Estimating per‐pixel thematic uncertainty in remote sensing classifications

Standard methodologies for estimating the thematic accuracy of hard classifications, such as those using the confusion matrix, do not provide indications of where thematic errors occur. However, spatial variation in thematic error can be a key variable affecting output errors when operations such as change detection are applied. One method of assessing thematic error on a per‐pixel basis is to use the outputs of a classifier to estimate thematic uncertainty. Previous studies that have used this approach have generally used a single classifier and so comparisons of the relative accuracy of classifiers for predicting per‐pixel thematic uncertainty have not been made. This paper compared three classification methods for predicting thematic uncertainty: the maximum likelihood, the multi‐layer perceptron and the probabilistic neural network. The results of the study are discussed in terms of selecting the most suitable classifier for mapping land cover or predicting thematic uncertainty.

[1]  Giles M. Foody,et al.  Mapping Land Cover from Remotely Sensed Data with a Softened Feedforward Neural Network Classification , 2000, J. Intell. Robotic Syst..

[2]  N. Campbell,et al.  Derivation and applications of probabilistic measures of class membership from the maximum-likelihood classification , 1992 .

[3]  Kyle Brown Per-Pixel Uncertainty for Change Detection Using Airborne Sensor Data , 2004 .

[4]  Edzer Pebesma,et al.  Spatial aggregation and soil process modelling , 1999 .

[5]  M. Ehlers,et al.  A framework for the modelling of uncertainty between remote sensing and geographic information systems , 2000 .

[6]  A. Comber,et al.  Assessment of a Semantic Statistical Approach to Detecting Land Cover Change Using Inconsistent Data Sets , 2004 .

[7]  Stefano Tarantola,et al.  Uncertainty and sensitivity analysis: tools for GIS-based model implementation , 2001, Int. J. Geogr. Inf. Sci..

[8]  Giles M. Foody,et al.  Deriving thematic uncertainty measures in remote sensing using classification outputs , 2006 .

[9]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[10]  Sucharita Gopal,et al.  Uncertainty and Confidence in Land Cover Classification Using a Hybrid Classifier Approach , 2004 .

[11]  John A. Richards,et al.  Analysis of remotely sensed data: the formative decades and the future , 2005, IEEE Transactions on Geoscience and Remote Sensing.

[12]  Giles M. Foody,et al.  An evaluation of some factors affecting the accuracy of classification by an artificial neural network , 1997 .

[13]  James R. McDonald,et al.  Using physics-based modeler outputs to train probabilistic neural networks for unexploded ordnance (UXO) classification in magnetometry surveys , 2001, IEEE Trans. Geosci. Remote. Sens..

[14]  I. Kanellopoulos,et al.  Strategies and best practice for neural network image classification , 1997 .

[15]  Archana Mahapatra,et al.  Modeling the Uncertainty in Orientation of IRS-1C/1D with A Rigorous Photogrammetric Model , 2004 .

[16]  F. Maselli,et al.  Use of probability entropy for the estimation and graphical representation of the accuracy of maximum likelihood classifications , 1994 .

[17]  Filipe Aires,et al.  Neural Network Uncertainty Assessment Using Bayesian Statistics: A Remote Sensing Application , 2004, Neural Computation.

[18]  J. D. Paola,et al.  The Effect of Neural-Network Structure on a Multispectral Land-Use/Land-Cover Classification , 1997 .

[19]  R. W. McClendon,et al.  LAND-USE CLASSIFICATION OF GRAY-SCALE AERIAL IMAGES USING PROBABILISTIC NEURAL NETWORKS , 2004 .

[20]  R. G. Pontlus Quantification Error Versus Location Error in Comparison of Categorical Maps , 2006 .

[21]  Mahmood R. Azimi-Sadjadi,et al.  Comparison of two different PNN training approaches for satellite cloud data classification , 2001, IEEE Trans. Neural Networks.

[22]  P. Gong,et al.  Mapping Ecological Land Systems and Classification Uncertainties from Digital Elevation and Forest-Cover Data Using Neural Networks , 1996 .

[23]  D. Dean,et al.  Combining location and classification error sources for estimating multi-temporal database accuracy , 2001 .

[24]  L.L.F. Janssen,et al.  Accuracy assessment of satellite derived land - cover data : a review , 1994 .

[25]  Bayya Yegnanarayana,et al.  Supervised texture classification using a probabilistic neural network and constraint satisfaction model , 1998, IEEE Trans. Neural Networks.

[26]  Paul M. Mather,et al.  Pruning artificial neural networks: An example using land cover classification of multi-sensor images , 1999 .

[27]  Mahmood R. Azimi-Sadjadi,et al.  A temporally adaptive classifier for multispectral imagery , 2004, IEEE Transactions on Neural Networks.

[28]  Donald F. Specht,et al.  Probabilistic neural networks , 1990, Neural Networks.

[29]  Giles M. Foody,et al.  Local characterization of thematic classification accuracy through spatially constrained confusion matrices , 2005 .

[30]  Giles M. Foody,et al.  Uncertainty in Remote Sensing and GIS: Fundamentals , 2006 .

[31]  Chris T. Kiranoudis,et al.  The performance of pixel window algorithms in the classification of habitats using VHSR imagery , 2006 .

[32]  P. Atkinson,et al.  Introduction Neural networks in remote sensing , 1997 .

[33]  Jon Atli Benediktsson,et al.  Neural Network Approaches Versus Statistical Methods in Classification of Multisource Remote Sensing Data , 1989, 12th Canadian Symposium on Remote Sensing Geoscience and Remote Sensing Symposium,.

[34]  Paul M. Mather,et al.  The use of backpropagating artificial neural networks in land cover classification , 2003 .