Multi-season unmixing of vegetation class fractions across diverse Californian ecoregions using simulated spaceborne imaging spectroscopy data

2021 ◽  
pp. 112558
Author(s):  
Akpona Okujeni ◽  
Clemens Jänicke ◽  
Sam Cooper ◽  
David Frantz ◽  
Patrick Hostert ◽  
...  
Keyword(s):  
Imaging Spectroscopy ◽  
Spectroscopy Data ◽  
Vegetation Class

scholarly journals Quantifying Vegetation Biophysical Variables from Imaging Spectroscopy Data: A Review on Retrieval Methods

2018 ◽  
Vol 40 (3) ◽  
pp. 589-629 ◽  
Author(s):  
Jochem Verrelst ◽  
Zbyněk Malenovský ◽  
Christiaan Van der Tol ◽  
Gustau Camps-Valls ◽  
Jean-Philippe Gastellu-Etchegorry ◽  
...  
Keyword(s):  
Imaging Spectroscopy ◽  
Spectroscopy Data

scholarly journals Assessing the Spectral Properties of Sunlit and Shaded Components in Rice Canopies with Near-Ground Imaging Spectroscopy Data

Sensors ◽  
2017 ◽  
Vol 17 (3) ◽  
pp. 578 ◽  
Author(s):  
Kai Zhou ◽  
Xinqiang Deng ◽  
Xia Yao ◽  
Yongchao Tian ◽  
Weixing Cao ◽  
...  
Keyword(s):  
Spectral Properties ◽  
Imaging Spectroscopy ◽  
Spectroscopy Data

scholarly journals Material-specific user colour profiles from imaging spectroscopy data

2011 ◽  
Author(s):  
Lin Gu ◽  
Cong Phuoc Huynh ◽  
Antonio Robles-Kelly ◽  
Jun Zhou
Keyword(s):  
Imaging Spectroscopy ◽  
Spectroscopy Data

scholarly journals EVALUATION OF VARIOUS SPECTRAL INPUTS FOR ESTIMATION OF FOREST BIOCHEMICAL AND STRUCTURAL PROPERTIES FROM AIRBORNE IMAGING SPECTROSCOPY DATA

2016 ◽  
Vol XLI-B7 ◽  
pp. 961-966
Author(s):  
L. Homolová ◽  
R. Janoutová ◽  
Z. Malenovský
Keyword(s):  
Imaging Spectroscopy ◽  
Beech Forest ◽  
Retrieval Algorithm ◽  
Support Vector ◽  
Spectroscopy Data ◽  
Area Index ◽  
The Czech Republic ◽  
Sensor Field ◽  
Airborne Imaging ◽  
Study Sites

In this study we evaluated various spectral inputs for retrieval of forest chlorophyll content (Cab) and leaf area index (LAI) from high spectral and spatial resolution airborne imaging spectroscopy data collected for two forest study sites in the Czech Republic (beech forest at Štítná nad Vláří and spruce forest at Bílý Kříž). The retrieval algorithm was based on a machine learning method – support vector regression (SVR). Performance of the four spectral inputs used to train SVR was evaluated: a) all available hyperspectral bands, b) continuum removal (CR) 645 – 710 nm, c) CR 705 – 780 nm, and d) CR 680 – 800 nm. Spectral inputs and corresponding SVR models were first assessed at the level of spectral databases simulated by combined leaf-canopy radiative transfer models PROSPECT and DART. At this stage, SVR models using all spectral inputs provided good performance (RMSE for Cab < 10 μg cm<sup>&minus;2</sup> and for LAI < 1.5), with consistently better performance for beech over spruce site. Since application of trained SVRs on airborne hyperspectral images of the spruce site produced unacceptably overestimated values, only the beech site results were analysed. The best performance for the Cab estimation was found for CR bands in range of 645 – 710 nm, whereas CR bands in range of 680 – 800 nm were the most suitable for LAI retrieval. The CR transformation reduced the across-track bidirectional reflectance effect present in airborne images due to large sensor field of view.

scholarly journals The CCB-ID approach to tree species mapping with airborne imaging spectroscopy

2018 ◽  
Author(s):  
Christopher B Anderson
Keyword(s):  
Tree Species ◽  
Rare Species ◽  
Imaging Spectroscopy ◽  
Class Imbalance ◽  
Accuracy Score ◽  
Spectroscopy Data ◽  
Ecological Data ◽  
Outlier Removal ◽  
Airborne Imaging ◽  
Species Mapping

Background. Biogeographers assess how species distributions and abundances affect the structure, function, and composition of ecosystems. Yet we face a major challenge: it is difficult to precisely map species across landscapes. Novel Earth observations could obviate this challenge. Airborne imaging spectrometers measure plant functional traits at high resolution, and these measurements can be used to identify tree species. Plant traits are often highly conserved within species, and highly variable between species, which provides the biophysical basis for species mapping. In this paper I describe a trait-based approach to species identification with imaging spectroscopy, CCB-ID, which was developed as part of a NIST-sponsored ecological data science evaluation (ECODSE). Methods. These methods were developed using NEON airborne imaging spectroscopy data. CCB-ID classifies tree species using trait-based reflectance variation and decision tree-based machine learning models, approximating a morphological trait and dichotomous key method traditionally used in botanical classification. First, outliers were removed using a spectral variance threshold. The remaining samples were transformed using principal components analysis and resampled by species to reduce common species biases. Gradient boosting and random forest classifiers were trained using the transformed and resampled feature data. Prediction probabilities were then calibrated using sigmoid regression, and sample-scale predictions were averaged to the crown scale. Results. This approach performed well according to the competition metrics, receiving a rank-1 accuracy score of 0.919, and a cross-entropy cost score of 0.447 on the test data. Accuracy and specificity scores were high for all species, but precision and recall scores were variable for rare species. PCA transformation improved accuracy scores compared to models trained using reflectance data, but outlier removal and data resampling exacerbated class imbalance problems. Discussion. CCB-ID accurately classified tree species using NEON imaging spectroscopy data, reporting the best classification scores among participants. However, it failed to overcome several well-known species mapping challenges, like precisely identifying rare species. Key takeaways include (1) training models to maximize metrics beyond accuracy (e.g. recall) could improve rare species predictions, (2) within-genus trait variation may drive spectral separability, precluding efforts to distinguish between functionally convergent species, (3) outlier removal and data resampling exacerbated class imbalance problems, and should be carefully implemented, (4) PCA transformation greatly improved model results, and (5) feature selection could further improve species classification models. CCB-ID is open source, designed for use with NEON data, and available to support future species mapping efforts.

Sign in / Sign up

Export Citation Format

Share Document