection apparatus 10 illustrated in FIG. 1 observes the state of the subject P on the bed 1 and notifies the nurse and the caregiver will be described.

  FIG. 5 is a flowchart showing the operation of the state detection device.

  In step 301, a process for setting the initial state of the subject P is performed on the determination unit 13, and a process for setting the notification target state is performed on the notification processing unit 15. The initial state of the subject person P set in step 301 is held in the determination unit 13 as a past state. Further, the notification target state is held inside the notification processing unit 15.

  Next, every time an image is acquired by the camera 2 (step 302), the image processing unit 12 detects an image of the subject P (step 303). The image processing unit 12 may detect the target person P by receiving one frame image every time a predetermined number of frames are captured by the camera 2. Next, in step 304, the image processing unit 12 determines whether or not the detection of the image of the subject P has succeeded.

  When the image of the subject person P is detected from the detection region R2 of the newly acquired image (affirmative determination in step 304), the image processing unit 12 further detects the moving direction of the image of the subject person P (step step). 305). For example, the image processing unit 12 may detect the moving direction of the image of the subject P based on a comparison between an image of a new frame and an image of the previous frame. The moving direction of the image of the subject P can also be detected based on the new image and a plurality of images acquired prior to the new image.

  In step 305, the moving direction detected by the image processing unit 12 is passed to the determination unit 13. On the other hand, in the case of negative determination in step 304, the image processing unit 12 notifies the determination unit 13 that the detection of the image of the subject P has failed.

  In step 307, the determination unit 13 determines the current state of the subject P based on the detection result received from the image processing unit 12, the past state held in the determination unit 13, and the state transition diagram described above. Determine. At this time, the determination unit 13 checks whether or not the state transition has occurred by comparing the condition of the state transition starting from the past state held inside with the detection result received from the image processing unit 12. Can be judged. For example, when the landing state is held in the determination unit 13 as the past state, the determination unit 13 collates the detection result received from the image processing unit 12 with the condition of the state transition T1 illustrated in FIG. To do. When the movement direction that matches the condition of the state transition T1 is received from the image processing unit 12, the determination unit 13 determines that the current state of the subject P is a wake-up state. On the other hand, when the moving direction received from the image processing unit 12 and the condition of the state transition T1 do not match, the determination unit 13 determines that the current state of the subject P is the same landing state as the past state. .

  Next, the determination unit 13 determines whether or not the current state obtained in Step 307 has changed from the past state (Step 308). When the determination unit 13 determines that the state has changed (Yes determination in step 308), the determination unit 13 notifies the notification processing unit 15 of the output unit 14 of the current state.

  The notification processing unit 15 determines whether or not the current state of the target person P notified from the determination unit 13 matches the state of the notification target held inside the notification processing unit 15 in Step 301 (Step S1). 309). When the current state of the subject person P and the state to be notified do not match (negative determination at step 309), the notification processing unit 15 does not perform a process of notifying the current state of the subject person P. In this case, the determination unit 13 holds the current state of the target person P in order to use it as a past state in subsequent processing (step 310). Thereafter, the processing returns to step 302, and processing for a newly captured image is started.

  On the other hand, when the current state of the target person P notified from the determination unit 13 matches the state of the notification target (affirmative determination in step 309), the notification processing unit 15 displays the current state of the target person P. Processing for notifying the information shown is performed (step 311). At this time, the notification processing unit 15 illustrated in FIG. 1 outputs a voice message, an alarm voice, or the like set in advance in the notification processing unit 15 via the speaker 16. Instead of performing sound output via the speaker 16, the notification processing unit 15 uses an existing notification system such as a nurse call system to notify that the state of the target person P has shifted to the notification target state. May be output.

  As described above, according to the state detection method disclosed herein, the current state of the subject P on the bed 1 can be determined, and the determination result can be output. In the flowchart shown in FIG. 5, step 302 is an example of a procedure for inputting an image obtained by imaging the target person P. Steps 303 to 306 are an example of a procedure for detecting the moving direction of the subject P. Steps 307 and 310 are an example of a procedure for determining the current state of the subject P. Steps 308 to 309 and step 311 are an example of a procedure for outputting information indicating the current state of the subject P.

  Next, various embodiments of the state detection device disclosed herein will be described.

  FIG. 6 shows another embodiment of the state detection device disclosed herein. 6 that are the same as those shown in FIG. 1 are given the same reference numerals, and descriptions thereof are omitted.

  The state detection device 20 illustrated in FIG. 6 includes a camera interface (IF) 21, a nurse call interface (IF) 23, a processor 24, and a memory 25. The processor 24, the camera IF 21, the nurse call IF 23, and the memory 25 are connected via a bus. The camera IF 21 is connected to the camera 2. The nurse call IF 23 is connected to the nurse call system 4.

  In the memory 25 shown in FIG. 6, an image recognition program 26, a state determination program 27, and a notification determination program 28 are stored. The processor 24 that reads the image recognition program 26 and operates according to the program is another embodiment of the image processing unit 12. Similarly, the processor 24 that reads the state determination program 27 and operates according to this program is another embodiment of the determination unit 13. The processor 24, the notification IF 23, and the nurse call system 4 that read the notification determination program 28 and operate according to this program are other embodiments of the notification processing unit 15.

  The illumination unit 3 illustrated in FIG. 6 illuminates the subject P on the bed 1 with, for example, infrared light in the near infrared region. In addition, in FIG. 6, the range illuminated by the illumination part 3 was shown with the code | symbol R3. The range R3 illuminated by the illuminating unit 3 preferably includes the imaging range of the camera 2. The camera 2 has sensitivity in a range including the wavelength of illumination light emitted from the illumination unit 3. The illumination control unit 5 blinks illumination light from the illumination unit 3 in synchronization with imaging by the camera 2. For example, the illumination control unit 5 controls the illumination unit 3 so that the illumination light is turned on when an odd frame is captured by the camera 2 and is turned off when an even frame is captured. In this manner, an image of the subject P illuminated with light in the near infrared region every other frame can be acquired. The image acquired in this way is passed to the processor 24 via the camera IF 21.

  As described above, the state detection device 20 configured to receive an image of the subject P illuminated by light in the near infrared region every other frame is useful when detecting the state of the subject P at night. is there. Hereinafter, processing for detecting an image of the subject P using blinking illumination light and processing for detecting the moving direction of the subject P based on the detected image of the subject P will be described.

  FIG. 7 is a diagram illustrating processing for detecting an image of the subject P. FIG. 8 is a flowchart showing processing for detecting an image of the subject P. FIG. 9 is a diagram illustrating processing for detecting the moving direction of the subject person.

  FIG. 7A shows an example of an image captured by the camera 2 in a state where the illumination by the illumination unit 3 is turned on. FIG. 7B shows an example of an image captured by the camera 2 with the illumination by the illumination unit 3 turned off.

  The subject P on the bed 1 is closer to the illumination unit 3 than the surrounding objects such as the bed 1. For this reason, in the example of FIG. 7A, the luminance of the image of the subject P is higher than that of the example of FIG. On the other hand, the distance between the surrounding object such as the bed 1 and the illumination unit 3 is larger than the distance between the illumination unit 3 and the subject P. For this reason, compared with the change of the brightness between the example of FIG. 7A and the example of FIG.

  For this reason, in the difference image created using the image captured with the illumination turned on and the image captured with the illumination turned off, as shown in FIG. The brightness difference from the object image is emphasized. From such a difference image, for example, by extracting a region having a brightness equal to or higher than a predetermined threshold, an image of the subject person P can be easily extracted. 7A to 7C, the boundary of the detection region R2 is indicated by a white line.

  The flowchart shown in FIG. 8 shows a procedure related to processing for detecting an image of the subject P in the image recognition program 26. Each step shown in FIG. 8 is executed by the processor 24 that has read the image recognition program 26.

  In step 321 and step 322, the processor 24 acquires images captured with the illumination turned on and turned off via the camera IF 21. Next, the processor 24 generates a difference image from the two-frame images acquired in steps 321 and 322 (step 323). Next, the processor 24 extracts a region whose brightness is equal to or greater than a predetermined threshold value from the difference image (step 324).

  When the area of the extracted region is less than another predetermined threshold value (No determination at Step 325), the processor 24 outputs that the detection of the image of the subject P has failed (Step 327). In this case, the processor 24 starts processing for determining the state of the subject P based on the state determination program 27.

  On the other hand, when the area of the extracted region is equal to or larger than another predetermined threshold (Yes in Step 325), the processor 24 outputs that the detection of the image of the subject P has succeeded (Step 326). In this case, the processor 24 starts a process of detecting the moving direction of the subject P.

  When receiving an image of the subject P illuminated every other frame, the state detection device 20 detects the image of the subject P by a simple process as shown in Steps 321 to 327 described above. be able to.

  The moving direction of the subject person P can be detected by performing the following process on the image of the subject person P extracted in this way.

  FIGS. 9A and 9C show examples of postures that the subject person P can take when trying to get up on the bed 1. FIGS. 9B and 9D are generated from images captured by the camera 2 when the subject P is in the posture shown in FIGS. 9A and 9C, respectively. It is an example of a difference image. 9A and 9C corresponds to the vertical range of the detection region R2 included in the imaging range of the camera 2. Further, in the examples of FIGS. 9B and 9D, the image of the subject P is shaded.

  FIG. 9A shows an example of a posture in which the subject P slightly lifts his head. At this time, the difference image obtained from the image captured by the camera 2 includes a slight part of the image corresponding to the back of the subject P in the detection region R2, as shown in FIG. 9B. .

  FIG. 9C is an example of a posture in which the subject P is halfway up. At this time, as shown in FIG. 9D, the difference image obtained from the image captured by the camera 2 corresponds to the image of the back of the subject P and the upper half of the subject P in the detection region R2. Part of the image is included.

  The moving direction of the subject P can be obtained from, for example, the movement of the center of gravity of the portion detected as the image of the subject P in each difference image. For example, when the difference images shown in FIGS. 9B and 9D are obtained in order, the processor 24 may obtain the moving direction of the target person P as follows. First, the processor 24 obtains the position of the center of gravity of the portion corresponding to the image of the subject P based on each difference image. Then, the moving direction toward the center of gravity obtained from the difference image shown in FIG. 9D may be detected from the center of gravity obtained from the difference image shown in FIG. 9B. In this way, the processor 24 detects the moving direction that enters the detection region R2 from the lower boundary of the detection region R2, based on the difference images shown in FIGS. 9B and 9D.

  As described above, when using illumination light in the near-infrared region, the back of the head of the subject P may not appear clearly in the difference image, for example, when the hair of the subject P is black. However, even when an image corresponding to the back of the head is not detected, when a part of the upper body of the subject P enters the imaging range of the camera 2, an image of the upper body of the subject P appears in the difference image. Then, by performing the above-described processing on the difference image including the upper body image of the subject person P, the processor 24 moves the subject person P from the movement direction of the center of gravity of the image corresponding to a part of the upper body. Can be detected. Further, the processor 24 may use any method for detecting the moving direction of the subject person P as long as it detects a change in the position of the image of the subject person P.

  By the way, near-infrared light is difficult to be detected by human vision. In addition, since the illumination unit 3 and the camera 2 illustrated in FIG. 6 mainly perform illumination and imaging from behind the subject P, the subject P does not have the impression of being monitored by the subject P. You can watch over the state.

  Moreover, the method of detecting the image of the subject P using the above-described difference image is useful not only for nighttime watching but also for daytime watching. Because, in the process of generating the difference image as described above, the image of the object by the light other than the illumination unit 3 is canceled, and only the image by the light from the illumination unit 3 can be extracted. This is because the target person P can be extracted by the same processing as in the case of watching. For example, an image of an object confused with the target person P such as another person passing through the foot side of the bed 1 or a partition curtain is canceled by the light other than the illumination unit 3 in the process of generating the difference image, and the illumination unit Since the distance from 3 is large, the image of the light of the illumination unit 3 has a brightness that can be easily distinguished from the image of the subject P. Thereby, when detecting the image of the subject P based on the difference image as described above, it is possible to reduce erroneous detection due to the influence of the image of a person other than the subject P.

  The image recognition program 27 detects, for example, the method of detecting the image of the subject P using the above-described difference image and subject recognition of the image captured using visible light to detect the image of the subject P. It is also possible to include both methods. Further, a camera having sensitivity in the far infrared region may be used instead of the camera 2 described above. In this case, the image of the subject P can be detected by extracting a temperature region corresponding to the human body temperature from the temperature image acquired using the far-infrared camera. Moreover, you may perform the process which produces | generates a difference image as mentioned above on the camera 2 side. In this case, the generated difference image is passed to the processor 24 via the camera IF 21.

  Next, a method for extracting information useful as an index for determining the state of the subject P from an image obtained by imaging the subject P on the bed will be described.

  FIG. 10 is a diagram illustrating processing for detecting the orientation of the subject's body. 10A and 10C show examples of postures that the subject P on the bed 1 can take. FIGS. 10B and 10D show difference images generated from images captured by the camera 2 when the subject P is in the posture shown in FIGS. 10A and 10C, respectively. It is an example. 10A and 10C corresponds to the vertical range of the detection region R2 included in the imaging range of the camera 2. In the examples of FIGS. 10B and 10D, the image of the subject P is shown with shading.

  FIG. 10A is an example of a posture in which the subject P gets up on the bed 1. At this time, as shown in FIG. 10B, the difference image obtained from the image captured by the camera 2 includes images corresponding to the back of the back of the subject P and the back of the upper body in the center of the detection region R2. ing.

  FIG. 10C is an example of a posture in which the subject P is sitting on the end of the bed 1. At this time, the difference image obtained from the image captured by the camera 2 includes an image of the temporal region of the subject person P and the upper body of the subject person P on the left side of the detection region R2, as shown in FIG. Part of the image corresponding to the side is included.

  As can be seen from the comparison between FIGS. 10A and 10C, the orientation of the body is different when the subject P is getting up on the bed 1 and when the subject P is sitting on the end of the bed 1. . Such a difference in body orientation is reflected as a difference in the shape of the image corresponding to the upper half of the subject P included in the difference images shown in FIGS. 10 (B) and 10 (D). Such a difference in shape can be represented by, for example, a ratio of the vertical size and the horizontal size of the image corresponding to the upper body. 10B and 10D, in the image corresponding to the upper body of the subject P, the vertical size and the horizontal size of the portion included in the detection region R2 are denoted by a, Indicated by b.

  Therefore, for example, when the processor 24 executes the following three processes, it is possible to detect the orientation of the body of the subject P based on the shape of the image of the subject P included in the difference image. .

  First, the processor 24 acquires the sizes a and b in the vertical direction and the horizontal direction for a portion corresponding to the upper half of the subject P among the images of the subject P included in the difference image. Next, for example, the processor 24 divides the vertical size a by the horizontal size b to obtain the aspect ratio c. Next, the processor 24 compares the aspect ratio c with a predetermined threshold Th, and determines the orientation of the body of the subject P according to the comparison result. For example, when the aspect ratio c is larger than a predetermined threshold Th, the processor 24 can determine that the subject P is facing the camera 2 side. In this case, the processor 24 determines that the body direction of the subject P is horizontal. On the other hand, when the aspect ratio c is smaller than the predetermined threshold Th, the processor 24 can determine that the subject P is facing the camera 2 with his back or front. In this case, the processor 24 determines that the body direction of the subject P is the front. Note that the value of the threshold Th described above may be determined based on, for example, the aspect ratio obtained for the front-facing images of various subjects and the aspect ratio obtained for the lateral-facing images of various subjects. . Further, as described above, the target of the process for determining the orientation of the body of the subject based on the aspect ratio of the image of the subject P is not limited to the difference image described above. For example, the processor 24 can perform a process of determining the orientation of the subject's body based on the aspect ratio of the image of the subject P detected from the image captured using visible light.

  Then, the orientation of the body of the subject P obtained as described above can be used for processing for determining the state of the subject P as follows.

  As can be seen from the examples shown in FIGS. 10A and 10B, in the wake-up state, there is a high possibility that the subject P on the bed 1 is facing the front. Therefore, it is desirable to add, as an AND condition, that the direction of the body of the subject P is the front direction to the condition for determining the state transitions T1 and T2 toward the rising state.

  On the other hand, as can be seen from the examples shown in FIGS. 10C and 10D, in the state of getting out of bed, there is a high possibility that the subject P on the bed 1 is facing sideways. Therefore, it is desirable to add to the condition for determining the state transitions T5 and T6 toward the bed leaving state that the direction of the body of the subject P is horizontal as an AND condition.

  FIG. 11 shows an example of the transition condition considering the body direction. The transition condition shown in FIG. 11 is obtained by adding the AND condition described above to the transition condition shown in the description of the state transition diagram shown in FIG.

  As shown in FIG. 11, the state determination program 27 stored in the memory 25 illustrated in FIG. 6 can include a transition condition that includes the determination result of the body orientation as a condition for some state transitions. The processor 24 operating according to the state determination program 27 including such a transition condition can determine with high accuracy that the subject P is in the wake-up state and that the subject P is in the get-off state. .

  Furthermore, it is also possible to detect an abnormal movement of the subject person P using the body direction of the subject person P obtained by the determination process described above.

  For example, the following changes are observed in the normal process in which the subject P moves from the rising state to the leaving state. The image of the subject P included in the image moves from the center of the detection region R2 to the left end or the right end, and the aspect ratio of the image corresponding to the upper body of the subject P is horizontal from the value corresponding to the front direction. It changes to the value corresponding to. On the other hand, when the following change is detected, it can be estimated that the movement of the subject P is abnormal. For example, when the image of the subject P moves toward the outside of the detection region R2 while maintaining the value corresponding to the frontal aspect ratio, the subject P may have fallen from the bed 1. is there.

  For example, the state determination program 27 determines whether there is a change in the aspect ratio of the image corresponding to the upper half of the subject P and the moving direction of the image of the subject P in the procedure for judging whether or not there is a state transition from the bed leaving state. A procedure for detecting discrepancies can be included. The processor 24 that executes the state determination program 27 detects an abnormal state change different from the normal state transition T7 from the bed leaving state to the absent state when detecting the contradiction as described above. . Further, when the abnormal state change is detected in this manner, the processor 24 notifies the nurse call system 4 of an alarm with a higher degree of urgency via the nurse call IF 23.

  By the way, the state detection apparatus 20 of this indication can also be arrange | positioned in a sickroom or a living room like the example shown in FIG. 6, and part or all can also be arrange | positioned in a nurse station etc.

  FIG. 12 shows another embodiment of the state detection device. Note that among the constituent elements shown in FIG. 12, those equivalent to the constituent elements shown in FIG. 6 are denoted by the same reference numerals, and description thereof is omitted.

  The state detection apparatus illustrated in FIG. 12 includes a local unit 30 installed in a hospital room or a living room, and a center unit 32 installed in a nurse station. The local unit 30 has a notification interface (IF) 31 instead of the nurse call IF 23 shown in FIG. 6 is stored in the memory 34 of the center unit 32, not the memory 25 of the local unit 30.

  The center unit 32 illustrated in FIG. 12 includes a notification interface (IF) 33 corresponding to the notification IF 31 described above, a memory 34, a processor 35, a nurse call IF 23, and a database interface (DBIF) 36. The processor 35, the notification IF 33, the memory 34, the nurse call IF 23, and the DBIF 36 are connected to each other via a bus. The processor 35 of the center unit 32 is connected to the nurse call system 4 via the nurse call IF 23. The processor 35 is connected to a target person information database (DB) 6 via a DBIF 36.

  In the state detection apparatus illustrated in FIG. 12, information including the state of each subject P is notified from the local unit 30 arranged in each hospital room to the center unit 32. The notification IF 31 included in the local unit 30 shown in FIG. 12 adds the installation location of the local unit 30 or the identification information of the subject person P when notifying the center unit 32 of the current state of the subject person P. . Hereinafter, a case will be described in which the notification unit 31 notifies the center unit 32 of a target person ID for identifying an individual target person and information indicating the state of the target person P.

  Further, the target person information DB 6 includes information for individually setting a state in which notification to the nurse or the assistant is required for each target person. Further, the notification determination program stored in the memory 34 includes a procedure for determining whether or not to notify the nurse call system 4 based on such a setting corresponding to the subject. By reading such a notification determination program, the processor 35 can control the notification to the nurse call system 4 based on the setting corresponding to the subject.

  FIG. 13 is a flowchart illustrating an example of the notification determination process based on the target person corresponding setting. Each time the steps included in the flowchart illustrated in FIG. 13 are received from the local unit 30, the processor 35 executes the steps based on the notification determination program described above.

  In step 331, the processor 35 receives information indicating the target person ID and the current state of the target person from the local unit 30 via the notification IF 33. Next, the processor 35 refers to the target person information DB 6 based on the target person ID, thereby acquiring the notification target state stored in correspondence with the target person ID (step 332). Next, the processor 35 collates the notification target state with the current state received at step 331 (step 333).

  When the notification target state matches the current state (Yes in step 334), the processor 35 sends the current call of the subject indicated by the subject ID to the nurse call system 4 via the nurse call IF 23. It is notified that the state is the state to be notified (step 335). Thereafter, the processor 35 ends the process for the notification received in Step 331.

  On the other hand, when the state to be notified does not match the current state (negative determination at step 334), the processor 35 does not notify the nurse call system 4 but performs the process for the notification received at step 331. finish.

  Such processing is performed by the processor 35 of the center unit 32, so that different criteria can be applied to each subject person to notify the nurse call system 4 of the state of the subject person P. In addition, the notification target state for each target person is collectively stored in the target person information DB 6. Therefore, the criteria for determining whether or not to notify the state of the subject person P by manipulating the information stored in the subject person information DB 6 is matched to the content of care required for each subject person. Can be changed.

  Further, the center unit 32 can share the process of determining the current state of the subject person P. In this case, the state determination program 27 is stored in the memory 34 of the center unit 32. Further, the processor 24 of the local unit 30 detects the moving direction of the subject person P and the body direction of the subject person P by performing processing based on the image recognition program 26. Then, the information including the moving direction of the target person P and the body direction of the target person P and the target person ID are notified to the center unit 32 via the notification IF 31. The processor 35 that has read the state determination program 27 receives the subject P indicated by the subject ID based on the moving direction of the subject P, the body orientation of the subject P, and the past state received via the notification IF 33. Determine the current state of.

  Similarly, the process of detecting the moving direction and the body direction of the subject person P can be shared by the center unit 32. In this case, each local unit 30 transmits an image captured by the camera 2 to the center unit 32. Then, the processor 35 of the center unit 32 collectively performs an image recognition process on an image obtained by capturing each target person, a process for determining the current state of the target person, and a process for outputting the current state of the target person. Is processed.

  The center unit 32 can also be realized using a cloud service.

  In addition, the image input into the state detection apparatus of this indication may be the image imaged from the direction which cross | intersects the longitudinal direction of a bed. Moreover, the output of another sensor etc. can also be taken in into the judgment criteria of the state detection apparatus of this indication. For example, the detection of the weight by a pressure sensor installed at the end of the bed or the like may be added to the condition for the state transition from the rising state to the leaving state. Thereby, the subject's bed leaving state can be detected with higher accuracy.

As described above, the following items are further disclosed with respect to the embodiment described with reference to FIGS.
(Appendix 1) An image processing unit that detects a moving direction of the subject by analyzing an image of the subject on the bed;
A discriminating unit for discriminating the current state of the subject based on the moving direction detected by the image processing unit and the past state of the subject;
An output unit that outputs information indicating the current state of the subject obtained by the determination unit;
A state detection device comprising:
(Supplementary Note 2) In the state detection device described in Supplementary Note 1,
The past state and the current state are: a landing state where the subject lies on the bed; a wake-up state where the subject wakes up on the bed; and A bed leaving state at the end of the bed, and a absent state in which the subject is away from the bed,
The said discrimination | determination part discriminate | determines the present state of the said subject using the different reference | standard for every said multiple types of said past states. The state detection apparatus characterized by the above-mentioned.
(Supplementary Note 3) In the state detection device according to Supplementary Note 1 or Supplementary Note 2,
The image processing unit determines the orientation of the subject's body based on the shape of the subject's image included in the image,
The determination unit determines a current state of the subject based on the moving direction obtained by the image processing unit, the orientation of the subject's body, and the past state. apparatus.
(Supplementary Note 4) In the state detection device described in Supplementary Note 2,
The reference different for each of the past states of the plurality of types is whether or not the moving direction of the subject corresponding to a state transition starting from the past state is detected. .
(Supplementary note 5) In the state detection device according to any one of supplementary notes 1 to 4,
The image of the subject is
The state detection device, wherein the subject is an image obtained by capturing an image of the subject with a camera arranged with an optical axis aligned with a longitudinal direction of the bed.
(Appendix 6) In the state detection device according to any one of Appendix 1 to Appendix 4,
The image of the subject is
It is a difference image generated from an image obtained by capturing the subject illuminated with light in the near-infrared region every other frame with a camera arranged with the optical axis aligned with the longitudinal direction of the bed. Condition detection device.
(Appendix 7) Receiving an image of the subject on the bed,
By detecting the moving direction of the subject by analyzing the image,
Based on the detected moving direction and the past state of the subject, determine the current state of the subject,
Outputting information indicating the current state of the subject obtained by the discrimination. A state detection method.

DESCRIPTION OF SYMBOLS 1 ... Bed; 2 ... Camera; 3 ... Illumination part; 4 ... Nurse call system; 5 ... Illumination control part; 6 ... Subject database (DB); 10, 20 ... State detection apparatus; Discrimination unit; 14 ... Output unit; 15 ... Notification processing unit; 16 ... Speaker; 21 ... Camera interface (IF); 23 ... Nurse call interface (IF); 24, 35 ... Processor; 25, 34 ... Memory; Recognition program; 27 ... Status determination program; 28 ... Notification determination program; 30 ... Local unit; 31, 33 ... Notification interface (IF); 32 ... Center unit; 36 ... Database interface (DBIF)

Claims (9)

Analyzing a plurality of images captured by the imaging device including the bed in the imaging range, whether the moving direction of the subject included in the image is the longitudinal direction or the lateral direction of the bed , An image processing unit for detecting the orientation of the body ;
A determination unit that determines the current state of the subject based on the moving direction detected by the image processing unit, the orientation of the subject's body, and the past state of the subject;
An output unit that outputs information indicating the current state of the subject obtained by the determination unit;
A state detection device comprising: The state detection device according to claim 1,
The image processing unit further detects whether or not the moving direction of the subject is a direction outward from the bed,
The discriminating unit includes a moving direction indicating either a longitudinal direction or a lateral direction detected by the image processing unit, a moving direction indicating whether the direction is outward from the bed, and the subject. A state detection device that determines a current state of the subject based on a combination of a body orientation of the subject and a past state of the subject. In the state detection device according to claim 1 or 2,
The past state and the current state are: a landing state where the subject lies on the bed; a wake-up state where the subject wakes up on the bed; and A bed leaving state at the end of the bed, and a absent state in which the subject is away from the bed,
The said discrimination | determination part discriminate | determines the present state of the said subject using the said different combination for every said multiple types of said past states. The state detection apparatus characterized by the above-mentioned. In the state detection device according to any one of claims 1 to 3,
Wherein the image processing unit, the orientation of the subject's body, the state detection device, wherein the determining based on the shape of the subject image contained in the image. In the state detection device according to claim 3,
The combination different for each of the plurality of types of past states is whether or not the moving direction of the subject corresponding to a state transition starting from the past state is detected. . Analyzing a plurality of images captured by the imaging device including the bed in the imaging range, whether the moving direction of the subject included in the image is the longitudinal direction or the lateral direction of the bed , Detect the direction of the body ,
Based on the detected moving direction, the direction of the subject's body, and the past state of the subject, determine the current state of the subject,
Outputting information indicating the current state of the subject obtained by the discrimination. A state detection method. In the state detection method of Claim 6,
Further, it is detected whether or not the moving direction of the subject is a direction toward the outside from the bed,
A detected moving direction indicating either the longitudinal direction or the lateral direction, a moving direction indicating whether the direction is from the bed or not, a direction of the subject's body, and the subject 's body A state detection method comprising: determining a current state of the target person based on a combination with a past state. Analyzing a plurality of images captured by the imaging device including the bed in the imaging range, whether the moving direction of the subject included in the image is the longitudinal direction or the lateral direction of the bed , An image processing step for detecting the orientation of the body ;
A determination step of determining the current state of the subject based on the moving direction detected by the image processing step, the orientation of the subject's body, and the past state of the subject;
An output step of outputting information indicating the current state of the subject obtained by the determination step;
A state detection program for causing a computer to execute. In the state detection program according to claim 8,
The image processing step further detects whether or not the moving direction of the subject is a direction toward the outside from the bed,
The determining step includes a moving direction indicating either a longitudinal direction or a lateral direction detected by the image processing step, a moving direction indicating whether the direction is outward from the bed, and the subject. A state detection program for determining a current state of the subject based on a combination of a body orientation of the subject and a past state of the subject.

Images