TY - JOUR
T1 - Counting pedestrians with a zenithal arrangement of depth cameras
AU - Vera, Pablo
AU - Monjaraz, Sergio
AU - Salas, Joaquín
N1 - Publisher Copyright:
© 2015, Springer-Verlag Berlin Heidelberg.
PY - 2016/2/1
Y1 - 2016/2/1
N2 - Counting people is a basic operation in applications that include surveillance, marketing, services, and others. Recently, computer vision techniques have emerged as a non-intrusive, cost-effective, and reliable solution to the problem of counting pedestrians. In this article, we introduce a system capable of counting people using a cooperating network of depth cameras placed in zenithal position. In our method, we first detect people in each camera of the array separately. Then, we construct and consolidate tracklets based on their closeness and time stamp. Our experimental results show that the method permits to extend the narrow range of a single sensor to wider scenarios.
AB - Counting people is a basic operation in applications that include surveillance, marketing, services, and others. Recently, computer vision techniques have emerged as a non-intrusive, cost-effective, and reliable solution to the problem of counting pedestrians. In this article, we introduce a system capable of counting people using a cooperating network of depth cameras placed in zenithal position. In our method, we first detect people in each camera of the array separately. Then, we construct and consolidate tracklets based on their closeness and time stamp. Our experimental results show that the method permits to extend the narrow range of a single sensor to wider scenarios.
KW - Cameras in zenithal position
KW - Counting pedestrians
KW - Depth cameras
KW - Network of cameras
UR - http://www.scopus.com/inward/record.url?scp=84957951619&partnerID=8YFLogxK
U2 - 10.1007/s00138-015-0739-1
DO - 10.1007/s00138-015-0739-1
M3 - Artículo
AN - SCOPUS:84957951619
SN - 0932-8092
VL - 27
SP - 303
EP - 315
JO - Machine Vision and Applications
JF - Machine Vision and Applications
IS - 2
ER -