Wild animal survey using UAS imagery and deep learning: modified Faster R-CNN for kiang detection in Tibetan Plateau

doi:10.1016/j.isprsjprs.2020.08.026

	Wild animal survey using UAS imagery and deep learning: modified Faster R-CNN for kiang detection in Tibetan Plateau
	Peng, Jinbang 1,2,3,4,5,6; Wang, Dongliang 1,3,4,5,6; Liao, Xiaohan 3,4,5; Shao, Quanqin 1; Sun, Zhigang 2,6; Yue, Huanyin 3,4,5; Ye, Huping 3,4,5
刊名	ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
	2020-11-01
卷号	169 页码:364-376
关键词	Wild animal survey Deep learning Object detection Unmanned aircraft systems (UAS)
ISSN号	0924-2716
DOI	10.1016/j.isprsjprs.2020.08.026
通讯作者	Wang, Dongliang(wangdongliang@igsnrr.ac.cn)
英文摘要	Wild animal surveys play a critical role in wild animal conservation and ecosystem management. Unmanned aircraft systems (UASs), with advantages in safety, convenience and inexpensiveness, have been increasingly used in wild animal surveys. However, manually reviewing wild animals from thousands of images generated by UASs is tedious and inefficient. To support wild animal detection in UAS images, researchers have developed various automatic and semiautomatic algorithms. Among these algorithms, deep learning techniques achieve outstanding performances in wild animal detection, but have some practical issues (e.g., limited animal pixels and sparse animal samples). Based on a typical deep learning pipeline, faster region based convolutional neural networks (Faster R-CNN), this study adopted several tactics, including feature stride shortening, anchor size optimization, and hard negative class, to overcome the practical issues in wild animal detection in UAS images. In this study, a kiang survey was conducted in UAS datasets (23,748 images) obtained by 14 flight campaigns in the eastern Tibetan Plateau. The validation experiments of our adopted tactics revealed the following: (1) feature stride shortening and anchor size optimization improved small animal detection performance in the animal patch set, increasing the F1 score from 0.84 to 0.86 and from 0.86 to 0.92, respectively; and (2) the hard negative class significantly suppressed false positives in the full UAS image set, increasing the F1 score from 0.44 to 0.86. The test results in the full UAS image set showed that the modified model with the adopted tactics can be applied to either a semiautomatic survey to accelerate manual verification by 25 times or an automatic survey with an F1 score of approximately 0.90. This study demonstrates that the combination of UAS and deep learning techniques can enable automatic/semiautomatic, accurate, inexpensive, and efficient wild animal surveys.
资助项目	Major Projects of High Resolution Earth Observation System of China (Civil Part) Research on key technologies of CHEOS remote sensing product validation and compilation of standards and specifications ; National Natural Science Foundation of China[41501416] ; National Natural Science Foundation of China[41771388] ; National Natural Science Foundation of China[41971359] ; National Key R&D Program of China[2017YFB0503005] ; National Key R&D Program of China[2017YFC0506500] ; Strategic Priority Research Program of the Chinese Academy of Sciences[XDA23100200] ; State Key Laboratory of Resources and Environmental Information System ; Tianjin Intelligent Manufacturing Project: Technology of Intelligent Networking by Autonomous Control UAVs for Observation and Application[Tianjin-IMP-2] ; Natural Science Foundation of Tianjin[18JCYBJC42300]
WOS关键词	UNMANNED AERIAL VEHICLES ; OBJECT DETECTION ; DENSITY ; COST ; CLASSIFICATION ; HABITAT
WOS研究方向	Physical Geography ; Geology ; Remote Sensing ; Imaging Science & Photographic Technology
语种	英语
出版者	ELSEVIER
WOS记录号	WOS:000584231200028
资助机构	Major Projects of High Resolution Earth Observation System of China (Civil Part) Research on key technologies of CHEOS remote sensing product validation and compilation of standards and specifications ; National Natural Science Foundation of China ; National Key R&D Program of China ; Strategic Priority Research Program of the Chinese Academy of Sciences ; State Key Laboratory of Resources and Environmental Information System ; Tianjin Intelligent Manufacturing Project: Technology of Intelligent Networking by Autonomous Control UAVs for Observation and Application ; Natural Science Foundation of Tianjin
内容类型	期刊论文
源URL	[http://ir.igsnrr.ac.cn/handle/311030/156355]
专题	中国科学院地理科学与资源研究所
通讯作者	Wang, Dongliang
作者单位	1.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, Key Lab Land Surface Pattern & Simulat, Beijing 100101, Peoples R China 2.Chinese Acad Sci, Key Lab Ecosyst Network Observat & Modeling Inst, Beijing 100101, Peoples R China 3.Chinese Acad Sci, State Key Lab Resources & Environm Informat Syst, Inst Geog Sci & Nat Resources Res, Beijing 100101, Peoples R China 4.Inst UAV Applicat Res, Tianjin 301800, Peoples R China 5.Chinese Acad Sci, Tianjin 301800, Peoples R China 6.Univ Chinese Acad Sci, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Peng, Jinbang,Wang, Dongliang,Liao, Xiaohan,et al. Wild animal survey using UAS imagery and deep learning: modified Faster R-CNN for kiang detection in Tibetan Plateau[J]. ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING,2020,169:364-376.
APA	Peng, Jinbang.,Wang, Dongliang.,Liao, Xiaohan.,Shao, Quanqin.,Sun, Zhigang.,...&Ye, Huping.(2020).Wild animal survey using UAS imagery and deep learning: modified Faster R-CNN for kiang detection in Tibetan Plateau.ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING,169,364-376.
MLA	Peng, Jinbang,et al."Wild animal survey using UAS imagery and deep learning: modified Faster R-CNN for kiang detection in Tibetan Plateau".ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING 169(2020):364-376.