You are here:
Pixels, Blocks of Pixels, and Polygons: Choosing a Spatial Unit for Thematic Accuracy Assessment
Citation:
Stehman, S. V. AND J. D. WICKHAM. Pixels, Blocks of Pixels, and Polygons: Choosing a Spatial Unit for Thematic Accuracy Assessment. REMOTE SENSING OF ENVIRONMENT. Elsevier Science Ltd, New York, NY, 115(12):3044-3055, (2011).
Impact/Purpose:
Accuracy assessment is an established component of the process of creating and distributing thematic maps. The fundamental basis of an accuracy assessment is a location-specific comparison between the map classification and the ground condition or “reference” classification. Reporting thematic map accuracy in the form of an error matrix is a standard practice when the map and reference classifications are categories such as land-cover classes (Story and Congalton 1986; Congalton and Green 1999, 2008). However, opinions vary on the appropriate spatial unit (e.g., pixel, block of pixels, or polygon) underlying the comparison of the map and reference classifications 49 that provides the data summarized by the error matrix. Stehman and Czaplewski’s (1998) review of 33 map accuracy assessments illustrates this lack of consensus regarding choice of assessment unit as they reported a pixel was used as the spatial unit 14 times, a block of pixels (e.g., 2x2, 3x3) was used 10 times, and a polygon was used 9 times. Strahler et al. (2006), Janssen and van der Wel (1994) and Richards (1996) support a pixel-based assessment, whereas Congalton and Green’s (1999) influential and oft-cited accuracy assessment book recommends using a block of pixels or a polygon.
Description:
Pixels, polygons, and blocks of pixels are all potentially viable spatial assessment units for conducting an accuracy assessment. We develop a statistical population-based framework to examine how the spatial unit chosen affects the outcome of an accuracy assessment. The population may be visualized as a difference map created by overlaying a complete coverage reference classification and the target map being evaluated. The per-class areas of agreement and disagreement derived from this population are summarized by a population error matrix and accuracy parameters (e.g., overall, user’s and producer’s accuracies). The population and accuracy parameters are strongly affected by the protocols implemented for the response design which include the choice of spatial unit, how within-unit homogeneity is addressed when assigning class labels, and the definition of agreement between the reference and map classification. Several complete coverage populations are used to illustrate how accuracy results are affected by the spatial unit chosen for the assessment. The sampling design implemented for accuracy assessment does not change the population or accuracy parameters, but the choice of spatial unit will influence decisions on whether to use strata and clusters in the design. A universally best spatial assessment unit does not exist, so it is critical to recognize how the population, accuracy parameters, and sampling design are impacted by the choice of spatial unit.