TWM463402U - Falling down detection system with automatic tracking - Google Patents

Description

Automatic tracking tracking detection system

The present invention relates to a human motion detection system, and more particularly to a fall detection system with automatic tracking.

Taiwan entered an aging society in the 82nd year of the Republic of China, and the aging is accompanied by the problem of elderly care. The statistics of the Health Department of the Executive Yuan show that the cause of death of the elderly over 65 years old in Taiwan is injured. Ranked seventh, the fall is the second leading cause of death in accident injuries. It can be seen that the fall is a fatal injury to the elderly. The fall of the elderly often leads to impaired physical function and loss of activities of daily living. If the fall event is detected early And deal with it immediately, you can reduce the damage caused by the fall.

A fall detection system using an accelerometer and a gyroscope is an Arduino development board that is equipped with a three-axis accelerometer and a dual-axis gyroscope to determine whether the user has fallen. Since the above-mentioned fall detection system needs to wear a sensor on the subject of the subject, it is a burden for the elderly, and in addition to the trouble of wearing and taking off, it is easy to lose the sensor or forget to wear it. Sensor problem.

Therefore, the purpose of the present invention is to provide a fall detection system with automatic tracking.

Therefore, the novel fall detection system with automatic tracking includes a rotating stage, a depth sensing unit, and a processing unit.

The depth sensing unit is disposed on the rotating stage, and can obtain a majority depth image for sensing a space in which at least one subject exists.

The processing unit is electrically connected to the rotating stage and the depth sensing unit, and has a built-in judgment program.

The processing unit identifies the position of the subject in the depth image, and combines the positions of the subject in each depth image in time series to generate a moving path, and the processing unit performs the moving path according to the moving path. For the rotation control of the rotary stage, the depth sensing unit is rotated horizontally so that the optimal sensing range of the depth sensing unit always covers the subject.

The processing unit performs skeleton recognition processing on the subject in the depth images to obtain a plurality of skeleton images respectively corresponding to the depth images, and the processing unit sequentially processes the skeleton images to obtain The displacement change amount of each node of the skeleton in each skeleton image, and the displacement change amount of each node is used to identify the fall motion of the subject via the determination program.

Preferably, the depth sensing unit comprises an infrared emitter and an infrared camera. The infrared emitter emits infrared rays to the space, and the infrared camera receives the infrared rays reflected by the objects in the space to obtain the depth images.

Preferably, the automatic tracking fall detection system further comprises A notification unit electrically connected to the processing unit, when the processing unit recognizes that the subject has fallen, controlling the notification unit to issue the fall information.

The utility model has the advantages that the depth image is obtained by the depth sensing unit, and after the image processing of the processing unit, the automatic tracking control of the subject and the fall of the subject are recognized. Automatically track the effect of the subject and fall detection.

1‧‧‧Rotary stage

2‧‧‧Deep sensing unit

21‧‧‧Infrared emitter

22‧‧‧Infrared camera

3‧‧‧Processing unit

31‧‧‧Judgement program

4‧‧‧Notification unit

Other features and effects of the present invention will be apparent from the following detailed description of the preferred embodiments, wherein: FIG. 1 is a schematic diagram illustrating a preferred embodiment of the novel automatic tracking tracking detection system. For example, a depth sensing unit and a rotating stage are disposed, and depth sensing is performed on a space in which a subject exists; FIG. 2 is an electrical connection relationship of each component of the preferred embodiment; A schematic diagram of a depth image and a skeleton image; and FIG. 4 is a diagram illustrating that the subject has three corresponding bending positions when the subject is bent, squatted, and fallen, respectively. The amount of displacement change in the underarm and fall movements.

Referring to FIG. 1 and FIG. 2, a preferred embodiment of the automatic tracking and falling detection system of the present invention comprises a rotating stage 1, a depth sensing unit 2, a processing unit 3, and a notification unit 4. In the preferred embodiment, the kinect system designed and developed by Microsoft is used. The hardware and software of the kinect system are the depth sensing unit 2 and a part of the processing unit 3, respectively.

The depth sensing unit 2 is disposed on the rotating stage 1 and A space 9 in which a subject 8 exists continues to perform sensing to obtain a plurality of depth images (each depth image is as shown in the left image of FIG. 3). The depth sensing unit 2 includes an infrared emitter 21 and an infrared camera 22. The infrared ray emitter 21 emits infrared rays to the space 9, and the infrared ray camera 22 receives the infrared ray reflected by the object in the space 9 to obtain the depth images.

The processing unit 3 is electrically connected to the rotating stage 1 and the depth sensing unit 2, and a determination program 31 is built in. The processing unit 3 simultaneously performs arithmetic processing of automatic tracking and fall detection.

In the processing of the automatic tracking, the processing unit 3 searches for the moving object that may be the subject 8 in the depth image according to the depth value information of each depth image, and determines the determined program 31 through the discriminable body part. Whether the moving object is the subject 8, thereby discriminating the position of the subject 8. The processing unit 3 further combines the positions of the subjects 8 in each depth image in time series to generate a moving path. The processing unit 3 performs rotation control on the rotating stage 1 according to the moving path, and rotates the depth sensing unit 2 horizontally to make the optimal sensing range of the depth sensing unit 2 always cover the subject. 8, to achieve automatic tracking of the function of the subject 8.

In the processing of the fall detection, the processing unit 3 performs skeleton recognition processing on the subject 8 in the depth images to obtain a plurality of skeleton images respectively corresponding to the depth images (each skeleton map) As shown in the right image of FIG. 3, the processing unit 3 sequentially processes the skeleton images to obtain the displacement change amount of each node of the skeleton in each skeleton image, At the same time, the displacement change amount of each node is subjected to the fall detection operation of the subject 8 via the determination program 31.

The position recognition and skeleton recognition processing of the subject 8 described above are all processed by the software of the kinect system.

However, the judgment of the fall action of the subject 8 is based on the comparison of the displacement change of the skeleton node caused by the continuous daily recording of the subject's daily home movement, wherein the main purpose is to sit similar to the fall. The lower, the bending and the underarm are analyzed, and a fall action judgment is established in the judgment program 31. Since the amount of displacement of the head node in the skeleton is the largest, the displacement amount of the head node in the skeleton is taken as an example. As shown in FIG. 4, the displacement of the head node when the subject 8 falls is obvious. It is different from the amount of displacement of the head node when bending and kneeling.

Referring to FIG. 2, the notification unit 4 is electrically connected to the processing unit 3. When the processing unit 3 recognizes that the subject 8 has fallen, the notification unit 4 is controlled to issue the fall information. In the preferred embodiment, the notification unit 4 sends an alarm sound to notify other people outside the space 9, and sends a message to the relatives and friends of the subject 8 to inform the subject 8 that the user has fallen. The shortest time to deal with the fall event.

In summary, the fall detection system with automatic tracking of the present invention obtains a plurality of depth images by the depth sensing unit 2 and performs automatic tracking control of the subject 8 after the image processing by the processing unit 3 And the fall of the subject 8 is identified to achieve automatic tracking of the subject 8 and the subject's fall detection. The above fall detection method not only makes the subject 8 easy to use without wearing any sensor, and the depth The image captured by the degree sensing unit 2 is a depth image, and the conventional camera does not clearly record the appearance privacy problem of the subject. In addition, the notification unit 4 recognizes the subject 8 in the processing unit 2 When a fall action occurs, the fall information can be sent out in time to inform others to make a timely fall event, so it is indeed possible to achieve the purpose of the novel.

However, the above is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, that is, the simple equivalent changes and modifications made in accordance with the scope of the present patent application and the contents of the patent specification, All remain within the scope of this new patent.

1‧‧‧Rotary stage

2‧‧‧Deep sensing unit

21‧‧‧Infrared emitter

22‧‧‧Infrared camera

8‧‧‧ Subjects

9‧‧‧ Space

Claims (3)

A fall detection system with automatic tracking includes: a rotating stage; a depth sensing unit disposed on the rotating stage, and capable of sensing a space in which at least one subject exists to obtain a majority depth And a processing unit electrically connecting the rotating stage and the depth sensing unit, and having a determination program built therein; wherein the processing unit identifies the position of the subject in the depth images, and The position of the subject in each depth image is combined in time series to generate a moving path, and the processing unit performs rotation control on the rotating stage according to the moving path to horizontally rotate the depth sensing unit to make the depth The optimal sensing range of the sensing unit is always covering the subject; wherein the processing unit performs skeleton identification processing on the subject in the depth images to obtain a plurality of images respectively corresponding to the depth images The skeleton image, the processing unit sequentially processes the skeleton images to obtain the displacement change amount of each node of the skeleton in each skeleton image, and simultaneously changes the displacement of each node. This operation falling determination program for recognition by the test subject. The automatic tracking tracking detection system according to claim 1, wherein the depth sensing unit comprises an infrared emitter, and an infrared camera, the infrared emitter emits infrared light for the space, and the infrared camera receives the space The infrared light reflected by the object Get these depth images. The automatic tracking tracking detection system according to claim 1 or 2, further comprising a notification unit electrically connected to the processing unit, when the processing unit recognizes that the subject has fallen, controlling the notification unit to outward Issue the fall message.