Citation

Teng, Ching Yee and Connie, Tee and Choo, Kan Yeep and Goh, Michael Kah Ong (2022) A Visual Approach Towards Wildlife Surveillance in Malaysia. In: 2022 10th International Conference on Information and Communication Technology (ICoICT), 2-3 August 2022, Bandung, Indonesia.

Text A_Visual_Approach_Towards.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Road development has caused increased wildlife-vehicle collision (WCV) in Malaysia. WCV does not only pose a high risk to road safety but also become a major threat to the survival of precious wildlife in the country. Every year, hundreds of endangered species such as the Malayan Tapir are killed on the road. The wild animals set foot on man areas because their movement corridor has been blocked by roads or highways. Besides, the wild animals are forced to come out from the forests to hunt for food due to excessive deforestation that destroy their natural habitats. In this study, wildlife surveillance system based in visual input analysis is presented. Footages from CCTV cameras placed near the roadways are used to perform surveillance monitoring. A data-driven object detection method based on deep transfer learning is used to automatically monitor and detect the occurrence of wildlife in the scene under observation. Once the sign of wildlife is detected, a warning will be sent to the corresponding conservation or road infrastructure departments to take immediate action. Empirical results show that the proposed method achieve a mean average precision (mAP50) of 95.6% and an average recall of 94% in detecting wildlife. This suggests that proposed method offers a promising solution to help to reduce the cases of WCV in the country through visual surveillance.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Deep learning, animal recognition, YOLO, transfer learning, computer vision
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	08 Dec 2022 08:19
Last Modified:	08 Dec 2022 08:19
URII:	http://shdl.mmu.edu.my/id/eprint/10751

Downloads

Edit (login required)