An assessment of the effectiveness of decision tree methods for land cover classification

Pal, Mahesh; Mather, Paul M.

doi:10.1016/s0034-4257(03)00132-9

Cited by 937 publications

(498 citation statements)

References 17 publications

Supporting

Mentioning

438

Contrasting

Unclassified

Order By: Relevance

“…RF was selected for this study because it generally outperforms conventional classifiers such as the Gaussian maximum likelihood classifier [61,62], while performing favorably, or equally well, to other non-parametric approaches; e.g., CART [63,64], Support Vector Machines [32,65,66], Artificial Neural Networks [67], and K-Nearest Neighbor [68]. It is a powerful non-linear and non-parametric classifier that allows for fusion and aggregation of high-dimensional data from various sources (e.g., optical, SAR, and topography [30,69,70]; SAR and topography [21,58,71]; and optical and topography [72][73][74]).…”

Section: Image Classificationmentioning

confidence: 99%

Mapping the Dabus Wetlands, Ethiopia, Using Random Forest Classification of Landsat, PALSAR and Topographic Data

et al. 2017

View full text Add to dashboard Cite

Abstract:The Dabus Wetland complex in the highlands of Ethiopia is within the headwaters of the Nile Basin and is home to significant ecological communities and rare or endangered species. Its many interrelated wetland types undergo seasonal and longer-term changes due to weather and climate variations as well as anthropogenic land use such as grazing and burning. Mapping and monitoring of these wetlands has not been previously undertaken due primarily to their relative isolation and lack of resources. This study investigated the potential of remote sensing based classification for mapping the primary vegetation groups in the Dabus Wetlands using a combination of dry and wet season data, including optical (Landsat spectral bands and derived vegetation and wetness indices), radar (ALOS PALSAR L-band backscatter), and elevation (SRTM derived DEM and other terrain metrics) as inputs to the non-parametric Random Forest (RF) classifier. Eight wetland types and three terrestrial/upland classes were mapped using field samples of observed plant community composition and structure groupings as reference information. Various tests to compare results using different RF input parameters and data types were conducted. A combination of multispectral optical, radar and topographic variables provided the best overall classification accuracy, 94.4% and 92.9% for the dry and wet season, respectively. Spectral and topographic data (radar data excluded) performed nearly as well, while accuracies using only radar and topographic data were 82-89%. Relatively homogeneous classes such as Papyrus Swamps, Forested Wetland, and Wet Meadow yielded the highest accuracies while spatially complex classes such as Emergent Marsh were more difficult to accurately classify. The methods and results presented in this paper can serve as a basis for development of long-term mapping and monitoring of these and other non-forested wetlands in Ethiopia and other similar environmental settings.

show abstract

Section: Image Classificationmentioning

confidence: 99%

Mapping the Dabus Wetlands, Ethiopia, Using Random Forest Classification of Landsat, PALSAR and Topographic Data

et al. 2017

View full text Add to dashboard Cite

show abstract

“…The method adopted for land-cover classification was the supervised algorithm Decision Tree [42], using the software ENVI 4.5 [43]. The Digital Elevation Model from Shuttle Radar Topographic Mission-SRTM (1 : 250.000 scale) was employed for deriving slope and for height control as an input dataset in the decision tree.…”

Section: (C) Image Classificationmentioning

confidence: 99%

Understanding patterns of land-cover change in the Brazilian Cerrado from 2000 to 2015

Espírito-Santo

Leite

Silva

et al. 2016

Phil. Trans. R. Soc. B

View full text Add to dashboard Cite

Clearing tropical vegetation impacts biodiversity, the provision of ecosystem services, and thus ultimately human welfare. We quantified changes in land cover from 2000 to 2015 across the Cerrado biome of northern Minas Gerais state, Brazil. We assessed the potential biophysical and socio-economic drivers of the loss of Cerrado, natural regeneration and net cover change at the municipality level. Further, we evaluated correlations between these land change variables and indicators of human welfare. We detected extensive land-cover changes in the study area, with the conversion of 23 446 km 2 and the natural regeneration of 13 926 km 2 , resulting in a net loss of 9520 km 2 . The annual net loss (−1.2% per year) of the cover of Cerrado is higher than that reported for the whole biome in similar periods. We argue that environmental and economic variables interact to underpin rates of conversion of Cerrado, most severely affecting more humid Cerrado lowlands. While rates of Cerrado regeneration are important for conservation strategies of the remaining biome, their integrity must be investigated given the likelihood of encroachment. Given the high frequency of land abandonment in tropical regions, secondary vegetation is fundamental to maintain biodiversity and ecosystem services. Finally, the impacts of Cerrado conversion on human welfare likely vary from local to regional scales, making it difficult to elaborate land-use policies based solely on socio-economic indicators. This article is part of the themed issue ‘Tropical grassy biomes: linking ecology, human use and conservation’.

show abstract

“…Researchers have proposed and experimented with many LCLU classification methods in recent years [25][26][27]. In comparison with various novel classifiers, the traditional maximum likelihood classifier has generally been used because of its ease in application, simple operation and good performance [28].…”

Section: Lclu Classification Methodsmentioning

confidence: 99%

Assessment of Land-Cover/Land-Use Change and Landscape Patterns in the Two National Nature Reserves of Ebinur Lake Watershed, Xinjiang, China

2017

View full text Add to dashboard Cite

Abstract:Land-cover and land-use change (LCLUC) alters landscape patterns and affects regional ecosystems. The objective of this study was to examine LCLUC and landscape patterns in Ebinur Lake Wetland National Nature Reserve (ELWNNR) and Ganjia Lake Haloxylon Forest National Nature Reserve (GLHFNNR), two biodiversity-rich national nature reserves in the Ebinur Lake Watershed (ELW), Xinjiang, China. Landsat satellite images from 1972, 1998, 2007 and 2013 were used to calculate the dynamics of a land-cover and land-use (LCLU) transition matrix and landscape pattern index using ENVI 5.1 and FRAGSTATS 3.3. The results showed drastic land use modifications have occurred in ELWNNR during the past four decades. Between 1972Between and 1998Between , 1998Between and 2007Between , and 2007Between and 2013.50 km 2 (7.93%), 122.70 km 2 (3.87%), and 195.40 km 2 (6.16%) of wetland were turned into salinized land. In GLHFNNR both low and medium density Haloxylon forest area declined while high density Haloxylon forest area increased. This contribution presents a method for characterizing LCLUC using one or more cross-tabulation matrices based on Sankey diagrams, demonstrating the depiction of flows of energy or materials through ecosystem network. The ecological landscape index displayed that a unique landscape patches have shrunk in size, scattered, and fragmented. It becomes a more diverse landscape. Human activities like farming were negatively correlated with the landscape diversity of wetlands. Furthermore, evidence of degraded wetlands caused by air temperature and annual precipitation, was also observed. We conclude that national and regional policies related to agriculture and water use have significantly contributed to the extensive changes; the ELWNNR and GLHFNNR are highly susceptible to LCLUC in the surrounding Ebinur Lake Watershed.

show abstract

An assessment of the effectiveness of decision tree methods for land cover classification

Cited by 937 publications

References 17 publications

Mapping the Dabus Wetlands, Ethiopia, Using Random Forest Classification of Landsat, PALSAR and Topographic Data

Mapping the Dabus Wetlands, Ethiopia, Using Random Forest Classification of Landsat, PALSAR and Topographic Data

Understanding patterns of land-cover change in the Brazilian Cerrado from 2000 to 2015

Assessment of Land-Cover/Land-Use Change and Landscape Patterns in the Two National Nature Reserves of Ebinur Lake Watershed, Xinjiang, China

Contact Info

Product

Resources

About