Multi-camera video cross-regional human motion pose target recognition method
A human body movement and multi-camera technology, applied in character and pattern recognition, image analysis, image enhancement, etc., can solve problems such as difficult to accurately identify human moving objects
Active Publication Date: 2021-07-06
王连圭
View PDF8 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, in a complex environment, it is difficult to accurately identify human moving objects due to the long-term occlusion of moving objects.
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0083] Refer figure 1 The present invention discloses a multi-camera video cross-regional human motion attitude target identification method, and the method is specifically as follows:
[0084] I. Collect multi-frame video images across regional human motion targets from intelligent monitoring systems, and block images, and construct of target feature matrices.
[0085] Second, the compressed perception is sparse: obtaining the observed value matrix by calculating the sparse representation coefficient of the compression particle observation value, realizing the effect of data compression, improving tracking accuracy; solving the combined sparse optimization model and alternating update: adopt adaptive threshold The alternate parameter generating algorithm is used to implement the parameters alternately update, reduce the storage space and accelerate the convergence speed.
[0086] Joint sparse optimization model
[0087]
[0088] For example, the first video frame image is the g...
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 discloses a multi-camera video cross-regional human motion posture target recognition method, and for the multi-camera video cross-regional human motion posture target recognition problem, a joint sparse optimization algorithm based on compressed sensing is proposed: firstly, a joint sparse optimization model is given , Obtain the observed value matrix by calculating the sparse representation coefficient of the compressed particle observed value matrix: Secondly, an image reconstruction based on compressed sensing is proposed, and an adaptive threshold value alternate iterative parameter reconstruction algorithm is used to adaptively select the threshold value to iterate the image in the iterative process Reconstruction, the original image is reconstructed from the observation matrix. However, when the moving target is seriously occluded in a complex monitoring scene, it is difficult to achieve accurate tracking of the moving target: therefore, a background subtraction method based on the optimal solution is proposed, and the binarized image is used to identify the moving target of the human body.
Description
Technical field [0001] The present invention relates to a motion attitude target recognition method, particularly a multi-camera video cross-regional human motion attitude target identification method. Background technique [0002] Multi-camera video cross-region identifying human body motion attitude has become a research hotspot and frontier topics. During the actual video surveillance process, the posture of the moving target will change at any time. Since the target motion is subject to noise interference, light changes, especially when the video surveillance scene is blocked, so that the human motion attitude target is accurately tracked It is difficult to get difficult. Therefore, in-depth research of the topic is of great theoretical significance. In addition, the subject has a wide range of application prospects in many ways. In terms of transportation, a wide range of continuous tracking of multiple cameras is used to make a large-scale continuous tracking of accidents o...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/62G06T7/11G06T7/136G06T7/194G06T7/246
CPCG06T7/11G06T7/136G06T7/194G06T7/246G06T2207/10016G06T2207/30196G06V40/20G06V10/513G06F18/2136
Inventor 王连圭
Owner 王连圭