Robust monitoring of small-scale forest disturbances in a tropical montane forest using Landsat time series
DeVries, B.R. ; Verbesselt, J. ; Kooistra, L. ; Herold, M. - \ 2015
Remote Sensing of Environment 161 (2015). - ISSN 0034-4257 - p. 107 - 121.
etm plus data - geostatistical approach - afromontane forests - southwest ethiopia - detecting trends - central-africa - satellite data - cover change - deforestation - imagery
Remote sensing data play an important role in the monitoring of forest changes. Methods are needed to provide objective estimates of forest loss to support monitoring efforts at various scales, and with increasing public availability of remote sensing data, accurate deforestation measurements at high temporal resolution are becoming more realistic. While several time series based methods have recently been described in the literature, there are few studies focusing on tropical forest areas, where low data availability and complex change processes present challenges to forest disturbance monitoring. Here, we present a robust data-driven method to track tropical deforestation and degradation based on Landsat time series data. Based on the previously reported Breaks For Additive Season and Trend Monitor (BFAST Monitor) method (Verbesselt etal., 2012), we show that BFAST Monitor, when applied to Landsat NDVI time series data using sequentially defined monitoring periods, can be used to track small-scale forest disturbances annually in an Afromontane forest system in southern Ethiopia. Using an ordinal logistic regression (OLR) approach, change magnitude, calculated based on differences between observed and expected values in a monitoring period, was found to be an essential predictor variable for disturbances. After applying a NDVI change magnitude threshold of - 0.065, overall accuracy was estimated to be 78%, and both producer's and user's accuracy of the disturbance class were estimated to be 73%. The method and results presented here are relevant to tropical countries engaged in REDD + for whom data availability and complex forest change dynamics limit the ability to reliably track forest disturbances over time.
Assessing global land cover reference datasets for different user communities
Tsendbazar, N.E. ; Bruin, S. de; Herold, M. - \ 2015
ISPRS Journal of Photogrammetry and Remote Sensing 103 (2015). - ISSN 0924-2716 - p. 93 - 114.
classification accuracy assessment - thematic map accuracy - validation data set - igbp discover - design - products - modis - challenges - imagery - area
Global land cover (GLC) maps and assessments of their accuracy provide important information for different user communities. To date, there are several GLC reference datasets which are used for assessing the accuracy of specific maps. Despite significant efforts put into generating them, their availability and role in applications outside their intended use have been very limited. This study analyses metadata information from 12 existing and forthcoming GLC reference datasets and assesses their characteristics and potential uses in the context of 4 GLC user groups, i.e., climate modellers requiring data on Essential Climate Variables (ECV), global forest change analysts, the GEO Community of Practice for Global Agricultural Monitoring and GLC map producers. We assessed user requirements with respect to the sampling scheme, thematic coverage, spatial and temporal detail and quality control of the GLC reference datasets. Suitability of the datasets is highly dependent upon specific applications by the user communities considered. The LC-CCI, GOFC-GOLD, FAO-FRA and Geo-Wiki datasets had the broadest applicability for multiple uses. The re-usability of the GLC reference datasets would be greatly enhanced by making them publicly available in an expert framework that guides users on how to use them for specific applications.
The impact of the means of context evocation on consumers' emotion associations towards eating occasions
Piqueras Fiszman, B. ; Jaeger, S.R. - \ 2014
Food Quality and Preference 37 (2014). - ISSN 0950-3293 - p. 61 - 70.
evoked consumption contexts - imagery - appropriateness - responses - questionnaires - satisfaction - attributes - experience - validity - ratings
The joint investigation of the product, the consumer, and the consumption context is necessary for furthering the understanding of eating occasions (snacks and main meals), including their construction and enjoyment. The study of people’s experience of eating occasions is less advanced than the understanding of acceptability, preference, and choice of individual food/beverage items and/or their combination in meals. The current research contributes to narrowing this gap by focusing on emotions as a dimension of eating experiences and enjoyment. Under evoked consumption contexts (breakfast, lunch, afternoon snack, dinner), the emotion associations for several products (potato crisps, chocolate brownie, and kiwifruit) were obtained from consumers (n = 399) using a questionnaire method. Emotion associations were explored in relation to: (1) the way in which the food stimulus was evaluated by participants (tasting food vs. seeing a food image); (2) the serving presentation of the food stimulus (image of food shown in isolation vs. image of food served on a plate with cutlery); and (3) the means in which the consumption context was evoked (written vs. written and pictorial). Consumers’ product emotion associations when tasting a food stimulus vs. seeing an image of the same food were highly similar. There was some evidence that more specific means of presenting the food stimuli (with tableware vs. without tableware) and consumption contexts (written and pictorially vs. written only) influenced perceived appropriateness of the product in the focal consumption context. This resulted, for example, in a higher frequency of use of negative emotion terms in the less appropriate consumption contexts. Overall, through the use of evoked consumption contexts this research has contributed new understanding of product-specific emotional associations during eating occasions from a methodological approach. In addition to the aforementioned results a more general finding was the apparent reliance by participants on past product experiences when completing the emotion questionnaire.
A lightweight hyperspectral mapping system and photogrammetric processing chain for unmanned aerial vehicles
Suomalainen, J.M. ; Anders, N.S. ; Iqbal, S. ; Roerink, G.J. ; Franke, G.J. ; Wenting, P.F.M. ; Hünniger, D. ; Bartholomeus, H. ; Becker, R. ; Kooistra, L. - \ 2014
Remote Sensing 6 (2014)11. - ISSN 2072-4292 - p. 11013 - 11030.
imagery - motion
During the last years commercial hyperspectral imaging sensors have been miniaturized and their performance has been demonstrated on Unmanned Aerial Vehicles (UAV). However currently the commercial hyperspectral systems still require minimum payload capacity of approximately 3 kg, forcing usage of rather large UAVs. In this article we present a lightweight hyperspectral mapping system (HYMSY) for rotor-based UAVs, the novel processing chain for the system, and its potential for agricultural mapping and monitoring applications. The HYMSY consists of a custom-made pushbroom spectrometer (400–950 nm, 9 nm FWHM, 25 lines/s, 328 px/line), a photogrammetric camera, and a miniature GPS-Inertial Navigation System. The weight of HYMSY in ready-to-fly configuration is only 2.0 kg and it has been constructed mostly from off-the-shelf components. The processing chain uses a photogrammetric algorithm to produce a Digital Surface Model (DSM) and provides high accuracy orientation of the system over the DSM. The pushbroom data is georectified by projecting it onto the DSM with the support of photogrammetric orientations and the GPS-INS data. Since an up-to-date DSM is produced internally, no external data are required and the processing chain is capable to georectify pushbroom data fully automatically. The system has been adopted for several experimental flights related to agricultural and habitat monitoring applications. For a typical flight, an area of 2–10 ha was mapped, producing a RGB orthomosaic at 1–5 cm resolution, a DSM at 5–10 cm resolution, and a hyperspectral datacube at 10–50 cm resolution.
Segmentation of Rumex obtusifolius using Gaussian Markov random fields
Atni Hiremath, S. ; Tolpekin, V.A. ; Heijden, G. van der; Stein, A. - \ 2013
Machine Vision Applications 24 (2013)4. - ISSN 0932-8092 - p. 845 - 854.
energy minimization - texture features - weed-control - graph cuts - classification - systems - imagery - vision
Rumex obtusifolius is a common weed that is difficult to control. The most common way to control weeds-using herbicides-is being reconsidered because of its adverse environmental impact. Robotic systems are regarded as a viable non-chemical alternative for treating R. obtusifolius and also other weeds. Among the existing systems for weed control, only a few are applicable in real-time and operate in a controlled environment. In this study, we develop a new algorithm for segmentation of R. obtusifolius using texture features based on Markov random fields that works in real-time under natural lighting conditions. We show its performance by comparing it with an existing real-time algorithm that uses spectral power as texture feature. We show that the new algorithm is not only accurate with detection rate of 97.8 % and average error of 56 mm in estimating the location of the tap-root of the plant, but is also fast taking just 0.18 s to process an image of size pixels making it feasible for real-time applications.
Feature level fusion of multi-temporal ALOS PALSAR and Landsat data for mapping and monitoring of tropical deforestation and forest degradation
Reiche, J. ; Souza, C. ; Hoekman, D.H. ; Verbesselt, J. ; Haimwant, P. ; Herold, M. - \ 2013
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 6 (2013)5. - ISSN 1939-1404 - p. 2159 - 2173.
brazilian amazonia - sar - imagery - classification - emissions - countries - accuracy - band
Many tropical countries suffer from persistent cloud cover inhibiting spatially consistent reporting of deforestation and forest degradation for REDD+. Data gaps remain even when compositing Landsat-like optical satellite imagery over one or two years. Instead, medium resolution SAR is capable of providing reliable deforestation information but shows limited capacity to identify forest degradation. This paper describes an innovative approach for feature fusion of multi-temporal and medium-resolution SAR and optical sub-pixel fraction information. After independently processing SAR and optical input data streams the extracted SAR and optical sub-pixel fraction features are fused using a decision tree classifier. ALOS PALSAR Fine Bean Dual and Landsat imagery of 2007 and 2010 acquired over the main mining district in central Guyana have been used for a proof-of-concept demonstration observing overall accuracies of 88% and 89.3% formapping forest land cover and detecting deforestation and forest degradation, respectively. Deforestation and degradation rates of 0.1% and 0.08% are reported for the observation period. Data gaps due to mainly clouds and Landsat ETM+ SLC-off that remained after compositing a set of single-period Landsat scenes, but also due to SAR layover and shadow could be reduced from 7.9% to negligible 0.01% while maintaining the desired thematic detail of detecting deforestation and degradation. The paper demonstrates the increase of both spatial completeness and thematic detail when applying the methodology, compared with potential Landsat-only or PALSAR-only approaches for a heavy cloud contaminated tropical environment. It indicates the potential for providing the required accuracy of activity data for REDD+ MRV.
Robust pixel-based classification of obstacles for robotic harvesting of sweet-pepper
Bac, C.W. ; Hemming, J. ; Henten, E. van - \ 2013
Computers and Electronics in Agriculture 96 (2013). - ISSN 0168-1699 - p. 148 - 162.
machine vision - selection - identification - performance - features - imagery - trees - fruit - parts
Sweet-pepper plant parts should be distinguished to construct an obstacle map to plan collision-free motion for a harvesting manipulator. Objectives were to segment vegetation from the background; to segment non-vegetation objects; to construct a classifier robust to variation among scenes; and to classify vegetation primarily into soft (top of a leaf, bottom of leaf and petiole) and hard obstacles (stem and fruit) and secondarily into five plant parts: stem, top of a leaf, bottom of a leaf, fruit and petiole. A multi-spectral system with artificial lighting was developed to mitigate disturbances caused by natural lighting conditions. The background was successfully segmented from vegetation using a threshold in a near-infrared wavelength (>900 nm). Non-vegetation objects occurring in the scene, including drippers, pots, sticks, construction elements and support wires, were removed using a threshold in the blue wavelength (447 nm). Vegetation was classified, using a Classification and Regression Trees (CART) classifier trained with 46 pixel-based features. The Normalized Difference Index features were the strongest as selected by a Sequential Floating Forward Selection algorithm. A new robust-and-balanced accuracy performance measure PRob was introduced for CART pruning and feature selection. Use of PRob rendered the classifier more robust to variation among scenes because standard deviation among scenes reduced 59% for hard obstacles and 43% for soft obstacles compared with balanced accuracy. Two approaches were derived to classify vegetation: Approach A was based on hard vs. soft obstacle classification and Approach B was based on separability of classes. Approach A (PRob = 58.9) performed slightly better than Approach B (PRob = 56.1). For Approach A, mean true-positive detection rate (standard deviation) among scenes was 59.2 (7.1)% for hard obstacles, 91.5 (4.0)% for soft obstacles, 40.0 (12.4)% for stems, 78.7 (16.0)% for top of a leaf, 68.5 (11.4)% for bottom of a leaf, 54.5 (9.9)% for fruit and 49.5 (13.6)% for petiole. These results are insufficient to construct an accurate obstacle map and suggestions for improvements are described. Nevertheless, this is the first study that reports quantitative performance for classification of several plant parts under varying lighting conditions.
|Magnetic Resonance in Food Science - Food for Thought
Duynhoven, J.P.M. van; Belton, P.S. ; Webb, G.A. ; As, H. van - \ 2013
London : RSC Books - ISBN 9781849736343 - 235
voedselwetenschappen - voedingsmiddelen - kernspintomografie - diagnostische technieken - voedselverwerking - spectroscopie - afbeelden - voedselkwaliteit - voedselveiligheid - food sciences - foods - magnetic resonance imaging - diagnostic techniques - food processing - spectroscopy - imagery - food quality - food safety
There are many challenges and problems in food science and magnetic resonance methods may be used to provide answers and deepen both fundamental and practical knowledge. This book presents innovations in magnetic resonance and in particular applications to understanding the functionality of foods, their processing and stability and their impact on health, perception and behaviour. Coverage includes structure and function, emphasizing respectively applications of spectroscopy/relaxometry and imaging/diffusometry; high resolution NMR spectroscopy as applied to quality and safety and foodomics; and dedicated information on perception and behaviour demonstrating the progress that has been made in applications of fMRI in this field.
Mapping tropical forest trees using high-resolution aerial digital photographs
Garzon-Lopez, C.X. ; Bohlman, S.A. ; Olff, H. ; Jansen, P.A. - \ 2013
Biotropica 45 (2013)3. - ISSN 0006-3606 - p. 308 - 316.
rain-forest - spatial-patterns - scale - dispersal - imagery - identification - biodiversity - limitation - management - dynamics
The spatial arrangement of tree species is a key aspect of community ecology. Because tree species in tropical forests occur at low densities, it is logistically challenging to measure distributions across large areas. In this study, we evaluated the potential use of canopy tree crown maps, derived from high-resolution aerial digital photographs, as a relatively simple method for measuring large-scale tree distributions. At Barro Colorado Island, Panama, we used high-resolution aerial digital photographs (~0.129 m/pixel) to identify tree species and map crown distributions of four target tree species. We determined crown mapping accuracy by comparing aerial and ground-mapped distributions and tested whether the spatial characteristics of the crown maps reflect those of the ground-mapped trees. Nearly a quarter (22%) of the common canopy species had sufficiently distinctive crowns to be good candidates for reliable mapping. The errors of commission (crowns misidentified as a target species) were relatively low, but the errors of omission (missed canopy trees of the target species) were high. Only 40 percent of canopy individuals were mapped on the air photographs. Despite failing to accurately predict exact abundances of canopy trees, crown distributions accurately reproduced the clumping patterns and spatial autocorrelation features of three of four tree species and predicted areas of high and low abundance. We discuss a range of ecological and forest management applications for which this method can be useful.
Heathland conservation status mapping through integration of hyperspectral mixture analysis and decision tree classifiers
Delalieux, S. ; Somers, B. ; Haest, B. ; Mücher, C.A. - \ 2012
Remote Sensing of Environment 126 (2012). - ISSN 0034-4257 - p. 222 - 231.
land-cover classification - vegetation - habitats - imagery - accuracy - indicators - management - selection - ecology - wetland
Monitoring the conservation status of natural habitats is an essential aspect of effective conservation management. Not only data on habitat occurrence are needed, but also detailed information on the structural and functional characteristics of the habitat patches is crucial for an adequate conservation status assessment. Classification of hyperspectral remote sensing images performs well in discriminating dominant land cover and vegetation classes, but the accuracy drops significantly for the classification of more subtle differences in conservation status that are related to structural characteristics. This study proposes a method to facilitate ecological conservation status assessment based on decision tree modeling of subpixel fraction estimates steered by ecological expert knowledge. In particular, it contributes to the spatially explicit assessment of an important structural aspect of dry heathland vegetation, namely the heather age structure, using Airborne Hyperspectral line-Scanner radiometer (AHS-160) data of the Kalmthoutse Heide in northern Belgium. We implemented a subpixel unmixing approach to identify the percentage of heather, sand and shadow in each heather pixel, and subsequently applied a decision tree classification to allocate each pixel to a certain age class. As such, our method provides a tool that contributes to the information required for an appropriate management and successful conservation of natural heathlands.
Same-different reaction times to odors: some unexpected findings
Moeller, P. ; Koester, E.P. ; Dijkman, N. ; Wijk, R.A. de; Mojet, J. - \ 2012
Chemosensory Perception 5 (2012)2. - ISSN 1936-5802 - p. 158 - 171.
perceived fragrance complexity - incidental-learning experiment - evoked memories - flavor memory - food memory - identification - perception - imagery - pleasantness - familiarity
Two experiments were carried out using olfactometers that delivered two stimuli with an interval of, respectively, 0.2 s (experiment 1) and 4.0 s (experiment 2) in a same–different paradigm. In experiment 1 (four men, age 38.5¿±¿15.2 and six women, age 25.8¿±¿1.2), four odors and in experiment 2 (nine men, age 23.4¿±¿2.6 and ten women, age 22.7¿±¿1.9), another eight odors were used in all pairs and pair-member orders. Subjects received each combination twice and responded as soon as possible after arrival of the second stimulus. Pair member similarity and odor pleasantness were measured in experiment 1 and odor complexity, familiarity, pleasantness, and self-reported odor imagining ability (high vs. low) in experiment 2. Results showed three independent effects: (1) “Same” responses took longer than “different” responses. (2) High imagers reacted faster than low imagers. (3) Reversing pair member order led to non-reciprocal similarity and reaction times. In different odor pairs, similarity and reaction time (Rt) correlated strongly and prime-familiarity and Rt correlated negatively. Edibility had an effect via prime-familiarity. Pleasantness had an effect only when a less pleasant odor followed a more pleasant one. All these latter effects were unrelated to the effects of participants’ imaging ability.
Using a genetic algorithm as an optimal band selector in the mid and thermal infrared (2.5-14 µm) to discriminate vegetation species
Ullah, S. ; Groen, T.A. ; Schlerf, M. ; Skidmore, A.K. ; Nieuwenhuis, W. ; Vaiphasa, C. - \ 2012
Sensors 12 (2012)7. - ISSN 1424-8220 - p. 8755 - 8769.
spectral discrimination - reflectance - spectroscopy - emissivity - imagery - leaves - identification - spectrometry - regression - plants
Genetic variation between various plant species determines differences in their physio-chemical makeup and ultimately in their hyperspectral emissivity signatures. The hyperspectral emissivity signatures, on the one hand, account for the subtle physio-chemical changes in the vegetation, but on the other hand, highlight the problem of high dimensionality. The aim of this paper is to investigate the performance of genetic algorithms coupled with the spectral angle mapper (SAM) to identify a meaningful subset of wavebands sensitive enough to discriminate thirteen broadleaved vegetation species from the laboratory measured hyperspectral emissivities. The performance was evaluated using an overall classification accuracy and Jeffries Matusita distance. For the multiple plant species, the targeted bands based on genetic algorithms resulted in a high overall classification accuracy (90%). Concentrating on the pairwise comparison results, the selected wavebands based on genetic algorithms resulted in higher Jeffries Matusita (J-M) distances than randomly selected wavebands did. This study concludes that targeted wavebands from leaf emissivity spectra are able to discriminate vegetation species.
A New Minimum-Volume Enclosing Algorithm for Endmember Identification and Abundance Estimation in Hyperspectral Data
Hendrix, E.M.T. ; García, I. ; Plaza, J. ; Martín, G. ; Plaza, A. - \ 2012
IEEE Transactions on Geoscience and Remote Sensing 50 (2012)7. - ISSN 0196-2892 - p. 2744 - 2757.
nonnegative matrix factorization - spectral mixture analysis - imaging spectroscopy - n-findr - extraction - imagery
Spectral unmixing is an important technique for hyperspectral data exploitation, in which a mixed spectral signature is decomposed into a collection of spectrally pure constituent spectra, called endmembers, and a set of correspondent fractions, or abundances, that indicate the proportion of each endmember present in the mixture. Over the last years, several algorithms have been developed for automatic or semiautomatic endmember extraction. Some available approaches assume that the input data set contains at least one pure spectral signature for each distinct material and further conduct a search for the most spectrally pure signatures in the high-dimensional space spanned by the hyperspectral data. Among these approaches, those aimed at maximizing the volume of the simplex that can be formed using available spectral signatures have found wide acceptance. However, the presence of spectrally pure constituents is unlikely in remotely sensed hyperspectral scenes due to spatial resolution, mixing phenomena, and other considerations. In order to address this issue, other available algorithms have been developed to generate virtual endmembers (not necessarily present among the input data samples) by finding the simplex with minimum volume that encloses all available observations. In this paper, we discuss maximum-volume versus minimum-volume enclosing solutions and further develop a novel algorithm in the latter category which incorporates the fractional abundance estimation as an internal step of the endmember searching process (i.e., it does not require an external method to produce endmember fractional abundances). The method is based on iteratively enclosing the observations in a lower dimensional space and removing observations that are most likely not to be enclosed by the simplex of the endmembers to be estimated. The performance of the algorithm is investigated and compared to that of other algorithms (with and without the pure pixel assumption) using synthetic a- d real hyperspectral data sets collected by a variety of hyperspectral imaging instruments.
Mapping invasive woody species in coastal dunes in the Netherlands: a remote sensing approach using LIDAR and high-resolution aerial photographs
Hantson, W.P.R. ; Kooistra, L. ; Slim, P.A. - \ 2012
Applied Vegetation Science 15 (2012)4. - ISSN 1402-2001 - p. 536 - 547.
hippophae-rhamnoides l - rosa-rugosa - vegetation - imagery - laser - classifications - encroachment - diversity - expansion - ecosystem
Questions Does remote sensing improve classification of invasive woody species in dunes, useful for shrub management? Does additional height information and an object-based classifier increase woody species classification accuracy? Location The dunes of Vlieland, one of the Wadden Sea Islands, the Netherlands. Methods Extensive monitoring using optical remote sensing and LIDAR deliver large amounts of high-quality data to observe and manage coastal dunes as a defence against the sea in the Netherlands. Using these additional data could increase the accuracy of vegetation mapping and monitoring in coastal areas. In this study, a remote sensing approach has been developed to deliver detailed and standardized maps of (invasive) woody species in the dunes of Vlieland using multispectral aerial photographs and vegetation height derived from LIDAR. Three classification methods were used: maximum likelihood (ML) classification using aerial photographs, ML classification combined with vegetation heights derived from LIDAR (ML+) and object-based (OB) classification. Results The use of vegetation height from the LIDAR data increased the overall classification accuracy from 39% to 50%, but particularly improved classification of the taller woody species. The object-based classification increased the overall accuracy of the ML+ from 50% to 60%. The object-based results are comparable to human visual analysis while offering automated analysis. Conclusions Overall, the object-based classification delivers detailed maps of the woody species that are useful for management and evaluation of alien and invasive species in dune ecosystems.
Integration of multi-sensor data to assess grassland dynamics in a Yellow River sub-watershed
Ouyang, W. ; Hao, F. ; Skidmore, A.K. ; Groen, T.A. ; Toxopeus, A.G. ; Wang, T. - \ 2012
Ecological Indicators 18 (2012)1. - ISSN 1470-160X - p. 163 - 170.
qinghai-xizang plateau - time-series - land-cover - west-africa - vegetation - modis - variability - patterns - imagery - china
Grasslands form the dominant land cover in the upper reaches of the Yellow River and provide a reliable indicator by being strongly correlated with regional terrestrial ecological status. Remote sensing can provide information useful for vegetation quality assessments, but no single sensor can meet the needs for the high temporal-spatial resolution required for such assessments on a watershed scale. To observe long-term grassland dynamics in the Longliu Watershed located in the upper reaches of the Yellow River, Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat images were integrated to obtain Normalized Difference Vegetation Index (NDVI) data. The MODIS images were used to identify patterns of monthly variation. With the temporal dynamics of NDVI provided by the MODIS images, an exponential regression model was obtained that described the relationship between Julian day and NDVI. Four time-series data sets from multi-spectral sensors were constructed to obtain regional grassland NDVI information from 1977 to 2006 in the Longliu Watershed. Using the daily NDVI correlation coefficient, NDVI values for different days were obtained from Landsat series images, standardised to the same day and integrated into TM format by using NDVI coefficients between the four different sensors. Thus, the NDVI data obtained from multi-sensors on different days were integrated into a comparable format. A regression analysis correlating the NDVI data from two sensors with fresh grass biomass showed that the integration procedure was reliable.
Capabilities and limitations of Landsat and land cover data for aboveground woody biomass estimation of Uganda
Avitabile, V. ; Baccini, A. ; Friedl, M.A. ; Schmullius, C. - \ 2012
Remote Sensing of Environment 117 (2012). - ISSN 0034-4257 - p. 366 - 380.
tropical forest biomass - thematic mapper data - inventory data - tm data - brazilian amazon - satellite estimation - stand structure - carbon balance - etm+ data - imagery
Aboveground woody biomass for circa-2000 is mapped at national scale in Uganda at 30-m spatial resolution on the basis of Landsat ETM + images, a National land cover dataset and field data using an object-oriented approach. A regression tree-based model (Random Forest) produces good results (cross-validated R² 0.81, RMSE 13 T/ha) when trained with a sufficient number of field plots representative of the vegetation variability at national scale. The Random Forest model captures non-linear relationships between satellite data and biomass density, and is able to use categorical data (land cover) in the regression to improve the results. Biomass estimates were strongly correlated (r = 0.90 and r = 0.83) with independent LiDAR measurements. In this study, we demonstrate that in certain contexts Landsat data provide the capability to spatialize field biomass measurements and produce accurate and detailed estimates of biomass distribution at national scale. We also investigate limitations of this approach, which tend to provide conservative biomass estimates. Specific limitations are mainly related to saturation of the optical signal at high biomass density and cloud cover, which hinders the compilation of a radiometrically consistent multi-temporal dataset. As a result, a Landsat mosaic created for Uganda with images acquired in the dry season during 1999–2003 does not contain phenological information useful for discriminating some vegetation types, such as deciduous formations. The addition of land cover data increases the model performance because it provides information on vegetation phenology. We note that Landsat data present higher spatial and thematic resolution compared to land cover and allow detailed and spatially continuous biomass estimates to be mapped. Fusion of satellite and ancillary data may improve biomass predictions but, to avoid error propagation, accurate, detailed and up-to-date land cover or other ancillary data are necessary. --------------------------------------------------------------------------------
Identifying plant species using mid-wave infrared (2.5-6µm) and thermal infrared (8-14µm) emissivity spectra
Ullah, S. ; Schlerf, M. ; Skidmore, A.K. ; Hecker, C. - \ 2012
Remote Sensing of Environment 118 (2012)4. - ISSN 0034-4257 - p. 95 - 102.
salt-marsh vegetation - hyperspectral data - biomass estimation - reflectance - discrimination - indexes - imagery - leaves - classification - spectroscopy
Plant species discrimination using remote sensing is generally limited by the similarity of their reflectance spectra in the visible, NIR and SWIR domains. Laboratory measured emissivity spectra in the mid infrared (MIR; 2.5µm-6µm) and the thermal infrared (TIR; 8µm-14µm) domain of different plant species, however, reveal significant differences. It is anticipated that with the advances in airborne and space borne hyperspectral thermal sensors, differentiation between plant species may improve. The laboratory emissivity spectra of thirteen common broad leaved species, comprising 3024 spectral bands in the MIR and TIR, were analyzed. For each wavelength the differences between the species were tested for significance using the one way analysis of variance (ANOVA) with the post-hoc Tukey HSD test. The emissivity spectra of the analyzed species were found to be statistically different at various wavebands. Subsequently, six spectral bands were selected (based on the histogram of separable pairs of species for each waveband) to quantify the separability between each species pair based on the Jefferies Matusita (JM) distance. Out of 78 combinations, 76 pairs had a significantly different JM distance. This means that careful selection of hyperspectral bands in the MIR and TIR (2.5µm-14µm) results in reliable species discrimination.
Penalized regression techniques for prediction: a case study for predicting tree mortality using remotely sensed vegetation indices
Lazaridis, D.C. ; Verbesselt, J. ; Robinson, A.P. - \ 2011
Canadian Journal of Forest Research 41 (2011)1. - ISSN 0045-5067 - p. 24 - 34.
nonorthogonal problems - hyperspectral data - ridge regression - cross-validation - lasso - infestation - shrinkage - selection - forests - imagery
Constructing models can be complicated when the available fitting data are highly correlated and of high dimension. However, the complications depend on whether the goal is prediction instead of estimation. We focus on predicting tree mortality (measured as the number of dead trees) from change metrics derived from moderate-resolution imaging spectroradiometer satellite images. The high dimensionality and multicollinearity inherent in such data are of particular concern. Standard regression techniques perform poorly for such data, so we examine shrinkage regression techniques such as ridge regression, the LASSO, and partial least squares, which yield more robust predictions. We also suggest efficient strategies that can be used to select optimal models such as 0.632+ bootstrap and generalized cross validation. The techniques are compared using simulations. The techniques are then used to predict insect-induced tree mortality severity for a Pinus radiata D. Don plantation in southern New South Wales, Australia, and their prediction performances are compared. We find that shrinkage regression techniques outperform the standard methods, with ridge regression and the LASSO performing particularly well.
Leaf area index estimation with MODIS reflectance time series and model inversion during full rotations of Eucalyptus plantations
Maire, G. Le; Marsden, C. ; Verhoef, W. ; Ponzoni, F.J. ; Seen, D. Lo; Bégué, A. ; Stape, J.L. ; Nouvellon, Y. - \ 2011
Remote Sensing of Environment 115 (2011)2. - ISSN 0034-4257 - p. 586 - 599.
optical-properties - vegetation index - bidirectional reflectance - use efficiency - canopy - imagery - forest - chlorophyll - resolution - globulus
The leaf area index (LAI) of fast-growing Eucalyptus plantations is highly dynamic both seasonally and inter-annually, and is spatially variable depending on pedo-climatic conditions. LAI is very important in determining the carbon and water balance of a stand, but is difficult to measure during a complete stand rotation and at large scales. Remote-sensing methods allowing the retrieval of LAI time series with accuracy and precision are therefore necessary. Here, we tested two methods for LAI estimation from MODIS 250m resolution red and near-infrared (NIR) reflectance time series. The first method involved the inversion of a coupled model of leaf reflectance and transmittance (PROSPECT4), soil reflectance (SOILSPECT) and canopy radiative transfer (4SAIL2). Model parameters other than the LAI were either fixed to measured constant values, or allowed to vary seasonally and/or with stand age according to trends observed in field measurements. The LAI was assumed to vary throughout the rotation following a series of alternately increasing and decreasing sigmoid curves. The parameters of each sigmoid curve that allowed the best fit of simulated canopy reflectance to MODIS red and NIR reflectance data were obtained by minimization techniques. The second method was based on a linear relationship between the LAI and values of the GEneralized Soil Adjusted Vegetation Index (GESAVI), which was calibrated using destructive LAI measurements made at two seasons, on Eucalyptus stands of different ages and productivity levels. The ability of each approach to reproduce field-measured LAI values was assessed, and uncertainty on results and parameter sensitivities were examined. Both methods offered a good fit between measured and estimated LAI (R2 = 0.80 and R2 = 0.62 for model inversion and GESAVI-based methods, respectively), but the GESAVI-based method overestimated the LAI at young ages.
Update Beeldenbank : ziekten, plagen en onkruiden
Os, G.J. van - \ 2011
plantenziekten - plantenplagen - databanken - informatiesystemen - afbeelden - informatieontsluiting - agrarisch onderwijs - kennisoverdracht - plant diseases - plant pests - databases - information systems - imagery - information retrieval - agricultural education - knowledge transfer
Ca. 250 nieuwe items (foto’s met bijbehorende informatie over symptomen, oorzaak en maatregelen) over herbicideschade, gebreksziekten en natuurlijke vijanden worden toegevoegd aan de openbare beeldenbank ziekten, plagen en onkruiden. Behalve voor onderwijsdoeleinden is de beeldenbank ook interessant voor boeren, telers, voorlichters en particulieren. In mei 2008 is de beeldenbank publiek ontsloten op: www.groenkennisnet.nl