Revista de Teledetección - Núm. 49 (2017) Special Issue (2016)

Tabla de contenidos

Invited articles

A review of the role of active remote sensing and data fusion for characterizing forest in wildlife habitat models
Classification of forest development stages from national low-density lidar datasets: a comparison of machine learning methods

Research articles

A comparative study of regression methods to predict forest structure and canopy fuel variables from LiDAR full-waveform data
Modelling canopy fuel and forest stand variables and characterizing the influence of thinning in the stand structure using airborne LiDAR
Characterization of wildland-urban interfaces using LiDAR data to estimate the risk of wildfire damage
Determination of forest biomass using remote sensing techniques with radar images. Pilot study in area of the province of Huelva. REDIAM

Practical cases

Combined use of LIDAR and hyperspectral measurements for remote sensing of fluorescence and vertical profile of canopies

URI permanente para esta colecciónhttps://riunet.upv.es/handle/10251/80387

Examinar

Mostrando 1 - 7 de 7

Access status: Abierto ,
Combined use of LIDAR and hyperspectral measurements for remote sensing of fluorescence and vertical profile of canopies
(Universitat Politècnica de València, 2016-02-26) Ounis, A.; Bach, J.; Mahjoub, A.; Daumard, F.; Moya, I.; Goulas, Y.; Agence Nationale de la Recherche, Francia; Centre National d'Études Spatiales, Francia
[EN] We report the development of a new LIDAR system (LASVEG) for airborne remote sensing of chlorophyll fluorescence (ChlF) and vertical profile of canopies. By combining laser-induced fluorescence (LIF), sun-induced fluorescence (SIF) and canopy height distribution, the new instrument will allow the simultaneous assessment of gross primary production (GPP), photosynthesis efficiency and above ground carbon stocks. Technical issues of the fluorescence LIDAR development are discussed and expected performances are presented.
Access status: Abierto ,
Determination of forest biomass using remote sensing techniques with radar images. Pilot study in area of the province of Huelva. REDIAM
(Universitat Politècnica de València, 2016-02-26) Méndez, E.; Vales, J. J.; Pino, I.; Granado, L.; Montoya, G.; Prieto, R.; Carpintero, I. R.; Giménez de Azcárate, F.; Cáceres, F.; Moreira, J. M.; de la Fuente, D.; Sebastián, A.; Suárez, J.
[EN] Biomass is a very important forest resource in Andalusia. “Forest Biomass in Andalusia” web tool, developed by the Andalusian Government, provides information about the location and biomass stock for the main pine forest species. It is important to mention that information needs to be regularly, quickly, effectively and inexpensively updated. These requirements could be covered with the help of Earth Observation technologies. In this project, radar images have been acquired from ALOS-PALSAR sensor from different years (2008 and 2010) over two pilot areas located in Huelva. The aim of the study has been to develop a methodology to estimate wood volumes based on the statistic correlation between radar signal and wood volume, variable extracted of forest management plans contemporary to images. As result, correlations of 0.8 and 0.7 have been obtained for pine and eucalyptus respectively. Forest biomass has been calculated using species-specific allometric equations. Three key sources of information have been used: a sample of plots distributed homogeneously, an accurate digital terrain model and a current forest map. Furthermore, the study of the variability of estimated volumes between these dates has been carried out. Methodologies obtained could be extrapolated to the whole region.
Access status: Abierto ,
Characterization of wildland-urban interfaces using LiDAR data to estimate the risk of wildfire damage
(Universitat Politècnica de València, 2016-02-26) Robles, A.; Rodríguez-Garrido, M. A.; Alvarez-Taboada, M. F.
[EN] Galicia is a region in NW Spain which is usually affected by a high number of forest fires, and it should meet the current regulations regarding the distance between forests and buildings. This paper aims to identify and characterize woodlands and classify buildings according to their fire risk, for a 36 km2 area in Forcarei (Pontevedra, Spain). We used LiDAR data to generate three spatial models (DTM: Digital Terrain Model, DSM: Digital Surface Model and nDSM: Normalized Digital Surface Model) and two statistics to characterize the forest stands (density of dominant trees per hectare and their average height). The identification of forested areas was performed using an object-based classification method using the intensity image, the height model and an orthophotograph of the area, and a kappa coefficient of 0.82 was obtained in the validation. The woodlands were reclassified according to the magnitude of a possible fire, based on the density and the average height of the woodlands. The forest stands were mapped according to the magnitude of a possible fire and it was found that 1.18 km2 would be susceptible to a low magnitude fire, 3.75 km2 to a medium magnitude fire and 2.25 km2 to a fire of a high magnitude. Afterwards, it was determined whether the buildings in the area complied with the legislation relating to minimum distance from the forested areas (30 meters). For those that did not meet this distance, the risk of damage in case of a wildfire was calculated. The result was that 43.01% of buildings in the area complied with the regulations, 9.95% were located in a very low risk area, 25.74% in a low risk location, 12.37% in a medium risk area and 8.93% were in a high or very high risk area.
Access status: Abierto ,
Modelling canopy fuel and forest stand variables and characterizing the influence of thinning in the stand structure using airborne LiDAR
(Universitat Politècnica de València, 2016-02-26) Hevia, A.; Álvarez-González, J. G.; Ruiz-Fernández, E.; Prendes, C.; Ruiz-González, A. D.; Majada, J.; González-Ferreiro, E.; European Regional Development Fund; Gobierno del Principado de Asturias; Xunta de Galicia
[EN] Forest fires are a major threat in NW Spain. The importance and frequency of these events in the area suggests the need for fuel management programs to reduce the spread and severity of forest fires. Thinning treatments can contribute for fire risk reduction, because they cut off the horizontal continuity of forest fuels. Besides, it is necessary to conduct a fire risk management based on the knowledge of fuel allocation, since fire behaviour and fire spread study is dependent on the spatial factor. Therefore, mapping fuel for different silvicultural scenarios is essential. Modelling forest variables and forest structure parameters from LiDAR technology is the starting point for developing spatially explicit maps. This is essential in the generation of fuel maps since field measurements of canopy fuel variables is not feasible. In the present study, we evaluated the potential of LiDAR technology to estimate canopy fuel variables and other stand variables, as well as to identify structural differences between silvicultural managed and unmanaged P. pinaster Ait. stands. Independent variables (LiDAR metrics) of greater explanatory significance were identified and regression analyses indicated strong relationships between those and field-derived variables (R2 varied between 0.86 and 0.97). Significant differences were found in some LiDAR metrics when compared thinned and unthinned stands. Results showed that LiDAR technology allows to model canopy fuel and stand variables with high precision in this species, and provides useful information for identifying areas with and without silvicultural management.
Access status: Abierto ,
A comparative study of regression methods to predict forest structure and canopy fuel variables from LiDAR full-waveform data
(Universitat Politècnica de València, 2016-02-26) Crespo-Peremarch, P.; Ruiz, L.A.; Balaguer-Beser, A.; Ministerio de Economía y Competitividad; European Regional Development Fund
[EN] Regression methods are widely employed in forestry to predict and map structure and canopy fuel variables. We present a study where several regression models (linear, non-linear, regression trees and ensemble) were assessed. Independent variables were calculated using metrics extracted from full-waveform LiDAR data, while the reference data used to generate the dependent variables for the prediction models were obtained from fieldwork in 78 plots of 16 m radius. Transformations of dependent and independent variables with feature selection were carried out to assess their influence in the prediction of response variables. In order to evaluate significant differences and rank regression models we used the non-parametric tests Wilcoxon and Friedman, and post-hoc analysis or post-hoc pairwise multiple comparison tests, such as Nemenyi, for Friedman test. Regressions using transformation of the dependent variable, like square-root or logarithmic, or the independent variable, increased R2 up to 6% with respect to linear regression using unprocessed response variables. CART (Classification and Regression Tree) method provided poor results, but it may be interesting for categorisation purposes. Square-root transformation of the dependent variable is the method having the best overall results, except for stand volume. However, not always has a significant improvement with respect to other regression methods.
Access status: Abierto ,
Classification of forest development stages from national low-density lidar datasets: a comparison of machine learning methods
(Universitat Politècnica de València, 2016-02-26) Valbuena, R.; Maltamo, M.; Packalen, P.; Finnish Forest Centre
[EN] The area-based method has become a widespread approach in airborne laser scanning (ALS), being mainly employed for the estimation of continuous variables describing forest attributes: biomass, volume, density, etc. However, to date, classification methods based on machine learning, which are fairly common in other remote sensing fields, such as land use / land cover classification using multispectral sensors, have been largely overseen in forestry applications of ALS. In this article, we wish to draw the attention on statistical methods predicting discrete responses, for supervised classification of ALS datasets. A wide spectrum of approaches are reviewed: discriminant analysis (DA) using various classifiers –maximum likelihood, minimum volume ellipsoid, naïve Bayes–, support vector machine (SVM), artificial neural networks (ANN), random forest (RF) and nearest neighbour (NN) methods. They are compared in the context of a classification of forest areas into development classes (DC) used in practical silvicultural management in Finland, using their low-density national ALS dataset. We observed that RF and NN had the most balanced error matrices, with cross-validated predictions which were mainly unbiased for all DCs. Although overall accuracies were higher for SVM and ANN, their results were very dissimilar across DCs, and they can therefore be only advantageous if certain DCs are targeted. DA methods underperformed in comparison to other alternatives, and were only advantageous for the detection of seedling stands. These results show that, besides the well demonstrated capacity of ALS for quantifying forest stocks, there is a great deal of potential for predicting categorical variables in general, and forest types in particular. In conclusion, we consider that the presented methodology shall also be adapted to the type of forest classes that can be relevant to Mediterranean ecosystems, opening a range of possibilities for future research, in which ALS may show great predictive potential.
Access status: Abierto ,
A review of the role of active remote sensing and data fusion for characterizing forest in wildlife habitat models
(Universitat Politècnica de València, 2016-02-26) Vogeler, J. C.; Cohen, W. B.; National Aeronautics and Space Administration, EEUU
[EN] Spatially explicit maps of wildlife habitat relationships have proven to be valuable tools for conservation and management applications including evaluating how and which species may be impacted by large scale climate change, ongoing fragmentation of habitat, and local land-use practices. Studies have turned to remote sensing datasets as a way to characterize vegetation for the examination of habitat selection and for mapping realized relationships across the landscape. Potentially one of the more difficult habitat types to try to characterize with remote sensing are the vertically and horizontally complex forest systems. Characterizing this complexity is needed to explore which aspects may represent driving and/or limiting factors for wildlife species. Active remote sensing data from lidar and radar sensors has thus caught the attention of the forest wildlife research and management community in its potential to represent three dimensional habitat features. The purpose of this review was to examine the applications of active remote sensing for characterizing forest in wildlife habitat studies through a keyword search within Web of Science. We present commonly used active remote sensing metrics and methods, discuss recent advances in characterizing aspects of forest habitat, and provide suggestions for future research in the area of new remote sensing data/techniques that could benefit forest wildlife studies that are currently not represented or may be underutilized within the wildlife literature. We also highlight the potential value in data fusion of active and passive sensor data for representing multiple dimensions and scales of forest habitat. While the use of remote sensing has increased in recent years within wildlife habitat studies, continued communication between the remote sensing, forest management, and wildlife communities is vital to ensure appropriate data sources and methods are understood and utilized, and so that creators of mapping products may better realize the needs of secondary users.

Examinar

Envíos recientes