Sources of error in accuracy assessment of thematic land-cover maps in the Brazilian Amazon

Powell, Rebecca; Matzke, Nicholas J.; Souza, Camilla Mariane Menezes; Clark, Matthew L.; Numata, Izaya; Hess, Laura L.; Roberts, Dar A.

doi:10.1016/j.rse.2003.12.007

Cited by 171 publications

(134 citation statements)

References 29 publications

Supporting

Mentioning

118

Contrasting

Unclassified

Order By: Relevance

“…As examples, aerial photography is commonly used because it offers a cost-effective option for covering extensive areas; videography can be used to determine the observed class or quantity (Powell et al, 2004;Wulder et al, 2007); and interpretation of very fine-resolution imagery such as IKONOS and QuickBird offers another, albeit expensive, option. Finally, airborne lidar data have been used to evaluate insect defoliation maps derived from spaceborne optical MODIS data (Eklundh et al, 2009), biomass estimated from spaceborne radar data (Solberg et al, 2010), and canopy height estimated from spaceborne lidar data (Duncanson et al, 2009).…”

Section: Response Designmentioning

confidence: 99%

Using remotely sensed data to construct and assess forest attribute maps and related spatial products

McRoberts

Cohen

Næsset

et al. 2010

Scandinavian Journal of Forest Research

110

View full text Add to dashboard Cite

Tremendous advances in the construction and assessment of forest attribute maps and related spatial products have been realized in recent years, partly as a result of the use of remotely sensed data as an information source. This review focuses on the current state of techniques for the construction and assessment of remote sensing-based maps and addresses five topic areas: statistical classification and prediction techniques used to construct maps and related spatial products, accuracy assessment methods, map-based statistical inference, and two emerging topics, change detection and use of lidar data. Multiple general conclusions were drawn from the review: (1) remotely sensed data greatly contribute to the construction of forest attribute maps and related spatial products and to the reduction of inventory costs; (2) parametric prediction techniques, accuracy assessment methods and probability-based (design-based) inferential methods are generally familiar and mature, although inference is surprisingly seldom addressed; (3) non-parametric prediction techniques and modelbased inferential methods lack maturity and merit additional research; (4) change detection methods, with their great potential for adding a spatial component to change estimates, will mature rapidly; and (5) lidar applications, although currently immature, add an entirely new dimension to remote sensing research and will also mature rapidly. Crucial forest sustainability and climate change applications will continue to push all aspects of remote sensing to the forefront of forest research and operations.

show abstract

Section: Response Designmentioning

confidence: 99%

Using remotely sensed data to construct and assess forest attribute maps and related spatial products

McRoberts

Cohen

Næsset

et al. 2010

Scandinavian Journal of Forest Research

110

View full text Add to dashboard Cite

show abstract

“…However, the non-thematic errors arising from misregistration can be a significant contribution to the error matrix that may then mask the real thematic accuracy information. Powell et al [19] determined that more than 30% of thematic errors for Landsat TM could be attributed to the positional errors. This issue is more severe in heterogeneous landscapes where several map classes are located together [11].…”

Section: Introductionmentioning

confidence: 99%

The Impact of Positional Errors on Soft Classification Accuracy Assessment: A Simulation Analysis

Congalton

Pan

2015

Remote Sensing

View full text Add to dashboard Cite

Abstract:Validating or accessing the accuracy of soft classification maps has rapidly developed over the past few years. This assessment employs a soft error matrix as generalized from the traditional, hard classification error matrix. However, the impact of positional error on the soft classification is uncertain and whether the well-accepted half-pixel registration accuracy is suitable for the soft classification accuracy assessment is unknown. In this paper, a simulation analysis was conducted to examine the influence of positional error on the overall accuracy (OA) and kappa in soft classification accuracy assessment under different landscape conditions (i.e., spatial characteristics and spatial resolutions). Results showed that with positional error ranging from 0 to 3 soft pixels, the OA-error varied from 0 to 44.6 percent while the kappa-error varied from 0 to 93.7 percent. Landscape conditions with smaller mean patch size (MPS) and greater fragmentation produced greater positional error impact on the accuracy measures at spatial resolutions of 1 and 2 unit distances. However, this trend did not hold for spatial resolutions of 5 and 10 unit distances. A half of a pixel was not sufficient to keep the overall accuracy error and kappa error under 10 percent. The results indicate that for soft classification accuracy assessment the requirement for registration accuracy is higher and depends greatly on the landscape characteristics. There is a great need to consider positional error for validating soft classification maps of different spatial resolutions. OPEN ACCESSRemote Sens. 2015, 7 580

show abstract

“…However, there are few published papers about landslide susceptibility assessment that provide information on the different classification methods. Measures should therefore be adopted to analyze the weakness of a particular classification strategy (Powell et al, 2004) or to compare two or more classification techniques (Foody, 2004).…”

Section: Introductionmentioning

confidence: 99%

“…The 'error matrix' (also known as confusion matrix, confusion table, and contingency table) is central to most measures of thematic map accuracy (Story and Congalton, 1986;Smits et al, 1999;Foody, 2004;Powell et al, 2004 andGupta et al, 2008). Once a prediction threshold has been adopted by the mathematical model, the binary prediction (failed/unfailed) can be compared with the landslide observed.…”

Section: Introductionmentioning

confidence: 99%

Statistical and spatial analysis of landslide susceptibility maps with different classification systems

Baeza¹,

Lantada

Amorim

2016

Environ Earth Sci

View full text Add to dashboard Cite

A landslide susceptibility map is an essential tool for land-use spatial planning and management in mountain areas. However, a classification system used for readability determines the final appearance of the map and may therefore influence the decision-making tasks adopted. The present paper addresses the spatial comparison and the accuracy assessment of some well-known classification methods applied to a susceptibility map that was based on a discriminant statistical model in an area in the Eastern Pyrenees. A number of statistical approaches (Spearman's correlation, Kappa index, Factorial and Cluster analyses and Landslide density index) for map comparison were performed to quantify the information provided by the usual image analysis.The results showed the reliability and consistency of the Kappa index against Spearman's correlation as accuracy measures to assess the spatial agreement between maps. Inferential tests between unweighted and linear weighted Kappa results showed that all the maps were more reliable in classifying areas of highest susceptibility and less reliable in classifying areas of low to moderate susceptibility. The spatial variability detected and quantified by Factorial and Cluster analyses showed that the maps classified by Quantile and Natural Break methods were the closest whereas those classified by Landslide Percentage and Equal Interval methods displayed the greatest differences. The difference image analysis showed that the five classified maps only matched 9% of the area. This area corresponded to the steeper slopes and the steeper watershed angle with forestless and sunny slopes at low altitudes. This means that the five maps coincide in identifying and classifying the most dangerous areas. The Equal Interval map overestimated the susceptibility of the study area and the Landslide Percentage map was considered to be a very optimistic model. The spatial pattern of the Quantile and Natural Break maps was very similar but the latter was more consistent and predicted potential landslides more efficiently and reliably in the study area.

show abstract

Sources of error in accuracy assessment of thematic land-cover maps in the Brazilian Amazon

Cited by 171 publications

References 29 publications

Using remotely sensed data to construct and assess forest attribute maps and related spatial products

Using remotely sensed data to construct and assess forest attribute maps and related spatial products

The Impact of Positional Errors on Soft Classification Accuracy Assessment: A Simulation Analysis

Statistical and spatial analysis of landslide susceptibility maps with different classification systems

Contact Info

Product

Resources

About