TW201539388A - Suicide prevention notification system - Google Patents

Abstract

A suicide prevention notification system comprises an embedded video camera mounted above the window (door) places to continuously video-record place and human body images to determine behavior mode of human. Moreover, the embedded video camera has a signal output unit. When the behavior mode of human is established, the signal output unit will issue notification signals. Accordingly, if a person is not a suicide person, an alert is disarmed to disable the signal output unit. If a person suicides, the alert is activated to further perform determination for abnormal motions of human body images. When the motions of human body images matches the behavior mode preset by the memory, the signal output unit is enabled to immediately issue a notification signal to notify of the suicide incident.

Description

Anti-suicide notification system

The present invention relates to an anti-suicide notification system that intelligently judges a suicide event and can notify a suicide event to a predetermined alarm platform.

According to the fact that suicides account for a high proportion of deaths, and suicides are difficult to prevent. They can only use passive methods to carry out remedial measures. They do not actively detect them for prevention. The schematic diagram of RFID used in Figure 1 is shown. The utility model comprises: a radio frequency identification tag 10, wherein the suicide person wears a tag, and the content of the tag records relevant information of the suicide person, such as age, name, type of diagnosis and condition, etc.; and the place where the person who suicide is traced 11 The site 11 is usually a mental hospital, an annuity home or a prison, and is controlled by a mentally ill person, an elderly person or a prisoner in a closed state, and a mentally ill, an elderly person or a prisoner is liable to become a suicide person, and can be based on the place 11 and related materials. Passively inferring the possibility of suicide, it is impossible to actively detect whether it is a suicide event, and RFID is a prior art, and its principle is not described.

However, in the prior art (Prior Art), it is still generally the following problems: (1). Because the suicide person knows the RFID that cannot be transmitted through the behavior pattern of the place 11, the suicide person The behavioral pattern of the presence 11 cannot determine whether it is a suicide incident. It can only infer a suicide incident to carry out remedial measures to reduce the degree of danger, and cannot prevent it in advance. (2). Because the suicide incident is inferred rather than judged, the error rate of reporting suicides is increased, which not only does not help the handling of suicides, but also reduces the reliability of the monitors in reporting suicides.

The present inventors have actively studied and improved in view of the above problems, and have overcome the shortcomings thereof, and have been the subject of the creation of the present invention.

The main purpose of the present invention is to provide an anti-suicide notification system for solving the behavior pattern that cannot be discriminated by the prior art, and can utilize the behavior pattern of the person to the window (door) to intelligently judge the suicide event, thereby having The effect of preventing suicide beforehand.

Another object of the present invention is to provide an anti-suicide notification system for intelligently determining suicide events, reducing the error rate of reporting suicides, and further enhancing the reliability of reporting suicide events.

In order to achieve the above objective, the technical means adopted by the present invention comprises: an embedded camera mounted on a window (door) location, the embedded camera comprising: a capture image unit, which continuously records One of the locations monitors the image, the surveillance image is composed of a plurality of consecutive images; a motion image detection and control unit, the motion image detection and control unit obtains the surveillance image by the captured image unit, and the memory thereof Pre-storing the feature vector of the window (door) and the feature vector of the at least one human body image, so that the motion image detection and control unit detects the monitoring by analyzing the difference between the plurality of consecutive pictures a moving image in the image corresponding to the feature vector of the window (door) of the memory and the feature vector of the human body image, by calculating the feature vector of the window (door) and the feature vector of the human image a change in the behavior of the discriminating person; and a signal output unit electrically coupled to the motion image detection and control unit when the dynamic After the image detection and control unit determines that the behavior mode of the person is established, the signal output unit will send a notification signal; thereby, the captured image unit can continuously capture the monitored image and transmit it to the motion image detection and control unit. The motion picture detection and control unit system analyzes the difference between the front and back images, calculates the vector of the front and back images, determines whether there is a moving object, and compares the moving object with the human body image in the memory. In order to determine whether the moving object is a human, and then according to the continuous moving direction of the moving object to determine the behavior mode, if the person is a non-suicidal person, the guard is released, so that the signal output unit does not start, and the person who is suicidal Then enter the alert to further determine the abnormal motion of the human body image. If the motion of the human body image conforms to the behavior mode preset by the memory, the signal output unit is activated, and a notification signal is immediately sent to notify the suicide event. .

According to the foregoing feature, the dynamic image detecting and controlling unit can be composed of a digital signal processing chip and an embedded processor; the signal output unit transmits the notification signal by the following means, including: cable, radio signal, cable One of the network, wireless network and Bluetooth signal.

According to the foregoing feature, an alarm platform is further provided, and the alarm platform receives the notification signal sent by the signal output unit through an internet or a regional network.

According to the pre-existing feature, the behavior mode may be one of the neck down quickly, the suicidal intention, the action is out of time or staying at the door for more than the set time.

According to the foregoing feature, in a feasible embodiment, the feature vector of the window (door) and the feature vector of the human body image can detect the feature vector pair of the human body image through a time-varying model preset by the moving image. The feature vector of the window (door) location produces a change in the velocity of the velocity, while determining the behavior pattern of the neck rapidly downward.

In another possible embodiment, the feature vector of the window (door) and the feature vector of the human body image pass through a time-varying model and a spatial model preset by the moving image, and the spatial model sets a skeleton length to short axis ratio of the human body image. The angle between the length of the skeleton and the length of the skeleton, and the change time of the ratio of the length of the skeleton to the length of the skeleton and the angle between the length of the skeleton and the axis of the skeleton can be detected, and the feature vector of the human body image can be detected to generate time for the feature vector of the window (door) The length of the skeleton changes, and the behavior pattern of suicidal intention is judged.

In another feasible embodiment, the feature vector of the window (door) and the feature vector of the human body image can be detected by the spatial model preset by the moving image, and the feature vector of the human body image can be detected to the window (door) location. The eigenvectors produce interfering contact surface changes and determine behavioral patterns that are out of date.

In another possible embodiment, the feature vector of the window (door) and the feature vector of the human body image pass through a time-varying model and a spatial model preset by the moving image, and the spatial model sets a feature vector of the human body image. The starting parameter of the feature vector of the window (door), the moving track to the stopping parameter, and setting the starting parameter, the moving track to the changing time of the stopping parameter by the time varying model, and detecting the feature vector of the body image The feature vector of the window (door) location produces a change in the movement trajectory of the time, and determines the behavior pattern that stays at the door edge for more than the set time.

By means of the above-mentioned technical means, the invention adopts an anti-suicide notification system, and can also use the human body image to intelligently judge the suicide event in the window (door) behavior mode, which not only has the prevention of suicide beforehand, but also reduces the notification suicide. The error rate of the incident and the reliability of the suicide reported by the alert platform.

First, referring to FIG. 2, FIG. 2A and FIG. 3, a preferred embodiment of the anti-suicide notification system of the present invention comprises: an embedded camera 20 (Embedded Camera) having an embedded system, the embedding The camera 20 is mounted above a window (door) location 30. The embedded camera 20 includes a capture image unit 21 that continuously records a surveillance image 40 of the location. The surveillance image 40 is continuous by a plurality of images. The image 41 is composed of a moving image detecting and controlling unit 22, and the moving image detecting and controlling unit 22 obtains the monitoring image 40 by the captured image unit 21, and the window 23 is pre-stored in the memory 23. The feature vector of the location 30 and the feature vector of the at least one human body image enable the motion image detection and control unit 22 to detect a motion image 42 in the surveillance image by analyzing the difference between the plurality of consecutive images 41. The motion image 42 is generated corresponding to the feature vector of the window (door) location 30 of the memory 23 and the feature vector of the human body image, by calculating the feature vector of the window (door) location 30 and the feature vector of the human body image. The change, for example: the foregoing feature vector can be a common object description manner such as object size, skeleton, direction, texture, color, geometric shape, etc., and can also determine the average behavioral brightness, chromaticity, etc. of the image, and determine the behavior pattern of the person. In this embodiment, the dynamic image detection and control unit 22 can be a digital signal processing chip and an embedded processor; and a signal output unit 24 electrically coupled to the motion image detection and control unit 22 After the motion picture detection and control unit 22 determines that the behavior mode of the person is established, the signal output unit 24 will send a notification signal. In this embodiment, the signal output unit 24 transmits the notification signal by using the following methods, including : One of cable, radio, cable, wireless, and Bluetooth signals.

In a preferred embodiment, the embedded camera 20 captures the light and brightness conditions of the scene, and adjusts the parameters of the shooting and analysis to improve the quality of the monitoring image 40. The digital signal processor performs high-speed operation and cooperates with the calculation program of the change between the feature vector of the window (door) location 30 and the feature vector of the human body image by the embedded processor, by using the digit The signal processor and the embedded processor detect the moving image 42, and the moving image 42 corresponds to the feature vector of the window (door) location 30 of the memory 23 and the feature vector of the human body image, thereby completing the discriminating human behavior pattern. .

The present invention may further include an alarm platform 50 that receives the notification signal sent by the signal output unit 24 via an internet network 60 or a regional network 60A. In this embodiment, the alarm platform 50 may include: A server is electrically connected to a display unit and a warning unit, and the notification signal is transmitted through the Internet 60 or the area network 60A to the server, so that the warning unit notifies a non-suicidal person that a suicide will occur, and The person who commits suicide can be seen in the display unit, but is not limited thereto.

As shown in FIG. 4, the image capturing unit 21 can continuously transmit the captured image 40 to the motion image detecting and controlling unit 22, so that the motion image detecting and controlling unit 22 performs system analysis. The difference between the rear images and the vector of the front and back images is calculated to determine whether there is a moving object, and the moving object is compared with the human body image in the memory 23 to determine whether the moving object is a human, and then move according to the The continuous moving direction of the object changes to determine the behavior mode. If the person is a non-suicidal person, the guard is released, so that the signal output unit 24 does not start, and if the person commits suicide, the patient enters the alert and further abnormal action of the human body image. Judging, if the action of the human body image conforms to the behavior mode set by the memory 23 in advance, the signal output unit 24 is activated to immediately send a notification signal to notify the suicide event, and thus, after entering the alert, further Determining a person's behavioral pattern can be one of the neck down quickly, the suicidal intention, the action is out of date, or staying at the door for more than a set time.

In a possible embodiment, the feature vector of the window (door) location 30 and the feature vector of the human body image are detected by the time-varying model preset by the motion image 42 to detect the feature vector of the human body image (the window) ( The feature vector of the door 30 generates a change in the distance of the speed, and determines a behavior pattern in which the neck is rapidly downward. As shown in FIG. 5, the window (door) place 30 has a space place 31 and locks the space place 31. The feature vector, when the human body image shows that the first distance D1 of any height is rapidly reduced to the second distance D2, and can be determined as a suicide person, for example, the suicide mode is established, and the behavior pattern of the neck rapidly downward is established.

In another possible embodiment, the feature vector of the window (door) location 30 and the feature vector of the human body image pass through a time-varying model and a spatial model preset by the motion image 42, and the spatial model sets the skeleton length of the human body image. The axial ratio L 1 / L 2 and the skeleton length and the short axis angle value θ, and the time-varying model is used to set the change time of the skeleton length-to-minor axis ratio L 1 / L 2 and the skeleton long-axis axis angle value θ, and the characteristics of the human body image can be detected. The vector changes the skeleton length of the feature vector of the window (door) location 30, and determines the behavior pattern of the suicidal intention. As shown in FIG. 6A, the human body image is hooked and connected at the window (door) location 30. When a towel or clothing R or the like is found, it can be judged as a suicide person. In addition, for example, the suspending suicide mode or as shown in FIG. 6B, the human body image is prepared for jumping out at the window (door) location 30, and may also be determined as a suicide person, for example, a suicide suicide mode, which produces a skeleton length to short axis ratio. The change in the angle θ between L 1 / L 2 and the length of the skeleton axis at the window (door) location 30 establishes a behavioral pattern of suicidal intention.

In another possible embodiment, the feature vector of the window (door) location 30 and the feature vector of the human body image are transmitted through the spatial model preset by the motion image 42 to detect the feature vector of the human body image (the door) The feature vector of the place 30 generates a change in the interfering contact surface P, and determines an unsuitable behavior mode of the action. As shown in FIG. 7A, the window (door) place 30 has a window (door) 32 and locks the window (door). a feature vector of 32, the human body image contacting the window (door) 32, and may also be identified as a suicide person, for example, a suicide attempt mode or as shown in FIG. 7B, the window (door) location 30 has a ground 33, and The feature vector of the ground 33 is locked, and when the human body image falls on the ground 33 to generate an interference contact surface P, it can also be determined as a suicide person, for example, Kaivas suicide, and an untimely behavior pattern is established.

In another possible embodiment, the feature vector of the window (door) location 30 and the feature vector of the human body image pass through a time-varying model and a spatial model preset by the motion image 42, and the spatial model sets a feature vector of the human body image. The start parameter 331 of the feature vector of the window (door) location 30, the movement trajectory 332 to the stop parameter 333, and the change time of the start parameter 331 and the movement trajectory 332 to the stop parameter 333 are set by the time-varying model. Measuring the feature vector of the human body image to change the movement trajectory of the feature vector of the window (door) location 30, and determining the behavior pattern of staying at the door edge for more than the set time, as shown in FIG. 8, the feature vector of the ground 33 At least the path of the initial parameter 331, the movement trajectory 332, and the stop parameter 333 is provided. When the human body image indicates that the predetermined path set for the ground 33 exceeds the stay time, the person who is suicidal may be identified, for example, melancholy and restless. If you are suing, jumping off, or opening a gas to commit suicide, you will set up a behavior mode where the door stays longer than the set time.

In the above various feasible embodiments, various suicidal behavior patterns may be used, and various suicidal behavior patterns may be concatenated into another new suicide algorithm, for example, at the corner of the door or on the hook, hanging clothes, and the like. Actions can use two different suicidal behavior patterns to concatenate into another suicide behavior pattern. As a result, the new suicide algorithm uses the suicide person to hook or hang to produce the skeleton length change action, and suicide in tandem The person touches the corner of the door or locks the door to create a contact surface, but not limited to this.

Therefore, the objective behavior pattern of the human body image to the window (door) location 30 can not only determine suicide events, but also overlap different behavior patterns or concatenate different behavior patterns to enhance the reliability of the notification suicide event, and also enable monitoring. The suicide model of suicides is determined to adopt different rescue strategies, which can be applied to mental hospitals, nursing homes or prisons to prevent and control the seizure of mental patients, the elderly or prisoners in a closed state.

In summary, the technical means disclosed in the present invention have the invention patents such as "novelty", "progressiveness" and "available for industrial use", and pray for the patent to encourage the invention. German sense.

The drawings and the descriptions of the present invention are merely preferred embodiments of the present invention, and those skilled in the art, which are subject to the spirit of the present invention, should be included in the scope of the patent application.

20‧‧‧Embedded camera
21‧‧‧Image capture unit
22‧‧‧ Motion Picture Detection and Control Unit
23‧‧‧ memory
24‧‧‧Signal output unit
30‧‧‧window (door) place
31‧‧‧ Space places
32‧‧‧Window (door)
33‧‧‧ Ground
331‧‧‧ starting parameters
332‧‧‧moving track
333‧‧‧ stop parameters
40‧‧‧Monitoring images
41‧‧‧ Pictures
42‧‧‧moving
50‧‧‧Alarm platform
60‧‧‧Internet
60A‧‧‧Regional Network
D1‧‧‧First distance
D2‧‧‧Second distance
L1/L2‧‧‧ skeleton long and short axis ratio θ‧‧‧ skeleton long and short axis angle value
P‧‧‧ contact surface
R‧‧‧Towels, clothing

Figure 1 is a schematic diagram of a conventional RFID. Figure 2 is a schematic representation of the invention. 2A is a circuit block diagram of the embedded camera of the present invention. Figure 3 is a schematic illustration of the surveillance image of the present invention. Figure 4 is a flow chart of the present invention. Figure 5 is a schematic illustration of the behavioral mode of the present invention for discriminating the neck rapidly downward. Fig. 6A is a schematic diagram showing the behavior pattern of the present invention for discriminating whether or not suicide is intended. Fig. 6B is a schematic diagram showing another behavior pattern of the present invention for discriminating whether or not suicide is intended. Fig. 7A is a schematic diagram showing the behavior mode of the discriminating action of the present invention. Fig. 7B is a schematic view showing another behavior mode in which the discriminating action of the present invention is out of time. Fig. 8 is a schematic view showing the mode in which the discriminating window (door) of the present invention stays longer than the set time line.

20‧‧‧ camera

30‧‧‧window (door) place

40‧‧‧Monitoring images

41‧‧‧ Pictures

42‧‧‧moving

50‧‧‧Alarm platform

60‧‧‧Internet

60A‧‧‧Regional Network

Claims (9)

An anti-suicide notification system includes: an embedded camera mounted on a window (door) location, the embedded camera comprising: a capture image unit that continuously records one of the locations to monitor the image The surveillance image is composed of a plurality of consecutive images; a motion image detection and control unit, the motion image detection and control unit obtains the surveillance image by the captured image unit, and the window is pre-stored in the memory The feature vector of the (door) location and the feature vector of the at least one human body image enable the motion image detection and control unit to detect a motion image in the surveillance image by analyzing a difference between the plurality of consecutive images. The moving image is corresponding to the feature vector of the window (door) of the memory and the feature vector of the human body image, by calculating the change between the feature vector of the window (door) and the feature vector of the human body image. Determining a human behavior mode; and a signal output unit electrically coupled to the motion image detection and control unit when the motion image detection After the control unit determines that the behavior mode of the person is established, the signal output unit sends a notification signal; thereby, the captured image unit can continuously transmit the captured image to the motion image detection and control unit, so that the dynamic The image detection and control unit system analyzes the difference between the front and back images, calculates the vector of the front and back images, determines whether there is a moving object, and compares the moving object with the human body image in the memory to determine the movement. Whether the object is a person, and then according to the continuous moving direction of the moving object to determine the behavior mode, if the person is a non-suicidal person, the guard is released, so that the signal output unit does not start, and the person enters the alert if the person commits suicide. Further, the abnormal motion of the human body image is determined. If the motion of the human body image conforms to the behavior mode preset by the memory, the signal output unit is activated, and a notification signal is immediately sent to notify the suicide event. The anti-suicide notification system of claim 1, wherein the motion image detection and control unit is a digital signal processing chip and an embedded processor. The anti-suicide notification system of claim 1, wherein the signal output unit transmits the notification signal by: a cable, a radio signal, a wired network, a wireless network, and a Bluetooth signal. one. The anti-suicide notification system of claim 3, further comprising an alarm platform that receives the notification signal sent by the signal output unit via an internet or regional network. The anti-suicide notification system according to claim 1, wherein the behavior mode may be one of a neck down rapidly, a suicidal intention, an inappropriate action, or staying at the door for more than a set time. The anti-suicide notification system according to claim 6, wherein the feature vector of the window (door) and the feature vector of the human body image can detect the human body through a time-varying model preset by the moving image. The feature vector of the image produces a velocity change in the feature vector of the window (door) location, and the behavior pattern of the neck is determined to be rapidly downward. The anti-suicide notification system according to claim 6, wherein the feature vector of the window (door) and the feature vector of the human body image pass through the time-varying model and the space model preset by the moving image, the space The model sets the ratio of the length of the skeleton of the human body image to the angle between the length of the skeleton and the length of the skeleton, and sets the change time of the ratio of the length of the skeleton to the length of the skeleton and the angle between the length of the skeleton, and detects the characteristic vector of the human body image. The feature vector of the (door) site generates a change in the length of the skeleton of the time, and determines the behavior pattern of the suicidal intention. The anti-suicide notification system of claim 6, wherein the feature vector of the window (door) and the feature vector of the human body image can detect the human body image through a spatial model preset by the moving image. The feature vector changes the contact surface of the feature vector of the window (door) location, and determines the behavior mode of the action that is out of time. The anti-suicide notification system according to claim 6, wherein the feature vector of the window (door) and the feature vector of the human body image pass through the time-varying model and the space model preset by the moving image, the space The model sets a starting parameter of the feature vector of the human body image at the window (door), a moving track to a stopping parameter, and sets a starting parameter and a moving track to a stop parameter change time by the time varying model. The feature vector of the human body image is detected to change the movement trajectory of the feature vector of the window (door) location, and the behavior mode of staying at the door edge for more than the set time is determined.