Large scale region-merging segmentation using the local mutual best fitting concept

Lassalle, Pierre; Inglada, Jordi; Michel, Julien; Grizonnet, Manuel; Malik, Julien

doi:10.1109/igarss.2014.6947590

Cited by 7 publications

(6 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We use the method described in Ref. [15] with OTB's GenericRegionMerging (Large Scale version) application [16]. To obtain a segmentation result adapted to our study, we act on both the homogeneity criteria and the maximum heterogeneity threshold.…”

Section: Experimental Design Materials and Methodsmentioning

confidence: 99%

Mapping land cover on Reunion Island in 2017 using satellite imagery and geospatial ground data

2020

View full text Add to dashboard Cite

We here present a reference database and three land use maps produced in 2017 over the Reunion island using a machine learning based methodology. These maps are the result of a satellite image analysis performed using the Moringa land cover processing chain developed in our laboratory. The input dataset for map production consists of a single very high spatial resolution Pleiades images, a time series of Sentinel-2 and Landsat-8 images, a Digital Terrain Model (DTM) and the aforementioned reference database. The Moringa chain adopts an object based approach: the Pleiades image provides spatial accuracy with the delineation of land samples via a segmentation process, the time series provides information on landscape and vegetation dynamics, the DTM provides information on topography and the reference database provides annotated samples (6256 polygons) for the supervised classification process and the validation of the results. The three land use maps follow a hierarchical nomenclature ranging from 4 classes for the least detailed level to 34 classes for the most detailed one. The validation of these maps shows a good quality of the results with overall accuracy rates ranging from 86% to 97%. The maps are freely accessible and used by researchers, land managers (State services and local authorities) and also private companies.

show abstract

Section: Experimental Design Materials and Methodsmentioning

confidence: 99%

Mapping land cover on Reunion Island in 2017 using satellite imagery and geospatial ground data

2020

View full text Add to dashboard Cite

show abstract

“…The Moringa processing chain is designed to provide object-based supervised classification, and operates by first performing the segmentation of the VHSR image to generate a suitable object layer. The method described in [13] , implemented in the large scale version of OTB's GenericRegionMerging application [14] , was used to perform the segmentation. To obtain a segmentation result adapted to our study, parameters for the homogeneity criteria and the maximum heterogeneity threshold were assessed using a grid search on several representative subsets of the VHSR pansharpened image.…”

Section: Experimental Design Materials and Methodsmentioning

confidence: 99%

“…Training samples were subsequently generated by intersecting the so-obtained segmentation with the reference polygons available in the GIS dataset, and attributed using the spatial means over every band and index listed in Table 2 . Random Forest (RF) classification algorithm [ 14 , 15 ] was chosen for classification considering its robustness when working with heterogeneous data, such as in our study (data from several sensors combined with altitude, slopes and textural indices).…”

Section: Experimental Design Materials and Methodsmentioning

confidence: 99%

Land cover maps of Antananarivo (capital of Madagascar) produced by processing multisource satellite imagery and geospatial reference data

et al. 2020

View full text Add to dashboard Cite

We describe a reference spatial database and four land use maps of Antananarivo city produced over 2017 reference year using a methodology combining machine learning and object based image analysis (OBIA). These maps are produced by processing satellite images using the Moringa land cover processing chain developed in our laboratory. We use a single very high spatial resolution (VHSR) Pleiades image, a time series of Sentinel-2 and Landsat-8 images, a Digital Terrain Model (DTM) and the aforementioned reference database. According to the Moringa workflow, the Pleiades image is used to generate a suitable object layer at VHSR using a segmentation algorithm. Each object is then classified using variables from the time series and information from the DTM. The reference database used to train the supervised classification algorithm is here described, as well as the 4 land cover maps produced at four different hierarchically nested nomenclature levels. For a number of classes going from 2 to 20, overall accuracies range from 94% to 74%. Such land cover products are very rare in Madagascar, so we have decided to make them openly accessible and usable by land managers and researchers.

show abstract

“…A outra ideia está relacionada com à utilização de tiles com sobreposição de forma a produzir áreas estáveis onde há a garantia de que, para uma dada quantidade de iterações, o crescimento de regiões será sempre igual. Esta solução remete à margem de estabilidade proposta em (Lassale et al, 2014) e requer que após uma quantidade pré-determinada de iterações, as iterações restantes do crescimento de regiões sejam realizadas uma de cada vez como processos distribuídos.…”

Section: 1trabalhos Futurosunclassified

“…Com o avanço da tecnologia de integração em altíssima escala organizações paralelas de computadores se tornaram disponíveis comercialmente a preços acessíveis. Assim, a computação paralela se tornou a principal opção para acelerar o procedimento de segmentação de imagens reduzindo consideravelmente o tempo gasto na análise de imagens baseadas em objetos (Moga et al, 2008, Montoya et al, 2003, Lenkiewicz et al 2009, Happ et al, 2013.…”

unclassified

Segmentação De Imagens Distribuída Baseada Em Mapreduce

HAPP¹

View full text Add to dashboard Cite

Large scale region-merging segmentation using the local mutual best fitting concept

Cited by 7 publications

References 1 publication

Mapping land cover on Reunion Island in 2017 using satellite imagery and geospatial ground data

Mapping land cover on Reunion Island in 2017 using satellite imagery and geospatial ground data

Land cover maps of Antananarivo (capital of Madagascar) produced by processing multisource satellite imagery and geospatial reference data

Segmentação De Imagens Distribuída Baseada Em Mapreduce

Contact Info

Product

Resources

About