Multitask Learning of Alfalfa Nutritive Value From UAV-Based Hyperspectral Images

被引:20
作者
Feng, Luwei [1 ]
Zhang, Zhou [2 ]
Ma, Yuchi [2 ]
Sun, Yazhou [2 ]
Du, Qingyun [1 ]
Williams, Parker [2 ]
Drewry, Jessica [2 ]
Luck, Brian [2 ]
机构
[1] Wuhan Univ, Sch Resources & Environm Sci, Wuhan 430079, Peoples R China
[2] Univ Wisconsin, Dept Biol Syst Engn, Madison, WI 53706 USA
基金
美国食品与农业研究所;
关键词
Task analysis; Hyperspectral imaging; Mathematical model; Agriculture; Predictive models; Computer architecture; Logic gates; Alfalfa; hyperspectral imagery; multitask learning; nutritive value; unmanned aerial vehicle (UAV); FORAGE QUALITY; PREDICTION; YIELD;
D O I
10.1109/LGRS.2021.3079317
中图分类号
P3 [地球物理学]; P59 [地球化学];
学科分类号
0708 ; 070902 ;
摘要
Alfalfa is a valuable and widely adapted forage crop, and its nutritive value directly affects animal performance and ultimately affects the profitability of livestock production. Traditional nutritive value measurement method is labor-intensive and time-consuming and thus hinders the determination of alfalfa nutritive values over large fields. The adoption of unmanned aerial vehicles (UAVs) facilitates the generation of images with high spatial and temporal resolutions for field-level agricultural research. Additionally, compared with other imaging modalities, hyperspectral data usually consist of hundreds of narrow spectral bands and allow the accurate detection, identification, and quantification of crop quality. Although various machine-learning methods have been developed for alfalfa quality prediction, they were all single-task models that learned independently for each quality trait and failed to utilize the underlying relatedness between each task. Inspired by the idea of multitask learning (MTL), this study aims to develop an approach that simultaneously predicts multiple quality traits. The algorithm first extracts shared information through a long short-term memory (LSTM)-based common hidden layer. To enhance the model flexibility, it is then divided into multiple branches, each containing the same or different number of task-specific fully connected hidden layers. Through comparison with multiple mainstream single-task machine-learning models, the effectiveness of the model is illustrated based on the measured alfalfa quality data and multitemporal UAV-based hyperspectral imagery.
引用
收藏
页数:5
相关论文
共 27 条
[1]   Improving Landsat predictions of rangeland fractional cover with multitask learning and uncertainty [J].
Allred, Brady W. ;
Bestelmeyer, Brandon T. ;
Boyd, Chad S. ;
Brown, Christopher ;
Davies, Kirk W. ;
Duniway, Michael C. ;
Ellsworth, Lisa M. ;
Erickson, Tyler A. ;
Fuhlendorf, Samuel D. ;
Griffiths, Timothy V. ;
Jansen, Vincent ;
Jones, Matthew O. ;
Karl, Jason ;
Knight, Anna ;
Maestas, Jeremy D. ;
Maynard, Jonathan J. ;
McCord, Sarah E. ;
Naugle, David E. ;
Starns, Heath D. ;
Twidwell, Dirac ;
Uden, Daniel R. .
METHODS IN ECOLOGY AND EVOLUTION, 2021, 12 (05) :841-849
[2]  
[Anonymous], US Department of Agriculture National Agricultural Statistics Service, 1985-2015
[3]   Crop Status Index as an indicator of wheat crop growth condition under abiotic stress situations [J].
Banerjee, Victor ;
Krishnan, P. ;
Das, Bappa ;
Verma, A. P. S. ;
Varghese, E. .
FIELD CROPS RESEARCH, 2015, 181 :16-31
[4]   Multitask learning [J].
Caruana, R .
MACHINE LEARNING, 1997, 28 (01) :41-75
[5]   Alfalfa Yield Prediction Using UAV-Based Hyperspectral Imagery and Ensemble Learning [J].
Feng, Luwei ;
Zhang, Zhou ;
Ma, Yuchi ;
Du, Qingyun ;
Williams, Parker ;
Drewry, Jessica ;
Luck, Brian .
REMOTE SENSING, 2020, 12 (12)
[6]   Boresight Calibration of GNSS/INS-Assisted Push-Broom Hyperspectral Scanners on UAV Platforms [J].
Habib, Ayman ;
Zhou, Tian ;
Masjedi, Ali ;
Zhang, Zhou ;
Flatt, John Evan ;
Crawford, Melba .
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2018, 11 (05) :1734-1749
[7]   Cascaded Recurrent Neural Networks for Hyperspectral Image Classification [J].
Hang, Renlong ;
Liu, Qingshan ;
Hong, Danfeng ;
Ghamisi, Pedram .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2019, 57 (08) :5384-5394
[8]   Forage quality variation in the US alfalfa core collection [J].
Jung, HG ;
Sheaffer, CC ;
Barnes, DK ;
Halgerson, JL .
CROP SCIENCE, 1997, 37 (04) :1361-1366
[9]   Dry season mapping of savanna forage quality, using the hyperspectral Carnegie Airborne Observatory sensor [J].
Knox, Nichola M. ;
Skidmore, Andrew K. ;
Prins, Herbert H. T. ;
Asner, Gregory P. ;
van der Werff, Harald M. A. ;
de Boer, Willem F. ;
van der Waal, Cornelis ;
de Knegt, Hendrik J. ;
Kohi, Edward M. ;
Slotow, Rob ;
Grant, Rina C. .
REMOTE SENSING OF ENVIRONMENT, 2011, 115 (06) :1478-1488
[10]  
Laser H., 2018, EXPT SIMULATION INTE, P93