Détail de l'auteur
Auteur Marco Heurich |
Documents disponibles écrits par cet auteur (8)
Ajouter le résultat dans votre panier Affiner la recherche Interroger des sources externesInstance segmentation of standing dead trees in dense forest from aerial imagery using deep learning / Aboubakar Sani-Mohammed in ISPRS Open Journal of Photogrammetry and Remote Sensing, vol 6 (December 2022)
Titre : Instance segmentation of standing dead trees in dense forest from aerial imagery using deep learning Type de document : Article/Communication Auteurs : Aboubakar Sani-Mohammed, Auteur ; Wei Yao, Auteur ; Marco Heurich, Auteur Année de publication : 2022 Article en page(s) : n° 100024 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] arbre mort
[Termes IGN] Bavière (Allemagne)
[Termes IGN] bois sur pied
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] détection automatique
[Termes IGN] gestion forestière durable
[Termes IGN] image à haute résolution
[Termes IGN] image aérienne
[Termes IGN] image infrarouge couleur
[Termes IGN] peuplement mélangé
[Termes IGN] puits de carbone
[Termes IGN] segmentation sémantiqueRésumé : (auteur) Mapping standing dead trees, especially, in natural forests is very important for evaluation of the forest's health status, and its capability for storing Carbon, and the conservation of biodiversity. Apparently, natural forests have larger areas which renders the classical field surveying method very challenging, time-consuming, labor-intensive, and unsustainable. Thus, for effective forest management, there is the need for an automated approach that would be cost-effective. With the advent of Machine Learning, Deep Learning has proven to successfully achieve excellent results. This study presents an adjusted Mask R-CNN Deep Learning approach for detecting and segmenting standing dead trees in a mixed dense forest from CIR aerial imagery using a limited (195 images) training dataset. First, transfer learning is considered coupled with the image augmentation technique to leverage the limitation of training datasets. Then, we strategically selected hyperparameters to suit appropriately our model's architecture that fits well with our type of data (dead trees in images). Finally, to assess the generalization capability of our model's performance, a test dataset that was not confronted to the deep neural network was used for comprehensive evaluation. Our model recorded promising results reaching a mean average precision, average recall, and average F1-Score of 0.85, 0.88, and 0.87 respectively, despite our relatively low resolution (20 cm) dataset. Consequently, our model could be used for automation in standing dead tree detection and segmentation for enhanced forest management. This is equally significant for biodiversity conservation, and forest Carbon storage estimation. Numéro de notice : A2022-871 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article DOI : 10.1016/j.ophoto.2022.100024 Date de publication en ligne : 10/11/2022 En ligne : https://doi.org/10.1016/j.ophoto.2022.100024 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=102165
in ISPRS Open Journal of Photogrammetry and Remote Sensing > vol 6 (December 2022) . - n° 100024[article]
Titre : Adaptive stopping criterion for top-down segmentation of ALS point clouds in temperate coniferous forests Type de document : Article/Communication Auteurs : Nina Amiri, Auteur ; Przemyslaw Polewski, Auteur ; Marco Heurich, Auteur ; Peter Krzystek, Auteur ; Andrew K. Skidmore, Auteur Année de publication : 2018 Article en page(s) : pp 265 - 274 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] Bavière (Allemagne)
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] inventaire forestier (techniques et méthodes)
[Termes IGN] inventaire forestier local
[Termes IGN] lasergrammétrie
[Termes IGN] Pinophyta
[Termes IGN] segmentation
[Termes IGN] semis de points
[Vedettes matières IGN] Inventaire forestierMots-clés libres : Bavarian Forest National Park Résumé : (auteur) The development of new approaches to individual tree crown delineation for forest inventory and management is an important area of ongoing research. The increasing availability of high density ALS (Airborne Laser Scanning) point clouds offers the opportunity to segment the individual tree crowns and deduce their geometric properties with a high level of accuracy. Top-down segmentation methods such as normalized cut are established approaches for delineation of single trees in ALS point clouds. However, overlapping crowns and branches of nearby trees frequently cause over- and under-segmentation due to the difficulty of defining a single criterion for stopping the partitioning process. In this work, we investigate an adaptive stopping criterion based on the visual appearance of trees within the point clouds. We focus on coniferous trees due to their well-defined crown shapes in comparison to deciduous trees. This approach is based on modeling the coniferous tree crowns with elliptic paraboloids to infer whether a given 3D scene contains exactly one or more than one tree. For each processed scene, candidate tree peaks are generated from local maxima found within the point cloud. Next, paraboloids are fitted at the peaks using a random sample consensus procedure and classified based on their geometric properties. The decision to stop or continue partitioning is determined by finding a set of non-overlapping paraboloids. Experiments were performed on three plots from the Bavarian Forest National Park in Germany. Based on validation data from the field inventory, results show that our approach improves the segmentation quality by up to 10% across plots with different properties, such as average tree height and density. This indicates that the new adaptive stopping criterion for normalized cut segmentation is capable of delineating tree crowns more accurately than a static stopping criterion based on a constant Ncut threshold value. Numéro de notice : A2018-670 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2018.05.006 Date de publication en ligne : 29/05/2018 En ligne : https://doi.org/10.1016/j.isprsjprs.2018.05.006 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=90405
in ISPRS Journal of photogrammetry and remote sensing > vol 141 (July 2018) . - pp 265 - 274[article]Exemplaires(3)
Code-barres Cote Support Localisation Section Disponibilité 081-2018071 RAB Revue Centre de documentation En réserve L003 Disponible 081-2018073 DEP-EXM Revue LASTIG Dépôt en unité Exclu du prêt 081-2018072 DEP-EAF Revue Nancy Dépôt en unité Exclu du prêt
Titre : Important LiDAR metrics for discriminating forest tree species in Central Europe Type de document : Article/Communication Auteurs : Yifang Shi, Auteur ; Tiejun Wang, Auteur ; Andrew K. Skidmore, Auteur ; Marco Heurich, Auteur Année de publication : 2018 Article en page(s) : pp 163 - 174 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie
[Termes IGN] Allemagne
[Termes IGN] arbre (flore)
[Termes IGN] classification par forêts d'arbres décisionnels
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] Europe centrale
[Termes IGN] forêt tempérée
[Termes IGN] morphologie mathématiqueRésumé : (Auteur) Numerous airborne LiDAR-derived metrics have been proposed for classifying tree species. Yet an in-depth ecological and biological understanding of the significance of these metrics for tree species mapping remains largely unexplored. In this paper, we evaluated the performance of 37 frequently used LiDAR metrics derived under leaf-on and leaf-off conditions, respectively, for discriminating six different tree species in a natural forest in Germany. We firstly assessed the correlation between these metrics. Then we applied a Random Forest algorithm to classify the tree species and evaluated the importance of the LiDAR metrics. Finally, we identified the most important LiDAR metrics and tested their robustness and transferability. Our results indicated that about 60% of LiDAR metrics were highly correlated to each other (|r| > 0.7). There was no statistically significant difference in tree species mapping accuracy between the use of leaf-on and leaf-off LiDAR metrics. However, combining leaf-on and leaf-off LiDAR metrics significantly increased the overall accuracy from 58.2% (leaf-on) and 62.0% (leaf-off) to 66.5% as well as the kappa coefficient from 0.47 (leaf-on) and 0.51 (leaf-off) to 0.58. Radiometric features, especially intensity related metrics, provided more consistent and significant contributions than geometric features for tree species discrimination. Specifically, the mean intensity of first-or-single returns as well as the mean value of echo width were identified as the most robust LiDAR metrics for tree species discrimination. These results indicate that metrics derived from airborne LiDAR data, especially radiometric metrics, can aid in discriminating tree species in a mixed temperate forest, and represent candidate metrics for tree species classification and monitoring in Central Europe. Numéro de notice : A2018-080 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2018.02.002 Date de publication en ligne : 07/02/2018 En ligne : https://doi.org/10.1016/j.isprsjprs.2018.02.002 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=89442
in ISPRS Journal of photogrammetry and remote sensing > vol 137 (March 2018) . - pp 163 - 174[article]Exemplaires(3)
Code-barres Cote Support Localisation Section Disponibilité 081-2018031 RAB Revue Centre de documentation En réserve L003 Disponible 081-2018033 DEP-EXM Revue LASTIG Dépôt en unité Exclu du prêt 081-2018032 DEP-EAF Revue Nancy Dépôt en unité Exclu du prêt
Titre : Large off-nadir scan angle of airborne LiDAR can severely affect the estimates of forest structure metrics Type de document : Article/Communication Auteurs : Jing Liu, Auteur ; Andrew K. Skidmore, Auteur ; Simon D. Jones, Auteur ; Tiejun Wang, Auteur ; Marco Heurich, Auteur ; Xi Zhu, Auteur ; Yifang Shi, Auteur Année de publication : 2018 Article en page(s) : pp 13 - 25 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie
[Termes IGN] angle de visée
[Termes IGN] Bavière (Allemagne)
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] instrument aéroporté
[Termes IGN] parc naturel régional
[Termes IGN] placette d'échantillonnage
[Termes IGN] structure d'un peuplement forestierRésumé : (Auteur) Gap fraction (Pgap) and vertical gap fraction profile (vertical Pgap profile) are important forest structural metrics. Accurate estimation of Pgap and vertical Pgap profile is therefore critical for many ecological applications, including leaf area index (LAI) mapping, LAI profile estimation and wildlife habitat modelling. Although many studies estimated Pgap and vertical Pgap profile from airborne LiDAR data, the scan angle was often overlooked and a nadir view assumed. However, the scan angle can be off-nadir and highly variable in the same flight strip or across different flight strips. In this research, the impact of off-nadir scan angle on Pgap and vertical Pgap profile was evaluated, for several forest types. Airborne LiDAR data from nadir (0°∼7°), small off-nadir (7°∼23°), and large off-nadir (23°∼38°) directions were used to calculate both Pgap and vertical Pgap profile. Digital hemispherical photographs (DHP) acquired during fieldwork were used as references for validation. Our results show that angular Pgap from airborne LiDAR correlates well with angular Pgap from DHP (R2 = 0.74, 0.87, and 0.67 for nadir, small off-nadir and large off-nadir direction). But underestimation of Pgap from LiDAR amplifies at large off-nadir scan angle. By comparing Pgap and vertical Pgap profiles retrieved from different directions, it is shown that scan angle impact on Pgap and vertical Pgap profile differs amongst different forest types. The difference is likely to be caused by different leaf angle distribution and canopy architecture in these forest types. Statistical results demonstrate that the scan angle impact is more severe for plots with discontinuous or sparse canopies. These include coniferous plots, and deciduous or mixed plots with between-crown gaps. In these discontinuous plots, Pgap and vertical Pgap profiles are maximum when observed from nadir direction, and then rapidly decrease with increasing scan angle. The results of this research have many important practical implications. First, it is suggested that large off-nadir scan angle of airborne LiDAR should be avoided to ensure a more accurate Pgap and LAI estimation. Second, the angular dependence of vertical Pgap profiles observed from airborne LiDAR should be accounted for, in order to improve the retrieval of LAI profiles, and other quantitative canopy structural metrics. This is especially necessary when using multi-temporal datasets in discontinuous forest types. Third, the anisotropy of Pgap and vertical Pgap profile observed by airborne LiDAR, can potentially help to resolve the anisotropic behavior of canopy reflectance, and refine the inversion of biophysical and biochemical properties from passive multispectral or hyperspectral data Numéro de notice : A2018-072 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2017.12.004 En ligne : https://doi.org/10.1016/j.isprsjprs.2017.12.004 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=89432
in ISPRS Journal of photogrammetry and remote sensing > vol 136 (February 2018) . - pp 13 - 25[article]Exemplaires(3)
Code-barres Cote Support Localisation Section Disponibilité 081-2018021 RAB Revue Centre de documentation En réserve L003 Disponible 081-2018023 RAB Revue Centre de documentation En réserve L003 Disponible 081-2018022 DEP-EAF Revue Nancy Dépôt en unité Exclu du prêt
Titre : Multi-model estimation of understorey shrub, herb and moss cover in temperate forest stands by laser scanner data Type de document : Article/Communication Auteurs : Hooman Latifi, Auteur ; Steven Hill, Auteur ; Bastian Schumann, Auteur ; Marco Heurich, Auteur ; Stefan Dech, Auteur Année de publication : 2017 Article en page(s) : pp 496 - 514 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie
[Termes IGN] classification par forêts d'arbres décisionnels
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] estimation statistique
[Termes IGN] forêt tempérée
[Termes IGN] habitat forestier
[Termes IGN] sous-boisRésumé : (Auteur) In temperate forests, the highest plant richness is regularly found in the understorey, i.e. shrub, tree regeneration, herbal and moss covers, which provides important food and shelter for other plant and animal species. Here, Light Detection And Ranging (LiDAR) remote sensing was investigated as a surrogate to laborious field surveys to improve understanding of the causal and predictive attributes of understorey. We designed a study in which we used a high-density LiDAR point cloud and applied a thinning algorithm to simulate two lower density point clouds including first and last returns and half of the remaining points (half-thinned data) and only first and last returns (F/L-thinned data). From each dataset, several over- and understorey-related statistical metrics were derived. Each of the three sets of LiDAR metrics was then combined with the forest habitat information to estimate the recorded proportions of shrub, herb and moss coverages. We used three different model procedures including zero-and-one-inflated beta regression (ZOINBR), ordinary least squares with logit-transformed response variables (logistic model) and a machine learning random forest (RF) method. The logistic and ZOINBR model results showed highly significant relationships between LiDAR metrics and habitat types in explaining understorey coverage. The highest coefficients of determination included r2 = 0.80 for shrub cover (estimated by F/L-thinned data and ZOINBR model), r2 = 0.53 for herb cover (estimated by half-thinned data and logistic model) and r2 = 0.48 for moss cover (estimated by half-thinned data and logistic model). RF models returned the best predictive performances (i.e. the lowest root mean square errors). Despite slight differences, no substantial difference was observed amongst the performances achieved by the original, half-thinned and F/L-thinned point clouds. Moreover, the ZOINBR models did not improve predictive performances compared with the logistic model, which suggests that the latter should be preferred due to its greater simplicity and parsimony. Despite the differences between our simulated data and the real-world LiDAR point clouds of different point densities, the results of this study are thought to mostly reflect how LiDAR and forest habitat data can be combined for deriving ecologically relevant information on temperate forest understorey vegetation layers. This, in turn, increases the applicability of prediction results for overarching aims such as forest and wildlife management. Numéro de notice : A2017-906 Affiliation des auteurs : non IGN Thématique : FORET/MATHEMATIQUE Nature : Article DOI : 10.1093/forestry/cpw066 Date de publication en ligne : 27/01/2017 En ligne : https://doi.org/10.1093/forestry/cpw066 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=93195
in Forestry, an international journal of forest research > vol 90 n° 4 (October 2017) . - pp 496 - 514[article]
