Morphology classification and behaviors identification of birds in scientific video

DOI

10.2991/icmt-13.2013.178How to use a DOI?

Keywords

birds behaviour analysis and identify target detection and tracking morphologyclassification feature vector support vector machine

Abstract

Through the establishment of video sensor networks, allowing researchers to do long-term field observations better in their offices. But the study of morphology classification and behavior identification of animals from mass video data is very costly and time-consuming repetitive work. For this, we did video pre-processing and multiple targets recognition and tracking. We built a classification model by support vector machine algorithm. Used 384 morphological records of bar-headed geese from the scientific video obtained in Qinghai Lake, we extracted a vector constructed 9 features in-cluing aspect ratio, shape factor and 7 Hu invariant moments respectively from each record and thus got 384 vectors in total, in which 289 were used for training and 95 for testing. The morphology classification accuracy reached 65.3%. And we used bar-headed geese movement parameters and bar-headed geese posture information; we analyzed and identified bar-headed geese four basic behaviors, and achieve a certain recognition rate. This will help the researches to do study the normal behavior and abnormal behavior of the bar-headed geese and the reserve personnel to do daily monitoring of the bar-headed geese.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Chen Can
AU  - Xiong Yan
AU  - Yan Baoping
PY  - 2013/11
DA  - 2013/11
TI  - Morphology classification and behaviors identification of birds in scientific video
BT  - Proceedings of 3rd International Conference on Multimedia Technology(ICMT-13)
PB  - Atlantis Press
SP  - 1449
EP  - 1457
SN  - 1951-6851
UR  - https://doi.org/10.2991/icmt-13.2013.178
DO  - 10.2991/icmt-13.2013.178
ID  - Can2013/11
ER  -