A concept of local gray entropy is introduced to solve the problem of the human body detection difficulty under the dim contrast environment. This paper analyses the trait of local gray entropy, and then proposes an algorithm of human target detection based on the trait, which is according to the principle that local gray entropy could reflect the discrete level accurately, and it has no relationship with the average gray. It establishes the background model, and extracts the background region and the foreground region by the background subtraction. The algorithm is able to effectively determine the domain window which is consistent with the conditions of the threshold and calculate the difference between the local gray entropy value of the background and of the moving objects according to the domain window obtained. With the limit of the difference threshold set, we can obtain the moving people region by using the test detection rate and the false-alarm rate as the evaluation index. The experimental results show that the algorithm of people target detection based on local gray entropy can obtain the people region more accurately under the dim contrast environment than others.
[1]
Wang Dongyun,et al.
Improved method of fabric defects inspection based on local entropy
,
2008,
2008 27th Chinese Control Conference.
[2]
Xinping Guan,et al.
Application of a new image segmentation method to detection of defects in castings
,
2009
.
[4]
C. E. SHANNON,et al.
A mathematical theory of communication
,
1948,
MOCO.
[5]
Sang Joon Kim,et al.
A Mathematical Theory of Communication
,
2006
.
[6]
Jitendra Malik,et al.
Robust Multiple Car Tracking with Occlusion Reasoning
,
1994,
ECCV.
[7]
Bill Triggs,et al.
Histograms of oriented gradients for human detection
,
2005,
2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).
[8]
Zhe Li,et al.
Blind Detection of Digital Forgery Image Based on the Local Entropy of the Gradient
,
2008,
IWDW.