Dynamic Bird Detection Using Image Processing and Neural Network

Collisions of aircraft and birds cause serious flight accidents, and various studies are underway to find a solution to the problem. In recent image recognition studies, state-of-the-art deep learning technologies have been actively applied. This paper proposes image preprocessing and bird detection...

Full description

Saved in:

Bibliographic Details
Published in:	2019 7th International Conference on Robot Intelligence Technology and Applications (RiTA) pp. 210 - 214
Main Authors:	Jo, Jeongjin, Park, Junwon, Han, Jinyoung, Lee, Minsun, Smith, Anthony H.
Format:	Conference Proceeding
Language:	English
Published:	IEEE 01-11-2019
Subjects:	Atmospheric modeling Birds Colored noise Data models Image color analysis Predictive models Vegetation
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Collisions of aircraft and birds cause serious flight accidents, and various studies are underway to find a solution to the problem. In recent image recognition studies, state-of-the-art deep learning technologies have been actively applied. This paper proposes image preprocessing and bird detection methods in all dynamic environments using Convolutional Neural Network (CNN) technology. Image preprocessing separates moving creatures from the dynamic background and removes the background. When image preprocessing is complete, the image of the moving object remaining in the frame is used as input data for the learning model to determine whether the bird is in the frame. We used the Inception -v3 neural network model to improve the accuracy of small object classifications.
DOI:	10.1109/RITAPP.2019.8932891