Détail de l'autorité
|
DEEPSURF / Pironon, Jacques
Autorités liées :
Nom :
DEEPSURF
titre complet :
échanges de chaleur et de matière entre les compartiments géologiques profonds, la zone critique et la surface, lors de leur utilisation pour la transition énergétique
URL du projet :
Auteurs :
Pironon, Jacques
|
Documents disponibles (4)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externesMultisource forest inventories: A model-based approach using k-NN to reconcile forest attributes statistics and map products / Ankit Sagar in ISPRS Journal of photogrammetry and remote sensing, vol 192 (October 2022)
Titre : Multisource forest inventories: A model-based approach using k-NN to reconcile forest attributes statistics and map products Type de document : Article/Communication Auteurs : Ankit Sagar Année de publication : 2022 Projets : LUE / Université de Lorraine, ARBRE / AgroParisTech (2007 -), DEEPSURF / Pironon, Jacques Article en page(s) : pp 175 - 188 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] classification barycentrique
[Termes IGN] données allométriques
[Termes IGN] données lidar
[Termes IGN] image Landsat-8
[Termes IGN] inventaire forestier national (données France)
[Vedettes matières IGN] Inventaire forestierRésumé : (auteur) Forest map products are widely used and have taken benefit from progresses in the multisource forest inventory approaches, which are meant to improve the precision of forest inventory estimates at high spatial resolution. However, estimating errors of pixel-wise predictions remains difficult, and reconciling statistical outcomes with map products is still an open and important question. We address this problem using an original approach relying on a model-based inference framework and k-nearest neighbours (k-NN) models to produce pixel-wise estimations and related quality assessment. Our approach takes advantage of the resampling properties of a model-based estimator and combines it with geometrical convex-hull models to measure respectively the precision and accuracy of pixel predictions. A measure of pixel reliability was obtained by combining precision and accuracy. The study was carried out over a 7,694 km2 area dominated by structurally complex broadleaved forests in centre of France. The targeted forest attributes were growing stock volume, basal area and growing stock volume increment. A total of 819 national forest inventory plots were combined with auxiliary data extracted from a forest map, Landsat 8 images, and 3D point clouds from both airborne laser scanning and digital aerial photogrammetry. k-NN models were built independently for both 3D data sources. Both selected models included 5 auxiliary variables, and were generated using 5 neighbours, and most similar neighbours distance measure. The models showed relative root mean square error ranging from 35.7% (basal area, digital aerial photogrammetry) in calibration to 63.4% (growing stock volume increment, airborne laser scanning) in the validation set. At pixel level, we found that a minimum of 86.4% of the predictions were of high precision as their bootstrapped coefficient of variation fall below calibration’s relative root mean square error. The amount of extrapolation varied from 4.3% (digital aerial photogrammetry) to 6.3% (airborne laser scanning). A relationship was found between extrapolation and k-NN distance, opening new opportunities to correct extrapolation errors. At the population level, airborne laser scanning and digital aerial photogrammetry performed similarly, offering the possibility to use digital aerial photogrammetry for monitoring purposes. The proposed method provided consistent estimates of forest attributes and maps, and also provided spatially explicit information about pixel predictions in terms of precision, accuracy and reliability. The method therefore produced high resolution outputs, significant for either decision making or forest management purposes. Numéro de notice : A2022-629 Affiliation des auteurs : LIF+Ext (2020- ) Thématique : FORET Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2022.08.016 En ligne : https://doi.org/10.1016/j.isprsjprs.2022.08.016 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101495
in ISPRS Journal of photogrammetry and remote sensing > vol 192 (October 2022) . - pp 175 - 188[article]
Titre : Characterizing the calibration domain of remote sensing models using convex hulls Type de document : Article/Communication Auteurs : Jean-Pierre Renaud Année de publication : 2022 Projets : DEEPSURF / Pironon, Jacques, ARBRE / AgroParisTech (2007 -) Article en page(s) : n° 102939 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Statistiques
[Termes IGN] données allométriques
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] échantillonnage
[Termes IGN] erreur systématique
[Termes IGN] étalonnage de modèle
[Termes IGN] étalonnage des données
[Termes IGN] extrapolation
[Termes IGN] placette d'échantillonnageRésumé : (auteur) The ever-increasing availability of remote sensing data allows production of forest attributes maps, which are usually made using model-based approaches. These map products are sensitive to various bias sources, including model extrapolation. To identify, over a case study forest, the proportion of extrapolated predictions, we used a convex hull method applied to the auxiliary data space of an airborne laser scanning (ALS) flight. The impact of different sampling efforts was also evaluated. This was done by iteratively thinning a set of 487 systematic plots using nested sub-grids allowing to divide the sample by two at each level. The analysis were conducted for all alternative samples and evaluated against 56 independent validation plots. Residuals of the extrapolated validation plots were computed and examined as a function of their distance to the model calibration domain. Extrapolation was also characterized for the pixels of the area of interest (AOI) to upscale at population level. Results showed that the proportion of extrapolated pixels greatly reduced with an increasing sampling effort. It reached a plateau (ca. 20% extrapolation) with a sampling intensity of ca. 250-calibration plots. This contrasts with results on model’s root mean squared error (RMSE), which reached a plateau at a much lower sampling intensity. This result emphasizes the fact that with a low sampling effort, extrapolation risk remains high, even at a relatively low RMSE. For all attributes examined (i.e., stand density, basal area, and quadratic mean diameter) estimations were generally found to be biased for validation plots that were extrapolated. The method allows an easy identification of map pixels that are out of the calibration domain, making it an interesting tool to evaluate model transferability over an area of interest (AOI). It could also serve to compare “competing” models at a variable selection phase. From a model calibration perspective, it could serve a posteriori, to evaluate areas (in the auxiliary space) that merit further sampling efforts to improve model reliability. Numéro de notice : A2022-581 Affiliation des auteurs : LIF+Ext (2020- ) Thématique : FORET/IMAGERIE/MATHEMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.jag.2022.102939 Date de publication en ligne : 28/07/2022 En ligne : https://doi.org/10.1016/j.jag.2022.102939 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101341
in International journal of applied Earth observation and geoinformation > vol 112 (August 2022) . - n° 102939[article]
Titre : Convex hull: another perspective about model predictions and map derivatives from remote sensing data Type de document : Article/Communication Auteurs : Jean-Pierre Renaud Editeur : Vienne [Autriche] : Technische Universität Wien Année de publication : 2021 Collection : Geowissenschaftliche Mitteilungen, ISSN 1811-8380 num. 104 Projets : ARBRE / AgroParisTech (2007 -) Conférence : SilviLaser 2021, 17th conference on Lidar Applications for Assessing and Managing Forest Ecosystems 28/09/2021 30/09/2021 Vienne + online Autriche open access proceedings Projets : DEEPSURF / Pironon, Jacques Importance : pp 71 - 73 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie
[Termes IGN] attribut non spatial
[Termes IGN] convexité
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] échantillonnage
[Termes IGN] erreur systématique
[Termes IGN] modèle de simulation
[Termes IGN] modèle linéaireMots-clés libres : enveloppe convexe Résumé : (auteur) [introduction] In forest inventories as well as in the process of building models, obtaining an efficient sample is a central goal to reach precise estimates of forest attributes (Hawbaker et al. 2009, Frazer et al. 2011, Grafström et al. 2014, Saarela et al. 2015, Bouvier et al. 2019). In a model-based approach, a plots sample must cover adequately the variability of the considered forest attributes in order to minimise prediction error. Different strategies have been proposed to efficiently distribute the field sampling units in the auxiliary space of the remote sensing data (e.g. Hawbaker et al. 2009, Grafström et al. 2014). Some authors have proposed to stratify Airborne Laser Scanning data (ALS) to optimize sampling (Hawbaker et al. 2009, Frazer et al. 2011), and Maltamo et al. (2011) compared different field plot selection strategies in order to optimise models precision. Interestingly, White et al. (2013) applied convex hull approach to show uncovered forest structures by the field calibration sampling units, since large prediction errors could be associated with model extrapolations, resulting in potentially biased map derivatives. In this research, we use convex hull to identify the proportion of extrapolated pixels, computed their distance to the calibration domain and estimated bias associated to the linear model predictions on an ALS case study. Numéro de notice : C2021-030 Affiliation des auteurs : LIF+Ext (2020- ) Thématique : FORET/IMAGERIE/MATHEMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.34726/wim.1919 Date de publication en ligne : 01/12/2021 En ligne : https://doi.org/10.34726/wim.1919 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98997 
Titre : High resolution mapping of forest resources and prediction reliability using multisource inventory approach Type de document : Article/Communication Auteurs : Ankit Sagar Editeur : Vienne [Autriche] : Technische Universität Wien Année de publication : 2021 Collection : Geowissenschaftliche Mitteilungen, ISSN 1811-8380 num. 104 Projets : ARBRE / AgroParisTech (2007 -) Conférence : SilviLaser 2021, 17th conference on Lidar Applications for Assessing and Managing Forest Ecosystems 28/09/2021 30/09/2021 Vienne + online Autriche open access proceedings, INCA 2021, 41th Indian National Cartographic Association international conference, Cartography for self-reliant India 27/10/2021 29/10/2021 Chandigarh Inde open access proceedings Projets : DEEPSURF / Pironon, Jacques Importance : pp 219 - 221 Langues : Anglais (eng) Descripteur : [Termes IGN] capital sur pied
[Termes IGN] données multisources
[Termes IGN] inventaire forestier (techniques et méthodes)
[Termes IGN] ressources forestières
[Vedettes matières IGN] Inventaire forestierRésumé : (auteur) [introduction] National forest inventory (NFI) provides precise forest resource estimates at national up to regional scale but could not support local estimates with high precision because of inadequate number of field plots. The forest managers and stakeholders prefer local estimates at fine spatial resolution (Chirici et al. 2020). Multi source-national forest inventory (MS-NFI) opens the possibility for wall-to-wall mapping of forest attributes with good precision at high spatial resolution. MS-NFI rely on the combination of NFI data with auxiliary data (remote sensing data, thematic map, etc.), and in many cases, this combination is modelled through a non-parametric k-nearest neighbour (k-NN) approach. k-NN is capable in predicting several attributes in a single model with a low prediction bias. The major drawbacks of k-NN are its inability to predict beyond the range of training data (Magnussen et al. 2010), the lack of well-established variance estimator (McRoberts et al. 2011) and its decreasing performance with increasing dimensionality. The estimation maps for the forest resources are important (Tomppo et al. 2008; Chirici et al., 2020), but their prediction uncertainties have also to be taken into consideration. Methods have been proposed recently to map the prediction uncertainty (Esteban et al, 2019) and these maps have been included into an inferential framework (Saarela et al, 2020). In this study we propose a method building upon bootstrap model-based estimator (McRoberts et al. 2011) to estimate forest attributes of interest at pixel level and address the problem of extrapolation and precision of estimation by providing maps for both at high spatial resolution. For sake of concision, results were presented for growing stock volume (GSV) only. Numéro de notice : C2021-031 Affiliation des auteurs : LIF+Ext (2020- ) Thématique : FORET Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.34726/wim.1986 Date de publication en ligne : 01/12/2021 En ligne : https://doi.org/10.34726/wim.1986 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98995 Documents numériques
peut être téléchargé
High resolution mapping ... - diaporama - pdf auteurAdobe Acrobat PDF
