The logic of multispectral classification and mapping of land

Robinove, Charles J.

doi:10.1016/0034-4257(81)90022-5

Cited by 45 publications

(16 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One does not have sufficient data about the state of the atmosphere when the image is recorded. This problem is further exaggerated when the comparison involves radiance values recorded at different dates and, therefore, during different atmospheric states (Escadafal, 1989;Robinove, 1982).…”

Section: Pretreatment Of Imagesmentioning

confidence: 97%

Detection of Change in Moroccan Rangelands with Multitemporal Spot Imagery

BennounaA¹,

Lefèvre-Fonollosa

Lacombe

et al. 2004

Arid Land Research and Management

View full text Add to dashboard Cite

Section: Pretreatment Of Imagesmentioning

confidence: 97%

Detection of Change in Moroccan Rangelands with Multitemporal Spot Imagery

BennounaA¹,

Lefèvre-Fonollosa

Lacombe

et al. 2004

Arid Land Research and Management

View full text Add to dashboard Cite

“…A brief review of each approach to supervised classification (Robinove, 1981) is first presented, following which these algorithms are compared with respect to statistical assumptions and the new processing requirements for modern, multi-source remote sensing and ancillary data sets. The comparison is summarized in Table 1 and discussed in more detail to illustrate some of the important advantages of the Evidential Reasoning approach to classification.…”

Section: Theoretical Comparison and Advantagesmentioning

confidence: 99%

An Empirical Comparison of Evidential Reasoning, Linear Discriminant Analysis, and Maximum Likelihood Algorithms for Alpine Land Cover Classification

Peddle

1993

Canadian Journal of Remote Sensing

View full text Add to dashboard Cite

RESUMELa mise au point de nouvelles methodes d'analyse des donnees de teledetection actuelles etfutures est necessaire si l'on veut surmonter les limites que presentent les methodes classiques de traitement et de classification des images. Dans le present article, l'auteurfait l'eualuation d'un classificateurpar raisonnement d'etndences [Evidential Reasoning] (MERCURY(f)) qui fait appel it la tbeorie des evidences de Dempster-Shafer et qui offre une souplesse accrue ainsi qu 'une nouvellefonctionnalite pour traiter des ensembles de donnees presentant des proprietes statistiques diverses, des dimensions superieures, de l'incertitude et des ecbelles de mesures differentes, facteurs qui dejouent les algorithmes classiques de classification par maximum de vraisemblance et d'analyse discriminante lineaire. Lefondement tbeorique du classificateur d'euidences est compare aux deux methodes precitees. Suit ensuite une comparaison ernpirique approfondie des trois classificateurs, fondee sur une serie de tests pour neufclasses de couverture du sol dans une region montagneuse situee dans te sud-ouest du Yukon. Douze combinaisons differentes de variables choisies parmi un ensemble de donnees de sources multiples, soit des images SPOT, des images de texture et des images de variables geomorpbometriques tirees d'un modele numerique d'eleoation, ont ete soumises it chacun des classificateurs, puis les resultats ont ete compares. La precision de la classification par maximum de vraisemblance a atteint son maximum avec l'utilisation de quatre variables et celle-ci s'est deterioree avec l'afout d'autres variables. On a observe la tendance inverse dans le cas des deux autres methodes. Sur les trois classificateurs, seulle classificateurpar raisonnement d'etndences pouvait traiter les mesures d'orientation directement, alors qu 'avec les deux autres, on a du faire appel it une transformation de l'incidence. La methode de classification des evidences a produit de meilleurs resultats de classification avec l'utilisation de l'orientation plut8t que l'incidence, ce qui donne it penser que la transformation des donnees a engendre une perte d'informations. Dans l'ensemble, le logiciel MERCURY(f) a produit les meilleurs resultats de classification dans 9 cas sur 12 et a, de plus, presente le meilleur taux de precision global, soit 91 pour cent. Les resultats de ces comparaisons illustrent, d'une part, le besoin d'innover en matiere de traitement d'images et d 'extraction de l'information et, d 'autre part, les avantages et la superiorite du classificateur d 'evidences pour la classification integree d'images de sources multiples. SUMMARY New procedures for analyzing modern andfuture remote sensing data sets are needed to address the methodological limitations ofconventional approaches to image processing and classification. Ibis paper evaluates an Evidential Reasoning classifier (MERCURY(f)) that uses the Dempster-Shafer Theory ofEvidence and provides greaterflexibility and new functionality to handle data sets with diverse statistical prop...

show abstract

“…Multispectral reflectance data from satellite sensors serve as an alternate representation of vegetation communities and can therefore, in theory, serve as a surrogate for gathering this information on the ground (Robinove 1981).…”

Section: Introductionmentioning

confidence: 99%

Using satellite imagery and LIDAR data to corroborate an adjudicated ordinary high water line

Genç

Smith

Dewitt

2005

International Journal of Remote Sensing

View full text Add to dashboard Cite

Determination of the ordinary high water line (OHWL) has been and continues to be an important issue. The OHWL defines the separation of sovereignty lands and private ownership on non-tidal water bodies. Determination of OHWL is conducted on a case-by-case basis in Florida through court challenges. A judge makes the decision on where the line exists based upon several criteria-including remote sensing data. This study investigated the possibility of using various remote sensing technologies to provide an efficient and accurate means of determining OHWL for a lake in central Florida. Landsat Enhanced Thematic Mapper (ETM) satellite imagery was compared with the higher resolution imagery IKONOS and Light Detection And Ranging (LIDAR) imagery in order to determine the water's edge and location of vegetation communities that may be correlated with OHWL. It was found that ETM imagery could be used only for mapping vegetation community transition zones and that this zone provided limited insight to OHWL. IKONOS imagery, on the other hand, was more promising for land cover mapping, but requires further study in order to draw general conclusions regarding its application to OHWL. LIDAR data provided the best results for determining OHWL, but also need further study over a larger area in order to draw final conclusions.

show abstract

The logic of multispectral classification and mapping of land

Cited by 45 publications

References 1 publication

Detection of Change in Moroccan Rangelands with Multitemporal Spot Imagery

Detection of Change in Moroccan Rangelands with Multitemporal Spot Imagery

An Empirical Comparison of Evidential Reasoning, Linear Discriminant Analysis, and Maximum Likelihood Algorithms for Alpine Land Cover Classification

Using satellite imagery and LIDAR data to corroborate an adjudicated ordinary high water line

Contact Info

Product

Resources

About