International Display Workshops General Incorporated Association

Proceedings of the International Display Workshops Volume 26 (IDW '19)

Proceedings of the International Display Workshops Volume 26 (IDW '19)

2019年11月27日 〜 2019年11月29日
Sapporo Convention Center

18:10 〜 18:25

[AIS2-4L] Outdoor Wild Bird Detection based on YOLO algorithm

*Bo-Cheng Zhu1, Tzung-Han Lin1, Yao-Chuan Tsai2, Kuang-Wen Hsieh2, Fuh-Min Fan2, Perng-Kwei Lei2 (1. National Taiwan University of Science and Technology (Taiwan), 2. National Chung-Hsing University (Taiwan))

Deep Learning, Object detection, Small Object

https://doi.org/10.36463/idw.2019.0036

This study focuses on outdoor bird detection in video surveillance to reduce the risk of avian influenza (AI) in poultry farms. Once a bird is detected, our system will trigger another action to drive away the bird. We utilized YOLO algorithm for object detection, and add diversity samples to the dataset to train the model, then recognition accuracy is improved.