Introducing New Indices for Accuracy Evaluation of Classified Images Representing Semi-Natural Woodland Environments.

A range of accuracy indices for determining the optimal outputs from the classification of multispectral remotely sensed data is evaluated. Airborne Thematic Mapper imagery of semi-natural woodland was used in conjunction with an in situ data set. Indices of classification accuracy were unable to distinguish substantial differences in classified images because they are based only on errors of omission, accounting for only a proportion of the errors in classification. The Classification Success Index (CSI) is introduced here to estimate the overall effectiveness of classification, considering all output classes and using both errors of omission and commission from the error matrix. The Individual Classification Success Index (ICSI) is introduced which accounts for the classification success of a specific class. Finally, the Group Classification Success Index (GCSI) measures classification success for the most important classes in the area of interest. These new indices were found to offer considerable improvement over existing approaches.

[1]  J. Campbell.  Introduction to remote sensing , 1987 .

[2]  Jacob Cohen,et al.  Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. , 1968 .

[3]  George Alan Blackburn,et al.  Filling the gaps: remote sensing meets woodland ecology , 1996 .

[4]  B. Berry.  Sampling, coding, and storing flood plain data. , 1962 .

[5]  George Alan Blackburn,et al.  An ecological survey of deciduous woodlands using airborne remote sensing and geographical information systems (GIS). , 1997 .

[6]  Russell G. Congalton,et al.  Assessing the accuracy of remotely sensed data : principles and practices , 1998 .

[7]  U. Helldén,et al.  A test of landsat-2 imagery and digital data for thematic mapping illustrated by an environmental study in northern Kenya, Lund University , 1980 .

[8]  Paul M. Mather,et al.  Computer Processing of Remotely-Sensed Images: An Introduction , 1988 .

[9]  P. H. Swain,et al.  Two effective feature selection criteria for multispectral remote sensing , 1973 .

[10]  R. Congalton.  A Quantitative Method to Test for Consistency and Correctness in Photointerpretation , 1983 .

[11]  C. Woodcock,et al.  The factor of scale in remote sensing , 1987 .

[12]  Stephen E. Fienberg,et al.  Discrete Multivariate Analysis: Theory and Practice , 1976 .

[13]  A. Hay.  Sampling designs to test land-use map accuracy , 1979 .

[14]  S. Kuntz,et al.  Applications of image texture in forest classification , 1994 .

[15]  George Alan Blackburn,et al.  Measurement of the spectral directional reflectance of forest canopies: A review of methods and a practical application , 1994 .

[16]  Philip H. Swain,et al.  Remote Sensing: The Quantitative Approach , 1981, IEEE Transactions on Pattern Analysis and Machine Intelligence.