Centre de documentation
scientifique

Mask R-CNN and OBIA fusion improves the segmentation of scattered vegetation in very high-resolution optical sensors / Emilio Guirado in Sensors, vol 21 n° 1 (January 2021)  

Public [article]  inSensors > vol 21 n° 1 (January 2021) . - n° 320 Titre : Mask R-CNN and OBIA fusion improves the segmentation of scattered vegetation in very high-resolution optical sensors Type de document : Article/Communication Auteurs : Emilio Guirado, Auteur ; Javier Blanco-Sacristán, Auteur ; Emilio Rodríguez-Caballero, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : n° 320 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] analyse d'image orientée objet [Termes IGN] apprentissage profond [Termes IGN] arbuste [Termes IGN] capteur optique [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] détection de changement [Termes IGN] image à très haute résolution [Termes IGN] segmentation d'image [Termes IGN] service écosystémique [Termes IGN] surveillance de la végétation [Termes IGN] zone aride Résumé : (auteur) Vegetation generally appears scattered in drylands. Its structure, composition and spatial patterns are key controls of biotic interactions, water, and nutrient cycles. Applying segmentation methods to very high-resolution images for monitoring changes in vegetation cover can provide relevant information for dryland conservation ecology. For this reason, improving segmentation methods and understanding the effect of spatial resolution on segmentation results is key to improve dryland vegetation monitoring. We explored and analyzed the accuracy of Object-Based Image Analysis (OBIA) and Mask Region-based Convolutional Neural Networks (Mask R-CNN) and the fusion of both methods in the segmentation of scattered vegetation in a dryland ecosystem. As a case study, we mapped Ziziphus lotus, the dominant shrub of a habitat of conservation priority in one of the driest areas of Europe. Our results show for the first time that the fusion of the results from OBIA and Mask R-CNN increases the accuracy of the segmentation of scattered shrubs up to 25% compared to both methods separately. Hence, by fusing OBIA and Mask R-CNNs on very high-resolution images, the improved segmentation accuracy of vegetation mapping would lead to more precise and sensitive monitoring of changes in biodiversity and ecosystem services in drylands. Numéro de notice : A2021-157 Affiliation des auteurs : non IGN Thématique : BIODIVERSITE/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.3390/s21010320 Date de publication en ligne : 05/01/2021 En ligne : https://doi.org/10.3390/s21010320 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97072 [article]

Transferability of multi- and hyperspectral optical biocrust indices / Emilio Rodríguez-Caballero in ISPRS Journal of photogrammetry and remote sensing, vol 126 (April 2017)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 126 (April 2017) . - pp 94 – 107 Titre : Transferability of multi- and hyperspectral optical biocrust indices Type de document : Article/Communication Auteurs : Emilio Rodríguez-Caballero, Auteur ; P. Escribano, Auteur ; C. Olehowski, Auteur ; et al., Auteur Année de publication : 2017 Article en page(s) : pp 94 – 107 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] biomasse [Termes IGN] classification dirigée [Termes IGN] état de surface du sol [Termes IGN] image hyperspectrale [Termes IGN] indice de détection [Termes IGN] surface du sol Résumé : (auteur) Biological soil crusts (biocrusts) are communities of cyanobacteria, algae, microfungi, lichens and bryophytes in varying proportions, which live within or immediately on top of the uppermost millimeters of the soil in arid and semiarid regions. As biocrusts are highly relevant for ecosystem processes like carbon, nitrogen, and water cycling, a correct characterization of their spatial distribution is required. Following this objective, considerable efforts have been devoted to the identification and mapping of biocrusts using remote sensing data, and several mapping indices have been developed. However, their transferability to different regions has only rarely been tested. In this study we investigated the transferability of two multispectral indices, i.e. the Crust Index (CI) and the Biological Soil Crust Index (BSCI), and two hyperspectral indices, i.e. the Continuum Removal Crust Identification Algorithm (CRCIA) and the Crust Development Index (CDI), in three sites dominated by biocrusts, but with differences in soil and vegetation composition. Whereas multispectral indices have been important and valuable tools for first approaches to map and classify biological soil crusts, hyperspectral data and indices developed for these allowed to classify biocrusts at much higher accuracy. While multispectral indices showed Kappa (κ) values below 0.6, hyperspectral indices obtained good classification accuracy (κ ∼ 0.8) in both the study area where they had been developed and in the newly tested region. These results highlight the capability of hyperspectral sensors to identify specific absorption features related to photosynthetic pigments as chlorophyll and carotenoids, but also the limitation of multispectral information to discriminate between areas dominated by biocrusts, vegetation or bare soil. Based on these results we conclude that remote sensing offers an important and valid tool to map biocrusts. However, the spectral similarity between the main surface components of drylands and biocrusts demand for mapping indices based on hyperspectral information to correctly map areas dominated by biocrusts at ecosystem scale. Numéro de notice : A2017-115 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2017.02.007 En ligne : http://dx.doi.org/10.1016/j.isprsjprs.2017.02.007 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=84510 [article]

Exemplaires(3)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
081-2017041	RAB	Revue	Centre de documentation	En réserve L003	Disponible
081-2017043	DEP-EXM	Revue	LASTIG	Dépôt en unité	Exclu du prêt
081-2017042	DEP-EAF	Revue	Nancy	Dépôt en unité	Exclu du prêt

A new adaptive method to filter terrestrial laser scanner point clouds using morphological filters and spectral information to conserve surface micro-topography / Emilio Rodríguez-Caballero in ISPRS Journal of photogrammetry and remote sensing, vol 117 (July 2016)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 117 (July 2016) . - pp 141 – 148 Titre : A new adaptive method to filter terrestrial laser scanner point clouds using morphological filters and spectral information to conserve surface micro-topography Type de document : Article/Communication Auteurs : Emilio Rodríguez-Caballero, Auteur ; A. Afana, Auteur ; S. Chamizo, Auteur ; et al., Auteur Année de publication : 2016 Article en page(s) : pp 141 – 148 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] données lidar [Termes IGN] filtrage numérique d'image [Termes IGN] microtopographie [Termes IGN] modèle numérique de terrain [Termes IGN] semis de points [Termes IGN] télémétrie laser terrestre Résumé : (auteur) Terrestrial laser scanning (TLS), widely known as light detection and ranging (LiDAR) technology, is increasingly used to provide highly detailed digital terrain models (DTM) with millimetric precision and accuracy. In order to generate a DTM, TLS data has to be filtered from undesired spurious objects, such as vegetation, artificial structures, etc., Early filtering techniques, successfully applied to airborne laser scanning (ALS), fail when applied to TLS data, as they heavily smooth the terrain surface and do not retain their real morphology. In this article, we present a new methodology for filtering TLS data based on the geometric and radiometric properties of the scanned surfaces. This methodology was built on previous morphological filters that select the minimum point height within a sliding window as the real surface. However, contrary to those methods, which use a fixed window size, the new methodology operates under different spatial scales represented by different window sizes, and can be adapted to different types and sizes of plants. This methodology has been applied to two study areas of differing vegetation type and density. The accuracy of the final DTMs was improved by ∼30% under dense canopy plants and over ∼40% on the open spaces between plants, where other methodologies drastically underestimated the real surface heights. This resulted in more accurate representation of the soil surface and microtopography than up-to-date techniques, eventually having strong implications in hydrological and geomorphological studies. Numéro de notice : A2016-583 Affiliation des auteurs : non IGN Thématique : IMAGERIE/POSITIONNEMENT Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2016.04.004 En ligne : https://doi.org/10.1016/j.isprsjprs.2016.04.004 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=81724 [article]