Guanghong Liu
Institute of Physical Education, Shenyang Sport University, Shenyang 110102, China
ABSTRACT
The increase of crowd come in and go out of fitness places can reflects the increase of the number of fitness people from the side. Fitness places often have installed camera, their daily video recording can be used as raw data of health situation in the area. In order to better statistical the number of people in fitness places with high density population, research means on the video should develop toward a more accurate goal that is easier to achieve. In the places with higher population density, target occlusion problem among each other is more prominent, which makes it difficult to detect and trace independent entity in a crowded area and the difficulty to precisely acquire the bodys movement trajectory is strengthened. On the basis of studying the characteristics of the video study object (crowd flow), this study establishes a linear regression model to estimate the population flows. The study first introduces the principle of video motion segmentation and the extraction method of eight categories of image features and then discusses the principles of regression estimation and significance test approach, finally verifies the reasonableness of theoretical models in the text by the data, which provides a theoretical basis for video analysis and provides a better technical foundation for the regional public fitness study.
PDF References Citation
How to cite this article
Guanghong Liu, 2013. Significance Test Algorithm of Crowd Flow in Public Fitness Areas Based On
Regression Analysis. Information Technology Journal, 12: 3374-3377.
DOI: 10.3923/itj.2013.3374.3377
URL: https://scialert.net/abstract/?doi=itj.2013.3374.3377
DOI: 10.3923/itj.2013.3374.3377
URL: https://scialert.net/abstract/?doi=itj.2013.3374.3377
REFERENCES
- Liu, X.R., J.L. Zhou and X.H. Li, 2008. Grey level dependent matrix based crowd density estimation in fourier spectrum. Contr. Automation, 24: 310-314.
Direct Link - Hou, Z.Q., 2005. A background reconstruction algorithm based on pixel intensity classification. J. Software, 16: 1568-1576.
Direct Link