scholarly journals Using hypertemporal Sentinel-1 data to predict forest growing stock volume

2021 ◽  
Author(s):  
Shaojia Ge ◽  
Erkki Tomppo ◽  
Yrjö Rauste ◽  
Ronald E. McRoberts ◽  
Jaan Praks ◽  
...  
Keyword(s):  
Time Series ◽  
Prediction Accuracy ◽  
Temporal Dynamics ◽  
Support Vector ◽  
List Type ◽  
Prediction Errors ◽  
Growing Stock ◽  
Reduction Techniques ◽  
Combined Images ◽  
Stock Volume

AbstractIn this study, we assess the potential of long time series of Sentinel-1 SAR data to predict forest growing stock volume and evaluate the temporal dynamics of the predictions. The boreal coniferous forests study site is located near the Hyytiälä forest station in central Finland and covers an area of 2,500 km2 with nearly 17,000 stands. We considered several prediction approaches (linear, support vector and random forests regression) and fine-tuned them to predict growing stock volume in several evaluation scenarios. The analyses used 96 Sentinel-1 images acquired over three years. Different approaches for aggregating SAR images and choosing feature (predictor) variables were evaluated. Our results demonstrate considerable decrease in RMSEs of growing stock volume as the number of images increases. While prediction accuracy using individual Sentinel-1 images varied from 85 to 91 m3/ha RMSE (relative RMSE 50-53%), RMSE with combined images decreased to 75.6 m3/ha (relative RMSE 44%). Feature extraction and dimension reduction techniques facilitated achieving the near-optimal prediction accuracy using only 8-10 images. When using assemblages of eight consecutive images, the GSV was predicted with the greatest accuracy when initial acquisitions started between September and January.HighlightsTime series of 96 Sentinel-1 images is analysed over study area with 17,762 forest stands.Rigorous evaluation of tools for SAR feature selection and GSV prediction.Improved periodic seasonality using assemblages of consecutive Sentinel-1 images.Analysis of combining images acquired in “frozen” and “dry summer” conditions.Competitive estimates using calculation of prediction errors with stand-area weighting.

scholarly journals Predicting Spatio-temporal Time Series Using Dimension Reduced Local States

2019 ◽  
Vol 30 (3) ◽  
pp. 713-735 ◽  
Author(s):  
Jonas Isensee ◽  
George Datseris ◽  
Ulrich Parlitz
Keyword(s):  
Time Series ◽  
Principle Component Analysis ◽  
Temporal Dynamics ◽  
Time Series Prediction ◽  
Noisy Data ◽  
Nearest Neighbour ◽  
Reduction Techniques ◽  
Spatio Temporal ◽  
Delay Coordinates ◽  
Dimension Reduction Techniques

Abstract We present a method for both cross-estimation and iterated time series prediction of spatio-temporal dynamics based on local modelling and dimension reduction techniques. Assuming homogeneity of the underlying dynamics, we construct delay coordinates of local states and then further reduce their dimensionality through Principle Component Analysis. The prediction uses nearest neighbour methods in the space of dimension reduced states to either cross-estimate or iteratively predict the future of a given frame. The effectiveness of this approach is shown for (noisy) data from a (cubic) Barkley model, the Bueno-Orovio–Cherry–Fenton model, and the Kuramoto–Sivashinsky model.

Study on Prediction of Chaotic Time Series Using Least Squares Support Vector Machines

2014 ◽  
Vol 1061-1062 ◽  
pp. 935-938
Author(s):  
Xin You Wang ◽  
Guo Fei Gao ◽  
Zhan Qu ◽  
Hai Feng Pu
Keyword(s):  
Time Series ◽  
Support Vector Machine ◽  
Support Vector Machines ◽  
Least Squares ◽  
Prediction Accuracy ◽  
Chaotic Time Series ◽  
Support Vector ◽  
Vector Machines ◽  
Online Prediction ◽  
Better Than

The predictions of chaotic time series by applying the least squares support vector machine (LS-SVM), with comparison with the traditional-SVM and-SVM, were specified. The results show that, compared with the traditional SVM, the prediction accuracy of LS-SVM is better than the traditional SVM and more suitable for time series online prediction.

scholarly journals Expansion of Eastern Redcedar (Juniperus virginiana L.) into the Deciduous Woodlands within the Forest–Prairie Ecotone of Kansas

Forests ◽  
2020 ◽  
Vol 11 (2) ◽  
pp. 154 ◽  
Author(s):  
G. A. Pabodha Galgamuwa ◽  
Jida Wang ◽  
Charles J. Barden
Keyword(s):  
Temporal Dynamics ◽  
Control Measures ◽  
Rapid Expansion ◽  
Deciduous Forests ◽  
Support Vector ◽  
Juniperus Virginiana ◽  
Eastern Redcedar ◽  
Growing Stock ◽  
Quercus Species ◽  
Spatio Temporal

North America’s midcontinent forest–prairie ecotone is currently exhibiting extensive eastern redcedar (ERC) (Juniperus virginiana L.) encroachment. Rapid expansion of ERC has major impacts on the species composition and forest structure within this region and suppresses previously dominant oak (Quercus) species. In Kansas, the growing-stock volume of ERC increased by 15,000% during 1965–2010. The overarching goal of this study was to evaluate the spatio-temporal dynamics of ERC in the forest–prairie ecotone of Kansas and understand its effects on deciduous forests. This was achieved through two specific objectives: (i) characterize an effective image classification approach to map ERC expansion, and (ii) assess ERC expansion between 1986 and 2017 in three study areas within the forest–prairie ecotone of Kansas, and especially expansion into deciduous forests. The analysis was based on satellite imagery acquired by Landsat TM and OLI sensors during 1986–2017. The use of multi-seasonal layer-stacks with a Support Vector Machine (SVM)-supervised classification was found to be the most effective approach to classify ERC distribution with high accuracy. The overall accuracies for the change maps generated for the three study areas ranged between 0.95 (95 CI: ±0.02) and 0.96 (±0.03). The total ERC cover increased in excess of 6000 acres in each study area during the 30-year period. The estimated percent increase of ERC cover was 139%, 539%, and 283% for the Tuttle Creek reservoir, Perry reservoir, and Bourbon County north study areas, respectively. This astounding rate of expansion had significant impacts on the deciduous forests where the conversion of deciduous woodlands to ERC, as a percentage of the total encroachment, were 48%, 56%, and 71%, for the Tuttle Creek reservoir, Perry reservoir, and Bourbon County north study areas, respectively. These results strongly affirm that control measures should be implemented immediately to restore the threatened deciduous woodlands of the region.

scholarly journals Evaluation of Using Sentinel-1 and -2 Time-Series to Identify Winter Land Use in Agricultural Landscapes

2018 ◽  
Vol 11 (1) ◽  
pp. 37 ◽  
Author(s):  
Julien Denize ◽  
Laurence Hubert-Moy ◽  
Julie Betbeder ◽  
Samuel Corgne ◽  
Jacques Baudry ◽  
...  
Keyword(s):  
Land Use ◽  
Time Series ◽  
Vegetation Cover ◽  
Temporal Dynamics ◽  
Agricultural Practices ◽  
Agricultural Landscapes ◽  
Support Vector ◽  
Kappa Index ◽  
Spatio Temporal ◽  
Agricultural Areas

Monitoring vegetation cover during winter is a major environmental and scientific issue in agricultural areas. From an environmental viewpoint, the presence and type of vegetation cover in winter influences the transport of pollutants to water resources. From a methodological viewpoint, characterizing spatio-temporal dynamics of land cover and land use at the field scale is challenging due to the diversity of farming strategies and practices in winter. The objective of this study was to evaluate the respective advantages of Sentinel optical and SAR time-series to identify land use in winter. To this end, Sentinel-1 and -2 time-series were classified using Support Vector Machine and Random Forest algorithms in a 130 km² agricultural area. From the classification, the Sentinel-2 time-series identified winter land use more accurately (overall accuracy (OA) = 75%, Kappa index = 0.70) than that of Sentinel-1 (OA = 70%, Kappa = 0.66) but a combination of the Sentinel-1 and -2 time-series was the most accurate (OA = 81%, Kappa = 0.77). Our study outlines the effectiveness of Sentinel-1 and -2 for identify land use in winter, which can help to change agricultural practices.

scholarly journals Decentralized Distribution-sampled Classification Models with Application to Brain Imaging

2019 ◽  
Author(s):  
Noah Lewis ◽  
Harshvardhan Gazula ◽  
Sergey M. Plis ◽  
Vince D. Calhoun
Keyword(s):  
Medical Imaging ◽  
Brain Imaging ◽  
Network Traffic ◽  
Prediction Accuracy ◽  
Data Transfer ◽  
Large Data ◽  
Support Vector ◽  
List Type ◽  
Imaging Data ◽  
Network Load

Abstract0.1backgroundIn this age of big data, large data stores allow researchers to compose robust models that are accurate and informative. In many cases, the data are stored in separate locations requiring data transfer between local sites, which can cause various practical hurdles, such as privacy concerns or heavy network load. This is especially true for medical imaging data, which can be constrained due to the health insurance portability and accountability act (HIPAA). Medical imaging datasets can also contain many thousands or millions of features, requiring heavy network load.0.2New MethodOur research expands upon current decentralized classification research by implementing a new singleshot method for both neural networks and support vector machines. Our approach is to estimate the statistical distribution of the data at each local site and pass this information to the other local sites where each site resamples from the individual distributions and trains a model on both locally available data and the resampled data.0.3ResultsWe show applications of our approach to handwritten digit classification as well as to multi-subject classification of brain imaging data collected from patients with schizophrenia and healthy controls. Overall, the results showed comparable classification accuracy to the centralized model with lower network load than multishot methods.0.4Comparison with Existing MethodsMany decentralized classifiers are multishot, requiring heavy network traffic. Our model attempts to alleviate this load while preserving prediction accuracy.0.5ConclusionsWe show that our proposed approach performs comparably to a centralized approach while minimizing network traffic compared to multishot methods.0.6HighlightsA novel yet simple approach to decentralized classificationReduces total network load compared to current multishot algorithmsMaintains a prediction accuracy comparable to the centralized approach

scholarly journals Evaluation of Different Algorithms for Estimating the Growing Stock Volume of Pinus massoniana Plantations Using Spectral and Spatial Information from a SPOT6 Image

Forests ◽  
2020 ◽  
Vol 11 (5) ◽  
pp. 540 ◽  
Author(s):  
Jingjing Zhou ◽  
Zhixiang Zhou ◽  
Qingxia Zhao ◽  
Zemin Han ◽  
Pengcheng Wang ◽  
...  
Keyword(s):  
Spatial Information ◽  
Vegetation Indices ◽  
Remote Sensing Data ◽  
Window Size ◽  
Pinus Massoniana ◽  
Support Vector ◽  
Moving Window ◽  
Growing Stock ◽  
Textural Parameters ◽  
Stock Volume

Precise growing stock volume (GSV) estimation is essential for monitoring forest carbon dynamics, determining forest productivity, assessing ecosystem forest services, and evaluating forest quality. We evaluated four machine learning methods: classification and regression trees (CART), support vector machines (SVM), artificial neural networks (ANN), and random forests (RF), for their reliability in the estimation of the GSV of Pinus massoniana plantations in China’s northern subtropical regions, using remote sensing data. For all four methods, models were generated using data derived from a SPOT6 image, namely the spectral vegetation indices (SVIs), texture parameters, or both. In addition, the effects of varying the size of the moving window on estimation precision were investigated. RF almost always yielded the greatest precision independently of the choice of input. ANN had the best performance when SVIs were used alone to estimate GSV. When using texture indices alone with window sizes of 3 × 5 × 5 or 9 × 9, RF achieved the best results. For CART, SVM, and RF, R2 decreased as the moving window size increased: the highest R2 values were achieved with 3 × 3 or 5 × 5 windows. When using textural parameters together with SVIs as the model input, RF achieved the highest precision, followed by SVM and CART. Models using both SVI and textural parameters as inputs had better estimating precision than those using spectral data alone but did not appreciably outperform those using textural parameters alone.

scholarly journals Fast Prediction with Sparse Multikernel LS-SVR Using Multiple Relevant Time Series and Its Application in Avionics System

2015 ◽  
Vol 2015 ◽  
pp. 1-10
Author(s):  
Yang M. Guo ◽  
Pei He ◽  
Xiang T. Wang ◽  
Ya F. Zheng ◽  
Chong Liu ◽  
...  
Keyword(s):  
Time Series ◽  
Prediction Model ◽  
Prediction Accuracy ◽  
Mean Squared Error ◽  
Computation Time ◽  
Support Vector ◽  
Good Prediction ◽  
Trend Prediction ◽  
Good Prediction Accuracy ◽  
Fast Prediction

Health trend prediction is critical to ensure the safe operation of highly reliable systems. However, complex systems often present complex dynamic behaviors and uncertainty, which makes it difficult to develop a precise physical prediction model. Therefore, time series is often used for prediction in this case. In this paper, in order to obtain better prediction accuracy in shorter computation time, we propose a new scheme which utilizes multiple relevant time series to enhance the completeness of the information and adopts a prediction model based on least squares support vector regression (LS-SVR) to perform prediction. In the scheme, we apply two innovative ways to overcome the drawbacks of the reported approaches. One is to remove certain support vectors by measuring the linear correlation to increase sparseness of LS-SVR; the other one is to determine the linear combination weights of multiple kernels by calculating the root mean squared error of each basis kernel. The results of prediction experiments indicate preliminarily that the proposed method is an effective approach for its good prediction accuracy and low computation time, and it is a valuable method in applications.

A NOVEL HYBRID FRACTAL INTERPOLATION-SVM MODEL FOR FORECASTING STOCK PRICE INDEXES

Fractals ◽  
2019 ◽  
Vol 27 (04) ◽  
pp. 1950055 ◽  
Author(s):  
HONG-YONG WANG ◽  
HONG LI ◽  
JIN-YE SHEN
Keyword(s):  
Time Series ◽  
Hybrid Model ◽  
Stock Price ◽  
Prediction Accuracy ◽  
Stock Index ◽  
Support Vector ◽  
Fractal Interpolation ◽  
Interpolation Model ◽  
Svm Model ◽  
Price Indexes

Forecasting stock price indexes has been regarded as a challenging task in financial time series analysis. In order to improve the prediction accuracy, a novel hybrid model that integrates fractal interpolation with support vector machine (SVM) models has been developed in this paper to forecast the time series of stock price indexes. For this, a new method to calculate the vertical scaling factors of the fractal interpolation iterated function system is first proposed and an improved fractal interpolation model is then established. The improved fractal interpolation model and the SVM model are integrated to predict the every 5-min high frequency index data of Shanghai Composite Index. The experimental results show that the hybrid model is suitable for forecasting the stock index time series with fractal characteristics. In addition, a comparison of the prediction accuracy is carried out among the hybrid model and other three commonly used models. The results show that the prediction performance of the hybrid model is superior to that of other three models.

Sign in / Sign up

Export Citation Format

Share Document