文章基本信息

标题：CLASSIFICATION OF WILD BIRD BY BEHAVIOR WITH FASTER R-CNN FOR COMPLICATED ENVIRONMENT LIKE ORCHARD
本地全文：下载
作者：CHEOL WON LEE ; AZAMJON MUMINOV ; DAEYOUNG NA 等
期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2019
卷号：97
期号：18
页码：4897-4908
出版社：Journal of Theoretical and Applied
摘要：Wild birds cause great significant damages to agricultural crops in orchards every year. The previous non-reactive controlwild bird detection algorithm methods have limitations that cannot be accurately detected because of the biological characteristics of wild birds. are limited by the nature of wild birds adapting rapidly to the environment through learning. Therefore, an aggressive real-time wild bird detection method is needed to respond to wild birds. In this paper, we propose a vision-based real-time wild bird detection method algorithm that operates in a complicated environment like orchard using the Wild Bird Behavior Classification(WBBC) modelFaster R-CNN of deep learning. That is tThe Wild Bird Behavior Classification (WBBC) model algorithm, classifies wild birds according to their behavior, which intended to improve the detection ability accuracy in a complicated environment. We also verified the performance benefit of Behavior Classification model and the performance of the WBBC algorithm through experiments. In our experiments, the Behavior Classification shows 3.6 percent growth than unused. of WBBC by comparing it with previous bird detection methods. The WBBC method algorithm has detected wild birds with an average of by 95.7 percent of% average accuracy in a variety of environments.
关键词：Computer Vision; Bird; Detection; Faster RCNN; Deep Learning