System and method for detecting falling behaviors from monitoring video
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A technology for monitoring video and behavior, applied in instruments, biological neural network models, calculations, etc., can solve problems such as additional wearable devices, additional installation of monitoring equipment, and low detection accuracy
Active Publication Date: 2021-01-12
CHENGDU UNIV OF INFORMATION TECH
View PDF11 Cites 6 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0005] Aiming at the above-mentioned deficiencies in the prior art, the present invention provides a system and method for detecting fall behavior from surveillance video, which solves the problem that the fall detection system in the prior art requires additional wearable equipment, the detection accuracy is not high, and additional monitoring is required. Equipment defects and deficiencies
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0085] like figure 1 As shown, the present invention provides a kind of system that detects falling behavior from surveillance video, comprises key frame extraction module, pedestrian detection module, human body key point detection module, continuous gesture feature encoding module and output module connected in sequence; Key frame extraction The module is used to obtain the monitoring video stream and extract the video frame; the pedestrian detection module is used to construct the pedestrian detection model and set the detection frame interval in the pedestrian detection model. According to the image of each video frame, the pedestrian detection model is used to monitor the picture Pedestrians in the pedestrian area are divided into sub-graphs of the human body area; the human body key point detection module is used to construct a key point detection model, and use the key point detection model to extract the human body key points of pedestrians from the human body area sub-...
Embodiment 2
[0088] Example 2, such as figure 2 As shown, the present invention provides a kind of method that detects falling behavior from monitoring video, comprises the following steps:
[0089] S1. Obtain a surveillance video stream and extract video frames;
[0090] S2, build pedestrian detection model, its implementation method is as follows:
[0091] S201, labeling image data for pedestrian detection training;
[0092] S202. According to the image data, based on the YOLOv3-tiny model, set the input image size to 448×448, the output category to 1, and the maximum training batches to 6000;
[0093] S203, using the pre-trained network parameters of the Darknet network to initialize the YOLOv3-tiny model;
[0094] S204, using the transfer learning method and the pedestrian detection data to train the YOLOv3-tiny model to obtain the pedestrian detection model;
[0095] S3, setting the detection frame interval in the pedestrian detection model, and according to the image of each video...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
PUM
Login to view more
Abstract
The invention, which belongs to the technical field of video monitoring, provides a system and method for detecting a falling behavior from a monitoring video. The system comprises a key frame extraction module, a pedestrian detection module, a human body key point detection module, a continuous attitude feature coding module and an output module that are connected successively. The method for detecting the falling behavior from the monitoring video comprises four parts of pedestrian detection from a monitoring picture, human body key point detection and attitude estimation, human body attitude image feature coding in a continuous time domain and a falling behavior classifier based on a CNN convolutional neural network. The detection system overcomes the defects that a falling behavior detection system in the prior art needs extra wearable equipment, is low in detection precision and needs extra monitoring equipment, is low in cost, does not need extra monitoring equipment, analyzes action behaviors of a human body from a video shot by an existing common monitoring camera, and detects whether a falling behavior occurs is detected.
Description
technical field [0001] The invention belongs to the technical field of video monitoring, and in particular relates to a system and method for detecting a fall behavior from a monitoring video. Background technique [0002] A fall is defined as a sudden change in body position, an involuntary fall to the ground or a lower level. Falling behavior is also a very common accident in daily life. It has a very high probability of occurring in various groups of people and may cause some fatal injuries, such as cerebral hemorrhage and fractures. If not treated in time, Permanent injury, even death, will result to the person who falls. At present, the methods of fall behavior detection mainly include three categories: wearable device-based fall detection systems, acoustic-based fall detection systems, and video surveillance-based fall detection systems. [0003] Fall detection systems based on wearable devices usually require people to wear a device equipped with motion sensors, suc...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to view more 
Login to view more