Monitoring the vertical distribution of HABs using hyperspectral imagery and deep learning models

被引:26
作者
Hong, Seok Min [1 ]
Baek, Sang-Soo [1 ]
Yun, Daeun [1 ]
Kwon, Yong-Hwan [2 ]
Duan, Hongtao [3 ]
Pyo, JongCheol [4 ]
Cho, Kyung Hwa [1 ]
机构
[1] Ulsan Natl Inst Sci & Technol, Sch Urban & Environm Engn, Ulsan 689798, South Korea
[2] Elect & Telecommun Res Inst, 218 Gajeong Ro, Daejeon 305700, South Korea
[3] Chinese Acad Sci, Nanjing Inst Geog & Limnol, Key Lab Watershed Geog Sci, Nanjing 210008, Peoples R China
[4] Korea Environm Inst, Ctr Environm Data Strategy, Sejong 30147, South Korea
关键词
Drone-borne hyperspectral image; Explainable deep learning model; Cyanobacteria; Vertical profile; REMOTE-SENSING REFLECTANCE; CHLOROPHYLL-A; CYANOBACTERIAL BLOOMS; DIURNAL CHANGES; INLAND WATERS; ALGAL BLOOMS; LAKE; MICROCYSTIS; PHYCOCYANIN; CLASSIFICATION;
D O I
10.1016/j.scitotenv.2021.148592
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
Remote sensing techniques have been applied to monitor the spatiotemporal variation of harmful algal blooms (HABs) in many inland waters. However, these studies have been limited to monitor the vertical distribution of HABs due to the optical complexity of inland water. Therefore, this study applied a deep neural network model to monitor the vertical distribution of Chlorophyll-a (Chl-a), phycocyanin (PC), and turbidity (Turb) using drone-borne hyperspectral imagery, in-situ measurement, and meteoroidal data. The pigment concentrations were measured between depths of 0 m and 5.0 m with 0.05 m intervals. Here, four state-of-the-art data driven model structures (ResNet-18, ResNet-101, GoogLeNet, and Inception v3) were adopted for estimating the vertical distributions of the harmful algal pigments. Among the four models, the ResNet-18 model showed the best performance, with an R-2 value of 0.70. In addition, Gradient-weighted Class Activation Mapping (Grad-CAM) substantially provided informative reflectance band ranges near 490 nm and 620 nm in the hyperspectral image for the vertical estimation of pigments. Therefore, this study demonstrated that the explainable deep learning model with drone-borne hyperspectral images has the potential to estimate Chl-a, PC, and Turb vertical distributions and to show influential features that contribute to describing the vertical profile phenomena. (C) 2021 Elsevier B.V. All rights reserved.
引用
收藏
页数:13
相关论文
共 79 条
[11]   Factors affecting algal blooms in a man-made lake and prediction using an artificial neural network [J].
Cho, Sohyun ;
Lim, Byungjin ;
Jung, Jaewoon ;
Kim, Sangdon ;
Chae, Hyunmi ;
Park, Jonghwan ;
Park, Seoksoon ;
Park, Jae K. .
MEASUREMENT, 2014, 53 :224-233
[12]   Regression with residual neural network for vanishing point detection [J].
Choi, Han-Soo ;
An, Keunhoi ;
Kang, Myungjoo .
IMAGE AND VISION COMPUTING, 2019, 91
[13]   Performance evaluation of in situ fluorometers for real-time cyanobacterial monitoring [J].
Choo, Florence ;
Zamyadi, Arash ;
Newton, Kelly ;
Newcombe, Gayle ;
Bowling, Lee ;
Stuetz, Richard ;
Henderson, Rita K. .
H2OPEN JOURNAL, 2018, 1 (01) :26-46
[14]   The influence of physical and physiological processes on the spatial heterogeneity of a Microcystis bloom in a stratified reservoir [J].
Chung, S. W. ;
Imberger, J. ;
Hipsey, M. R. ;
Lee, H. S. .
ECOLOGICAL MODELLING, 2014, 289 :133-149
[15]   Cyanobacterial toxins: risk management for health protection [J].
Codd, GA ;
Morrison, LF ;
Metcalf, JS .
TOXICOLOGY AND APPLIED PHARMACOLOGY, 2005, 203 (03) :264-272
[16]   Estimating corn leaf chlorophyll concentration from leaf and canopy reflectance [J].
Daughtry, CST ;
Walthall, CL ;
Kim, MS ;
de Colstoun, EB ;
McMurtrey, JE .
REMOTE SENSING OF ENVIRONMENT, 2000, 74 (02) :229-239
[17]   Response to variable light intensity in photoacclimated algae and cyanobacteria exposed to atrazine [J].
Deblois, Charles P. ;
Dufresne, Karine ;
Juneau, Philippe .
AQUATIC TOXICOLOGY, 2013, 126 :77-84
[18]  
Ficek D., 2011, REMOTE SENSING REFLE
[19]   THE FLUORESCENCE SPECTRA OF RED ALGAE AND THE TRANSFER OF ENERGY FROM PHYCOERYTHRIN TO PHYCOCYANIN AND CHLOROPHYLL [J].
FRENCH, CS ;
YOUNG, VK .
JOURNAL OF GENERAL PHYSIOLOGY, 1952, 35 (06) :873-890
[20]   Selective Inhibition on Growth and Photosynthesis of Harmful Cyanobacteria (Microcystis aeruginosa) by Water Soluble Substances ofDendranthema indicumFlowers [J].
Gao, Yunni ;
Zhang, Fang ;
Wu, Jing ;
Yang, Hui ;
Dong, Jing ;
Zhang, Man ;
Li, Xuejun .
WATER, 2020, 12 (07)