Centre de documentation
scientifique

See the forest and the trees: Effective machine and deep learning algorithms for wood filtering and tree species classification from terrestrial laser scanning / Zhouxin Xi in ISPRS Journal of photogrammetry and remote sensing, vol 168 (October 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 168 (October 2020) . - pp 1 - 16 Titre : See the forest and the trees: Effective machine and deep learning algorithms for wood filtering and tree species classification from terrestrial laser scanning Type de document : Article/Communication Auteurs : Zhouxin Xi, Auteur ; Christopher Hopkinson, Auteur ; Stewart B. Rood, Auteur ; Derek R. Peddle, Auteur Année de publication : 2020 Article en page(s) : pp 1 - 16 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] analyse comparative [Termes IGN] apprentissage automatique [Termes IGN] apprentissage profond [Termes IGN] classification [Termes IGN] données lidar [Termes IGN] données localisées 3D [Termes IGN] espèce végétale [Termes IGN] gestion forestière [Termes IGN] semis de points [Termes IGN] variation saisonnière Résumé : (auteur) Determining tree species composition in natural forests is essential for effective forest management. Species classification at the individual tree level requires fine-scale traits which can be derived through terrestrial laser scanning (TLS) point clouds. A generalizable species classification framework also needs to decouple seasonal foliage variation from deciduous species, for which wood filtering is applicable. Different machine learning and deep learning models are feasible for wood filtering and species classification. We investigated 13 machine learning and deep learning classifiers for 9 species, and 15 classifiers for filtering wood points from TLS plot scans. Each classifier was evaluated using the criteria of mean Intersection over Union accuracy (mIoU), training stability and time cost. On average, deep learning classifiers outperformed machine learning classifiers by 10% and 5% in terms of wood and species classification mIoU, respectively. PointNet++ provided the best species classifier, with the highest mIoU (0.906), stability, and moderate time cost. Among wood classifiers, UNet achieved the top mIoU (0.839) while ResNet-50 was recommended for rapid trial and error testing. Across the classifications, the factors of input resolution, attributes and features were also analyzed. Hot zones of species classification with PointNet++ were visualized to indicate how AI interpret species traits. Numéro de notice : A2020-533 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2020.08.001 Date de publication en ligne : 10/08/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2020.08.001 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=95718 [article]

Exemplaires(3)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
081-2020101	RAB	Revue	Centre de documentation	En réserve L003	Disponible
081-2020103	DEP-RECP	Revue	LASTIG	Dépôt en unité	Exclu du prêt
081-2020102	DEP-RECF	Revue	Nancy	Dépôt en unité	Exclu du prêt

ICESat/GLAS canopy height sensitivity inferred from Airborne Lidar / Craig Mahoney in Photogrammetric Engineering & Remote Sensing, PERS, vol 82 n° 5 (May 2016)  

Public [article]  inPhotogrammetric Engineering & Remote Sensing, PERS > vol 82 n° 5 (May 2016) . - pp 351 - 363 Titre : ICESat/GLAS canopy height sensitivity inferred from Airborne Lidar Type de document : Article/Communication Auteurs : Craig Mahoney, Auteur ; Christopher Hopkinson, Auteur ; Alex Held, Auteur ; Natascha Kljun, Auteur ; et al., Auteur Année de publication : 2016 Article en page(s) : pp 351 - 363 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] altimétrie satellitaire par laser [Termes IGN] analyse comparative [Termes IGN] biomasse forestière [Termes IGN] données ICEsat [Termes IGN] données lidar [Termes IGN] hauteur des arbres [Termes IGN] semis de points [Termes IGN] télémétrie laser aéroporté Résumé : (Auteur) Variations in laser properties and data acquisition times introduced inconsistencies in Geoscience Laser Altimeter System (GLAS) data. The effect of data inconsistencies, on two GLAS height retrieval methods, from three study sites, are investigated and validated against airborne laser scanning (ALS) percentile heights, from three data sources: all/first return point clouds, and raster canopy height models. GLAS/ALS controls were established as a basis against which the influence of laser number, transmission energy, and seasonality were assessed through comparison statistics. The favored GLAS height method best compared with ALS 95th percentile heights from an all return point cloud. Optimal GLAS data (R2 = 0.69, RMSE = 8.10 m) were noted when GLAS acquired data during summertime from high energy, laser three transmissions. As GLAS data can be used in global biomass assessments, there is a need to understand and quantify the influence of these data inconsistencies on canopy height estimates. Numéro de notice : A2016-410 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.14358/PERS.82.5.351 En ligne : http://dx.doi.org/10.14358/PERS.82.5.351 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=81276 [article]

Multisensor and multispectral Lidar characterization and classification of a forest environment / Christopher Hopkinson in Canadian journal of remote sensing, vol 42 n° 5 ([01/05/2016])  

Public [article]  inCanadian journal of remote sensing > vol 42 n° 5 [01/05/2016] . - pp 501 - 520 Titre : Multisensor and multispectral Lidar characterization and classification of a forest environment Type de document : Article/Communication Auteurs : Christopher Hopkinson, Auteur ; Laura Chasmer, Auteur ; Chris Gynan, Auteur ; Craig Mahoney, Auteur ; Michael Sitar, Auteur Année de publication : 2016 Article en page(s) : pp 501 - 520 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] canopée [Termes IGN] données lidar [Termes IGN] données localisées 3D [Termes IGN] erreur systématique [Termes IGN] feuille (végétation) [Termes IGN] modèle numérique de surface [Termes IGN] modèle numérique de surface de la canopée [Termes IGN] rayonnement proche infrarouge Résumé : (auteur) Airborne LiDAR is increasingly used in forest carbon, ecosystem, and resource monitoring. For practical design and manufacture reasons, the 1064 nm near-infrared (NIR) wavelength has been the most commonly adopted, and most literature in this field represents sampling characteristics in this wavelength. However, due to eye-safety and application-specific needs, other common wavelengths are 1550 nm and 532 nm. All provide canopy structure reconstructions that can be integrated or compared through space and time but the consistency or complementarity of 3D airborne LiDAR data sampled at multiple wavelengths is poorly understood. Here, we report on multispectral LiDAR missions carried out in 2013 and 2015 over a managed forest research site. The 1st used 3 independent sensors, and the 2nd used a single sensor carrying 3 lasers. The experiment revealed differences in proportions of returns at ground level, vertical foliage distributions, and gap probability across wavelengths. Canopy attenuation was greatest at 532 nm, presumably due to leaf tissue absorption. Relative to 1064 nm, foliage was undersampled at midheight percentiles at 1550 nm and 532 nm. Multisensor data demonstrated differences in foliage characterization due to combined influences of wavelength and acquisition configuration. Single-sensor multispectral data were more stable but demonstrated clear wavelength-dependent variation that could be exploited in intensity-based land cover classification without the aid of 3D derivatives. This work sets the stage for improvements in land surface classification and vertical foliage partitioning through the integration of active spectral and structural laser return information. Numéro de notice : A2016--128 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article DOI : 10.1080/07038992.2016.1196584 En ligne : http://dx.doi.org/10.1080/07038992.2016.1196584 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=85114 [article]

Moving toward consistent ALS monitoring of forest attributes across Canada: A consortium approach / Christopher Hopkinson in Photogrammetric Engineering & Remote Sensing, PERS, vol 79 n° 2 (February 2013)  

Public [article]  inPhotogrammetric Engineering & Remote Sensing, PERS > vol 79 n° 2 (February 2013) . - pp 159 - 173 Titre : Moving toward consistent ALS monitoring of forest attributes across Canada: A consortium approach Type de document : Article/Communication Auteurs : Christopher Hopkinson, Auteur ; Laura Chasmer, Auteur ; D. Colville, Auteur ; et al., Auteur Année de publication : 2013 Article en page(s) : pp 159 - 173 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] balayage laser [Termes IGN] Canada [Termes IGN] cohérence des données [Termes IGN] données lidar [Termes IGN] inventaire forestier (techniques et méthodes) [Termes IGN] qualité des données [Termes IGN] surveillance forestière [Termes IGN] télémétrie laser aéroporté [Termes IGN] traitement de données localisées Résumé : (Auteur) As airborne laser scanning (ALS) gains wider adoption to support forest operations in Canada, the consistency and quality of derivative products that support long-term monitoring and planning are becoming a key issues for managers. The Canadian Consortium for Lidar Environmental Applications Research (C-CLEAR) has supported almost 200 projects across Canada since 2000, with forest-related studies being a dominant theme. In 2010 and 2011, field operations were mobilized to support 13 ALS projects spanning almost the full longitudinal gradient of Canada's forests. This paper presents case studies for seven plus an overview of some best practices and data processing workflow tools that have resulted from these consortium activities. Although the projects and research teams are spread across Canada, the coordination and decade of experience provided through C-CLEAR have brought common methodological elements to all. It is clear that operational, analytical and reporting guidelines that adhere to community accepted standards are required if the benefits promised by ALS forestry are to be realized. A national Lidar Institute that builds upon the C-CLEAR model and focuses on developing standards, guidelines, and certified training would address this need. Numéro de notice : A2013-077 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article DOI : 10.14358/PERS.79.2.159 En ligne : https://doi.org/10.14358/PERS.79.2.159 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=32215 [article]

The Forward Propagation of Integrated System Component Errors within Airborne Lidar Data / T. Goulden in Photogrammetric Engineering & Remote Sensing, PERS, vol 76 n° 5 (May 2010)  

Public [article]  inPhotogrammetric Engineering & Remote Sensing, PERS > vol 76 n° 5 (May 2010) . - pp 589 - 601 Titre : The Forward Propagation of Integrated System Component Errors within Airborne Lidar Data Type de document : Article/Communication Auteurs : T. Goulden, Auteur ; Christopher Hopkinson, Auteur Année de publication : 2010 Article en page(s) : pp 589 - 601 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] données lidar [Termes IGN] erreur en position [Termes IGN] géoréférencement direct [Termes IGN] GPS-INS [Termes IGN] propagation d'erreur [Termes IGN] variance Résumé : (Auteur) Error estimates of lidar observations are obtained by applying the General Law of Propagation of Variances (GLOPOV) to the direct georeferencing equation. Within the formulation of variance propagation, the most important consideration is the values used to describe the error of the hardware component observations including the global positioning system, inertial measurement unit, laser ranger, and laser scanner (angular encoder noise and beam divergence). Data tested yielded in general, pessimistic predictions as 85 percent of residuals were within the predicted error level. Simulated errors for varying scan angles and altitudes produced horizontal errors largely influenced by IMU subsystem error as well as angular encoder noise and beam divergence. GPS subsystem errors contribute the largest proportion of vertical error only at shallow scan angles and low altitudes. The transformation of the domination of GPS related error sources to total vertical error occurs at scan angles of 23°, 13°, and 8° at flying heights of 1,200 m, 2,000 m, and 3,000 m AGL, respectively. Copyright ASPRS Numéro de notice : A2010-162 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.14358/PERS.76.5.589 En ligne : https://doi.org/10.14358/PERS.76.5.589 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=30357 [article]

Examining the influence of changing laser pulse repetition frequencies on conifer forest canopy returns / Laura Chasmer in Photogrammetric Engineering & Remote Sensing, PERS, vol 72 n° 12 (December 2006)  

Permalink