TY - JOUR
T1 - Robust vehicle detection and counting algorithm employing a convolution neural network and optical flow
AU - Gomaa, Ahmed
AU - Abdelwahab, Moataz M.
AU - Abo-Zahhad, Mohammed
AU - Minematsu, Tsubasa
AU - Taniguchi, Rin Ichiro
N1 - Funding Information:
Acknowledgments: This work was supported in part by Kyushu University and in part by the Egyptian Ministry of Higher Education (MoHE), Cairo, Egypt, and in part by the Egypt Japan University of Science and Technology (EJUST), Alexandria, Egypt.
Publisher Copyright:
© 2019 by the authors. Licensee MDPI, Basel, Switzerland.
PY - 2019/10/2
Y1 - 2019/10/2
N2 - Automatic vehicle detection and counting are considered vital in improving traffic control and management. This work presents an effective algorithm for vehicle detection and counting in complex traffic scenes by combining both convolution neural network (CNN) and the optical flow feature tracking-based methods. In this algorithm, both the detection and tracking procedures have been linked together to get robust feature points that are updated regularly every fixed number of frames. The proposed algorithm detects moving vehicles based on a background subtraction method using CNN. Then, the vehicle’s robust features are refined and clustered by motion feature points analysis using a combined technique between KLT tracker and K-means clustering. Finally, an efficient strategy is presented using the detected and tracked points information to assign each vehicle label with its corresponding one in the vehicle’s trajectories and truly counted it. The proposed method is evaluated on videos representing challenging environments, and the experimental results showed an average detection and counting precision of 96.3% and 96.8%, respectively, which outperforms other existing approaches.
AB - Automatic vehicle detection and counting are considered vital in improving traffic control and management. This work presents an effective algorithm for vehicle detection and counting in complex traffic scenes by combining both convolution neural network (CNN) and the optical flow feature tracking-based methods. In this algorithm, both the detection and tracking procedures have been linked together to get robust feature points that are updated regularly every fixed number of frames. The proposed algorithm detects moving vehicles based on a background subtraction method using CNN. Then, the vehicle’s robust features are refined and clustered by motion feature points analysis using a combined technique between KLT tracker and K-means clustering. Finally, an efficient strategy is presented using the detected and tracked points information to assign each vehicle label with its corresponding one in the vehicle’s trajectories and truly counted it. The proposed method is evaluated on videos representing challenging environments, and the experimental results showed an average detection and counting precision of 96.3% and 96.8%, respectively, which outperforms other existing approaches.
UR - http://www.scopus.com/inward/record.url?scp=85074163412&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85074163412&partnerID=8YFLogxK
U2 - 10.3390/s19204588
DO - 10.3390/s19204588
M3 - Article
C2 - 31652552
AN - SCOPUS:85074163412
VL - 19
JO - Sensors
JF - Sensors
SN - 1424-3210
IS - 20
M1 - 4588
ER -