TY - GEN
T1 - Vehicle Detection in UAV Videos Using CNN-SVM
AU - Valappil, Najiya Koderi
AU - Memon, Qurban A.
N1 - Publisher Copyright:
© 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - Conventional monitoring devices are usually kept at fixed locations which yields a fixed surveillance coverage. Unmanned aerial vehicles (UAVs) are receiving much attention from researchers in traffic monitoring due to their low cost, high flexibility, and wide view. Unlike stationary surveillance, the camera platform of UAVs is in constant motion and makes it difficult to process for data extraction. The inaccuracy in detection rates of vehicles from UAV videos becomes the motivation for combining optical flow methods with supervised learning algorithms. The proposed method incorporates steps that make use of the Kanade–Lucas optical flow method for moving object detection, connected graphs theory and CNN-SVM for further classification. Optical flow generated contains some background objects detected as vehicle when the camera platforms are moving. The classifier rules out the presence of any other moving objects to be detected as vehicles. The proposed method is tested on few stationary and moving aerial videos. The system is found to be 100% accurate in case of stationary aerial videos and 98% accurate in moving videos.
AB - Conventional monitoring devices are usually kept at fixed locations which yields a fixed surveillance coverage. Unmanned aerial vehicles (UAVs) are receiving much attention from researchers in traffic monitoring due to their low cost, high flexibility, and wide view. Unlike stationary surveillance, the camera platform of UAVs is in constant motion and makes it difficult to process for data extraction. The inaccuracy in detection rates of vehicles from UAV videos becomes the motivation for combining optical flow methods with supervised learning algorithms. The proposed method incorporates steps that make use of the Kanade–Lucas optical flow method for moving object detection, connected graphs theory and CNN-SVM for further classification. Optical flow generated contains some background objects detected as vehicle when the camera platforms are moving. The classifier rules out the presence of any other moving objects to be detected as vehicles. The proposed method is tested on few stationary and moving aerial videos. The system is found to be 100% accurate in case of stationary aerial videos and 98% accurate in moving videos.
KW - Aerial video
KW - CNN-SVM
KW - Kanade-Lucas optical flow
KW - Traffic surveillance
KW - Traffic video analysis
KW - Unmanned aerial vehicle
UR - http://www.scopus.com/inward/record.url?scp=85105883935&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85105883935&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-73689-7_22
DO - 10.1007/978-3-030-73689-7_22
M3 - Conference contribution
AN - SCOPUS:85105883935
SN - 9783030736880
T3 - Advances in Intelligent Systems and Computing
SP - 221
EP - 232
BT - Proceedings of the 12th International Conference on Soft Computing and Pattern Recognition, SoCPaR 2020
A2 - Abraham, Ajith
A2 - Ohsawa, Yukio
A2 - Gandhi, Niketa
A2 - Jabbar, M. A.
A2 - Haqiq, Abdelkrim
A2 - McLoone, Seán
A2 - Issac, Biju
PB - Springer Science and Business Media Deutschland GmbH
T2 - 12th International Conference on Soft Computing and Pattern Recognition, SoCPaR 2020 and 16th International Conference on Information Assurance and Security, IAS 2020
Y2 - 15 December 2020 through 18 December 2020
ER -