A Pedestrian Counting Method
A pedestrian counting and pedestrian technology, applied in computing, computer parts, instruments, etc., can solve problems such as reduced accuracy, large amount of calculation, and no consideration of occlusion.
Inactive Publication Date: 2016-06-01
SHANGHAI JIAOTONG UNIV
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, pedestrian counting methods based on pedestrian detection usually need to detect all the pedestrians in the scene for counting (see: BoWu, R. Nevatia. "Detection of multiple, partially included humans in single image by Bayesian combination of edgelet part detectors," ICCV2005, Tenth IEEE International Conference on, 1:90–97, 2005.) Therefore, it has a large amount of calculation and relatively poor real-time performance, so it cannot be directly applied to real-time monitoring systems.
The traditional pedestrian counting method based on pixel statistics (see: Lee, G.G., Kim, B.S., Kim, W.Y. "Automatic Estimation of Pedestrian Flow." ACM / IEEE International Conference on Distributed Smart Cameras, 291–296, 2007.) obtains the number of pedestrians by statistically calculating pedestrian characteristics, This avoids the pedestrian detection process and reduces the complexity of the calculation, but it does not consider the occlusion situation, resulting in a decrease in the accuracy of counting when pedestrian occlusion occurs
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
[0042] The video sequence used in this embodiment is a public security monitoring scene sequence.
[0043] The method for counting pedestrians involved in this embodiment includes the following specific steps:
[0044] Step 1: Carry out mixed Gaussian background modeling on the surveillance video to obtain pedestrian foreground images, such as figure 2 shown.
[0045] The second step: extract pedestrian optical flow features, such as image 3 shown.
[0046] Step 3: Normalize pedestrian foreground pixel features. The specific steps are: use pedestrian ellipse modeling, that is, each pedestrian is mapped into an ellipse according to the size of the area according to the ellipse mapping rule, and the area of the ellipse is regarded as the pedestrian e iThe area in the current camera view, denoted as area(e i ), for each foreground pixel belonging to the pedestrian, its corresponding normalization coefficient w i =1 / area(e i ),Such as Figure 4 shown.
[0047] The fou...
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 pedestrian counting method in the technical field of computer video processing. The method comprises the steps that foreground-pixel and light-stream features of pedestrians are extracted from a monitoring video; the extracted foreground-pixel features are normalized; pedestrian shield is detected and shield level is estimated; and pedestrian features are integrated with the shield level to count the pedestrians. The pedestrian counting method counts the pedestrians according to the features of the line-crossing pedestrians, thereby avoiding the process of pedestrian detection; counting is carried out in an ROI area around a virtual door, thereby reducing computational complexity; and the pedestrian shield is detected and the shield level is estimated, thereby improving counting accuracy under the condition that shield exists.
Description
technical field [0001] The invention belongs to the technical field of computer video processing, in particular to a method for counting pedestrians, in particular, a method for counting pedestrians crossing a line suitable for monitoring applications is designed. Background technique [0002] At present, pedestrian counting technology is playing an increasingly important role in monitoring, and pedestrians are the key monitoring objects in monitoring. [0003] Pedestrian counting based on pedestrian detection and pedestrian counting based on pixel statistics are two common pedestrian counting methods. However, pedestrian counting methods based on pedestrian detection usually need to detect all the pedestrians in the scene for counting (see: BoWu, R. Nevatia. "Detection of multiple, partially included humans in single image by Bayesian combination of edgelet part detectors," ICCV2005, Tenth IEEE International Conference on, 1:90–97, 2005.) Therefore, it has a large amount 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): G06K9/62
Inventor 杨华马文琪
Owner SHANGHAI JIAOTONG UNIV