This tool calculates the root-mean-square-error (RMSE) or root-mean-square-difference (RMSD) from two input rasters. If the two input rasters possess the same number of rows and columns, the RMSE is calucated on a cell-by-cell basis, otherwise bilinear resampling is used. In addition to RMSE, the tool also reports other common accuracy statistics including the mean verical error, the 95% confidence limit (RMSE x 1.96), and the 90% linear error (LE90), which is the 90% percentile of the residuals between two raster surfaces. The LE90 is the most robust of the reported accuracy statistics when the residuals are non-Gaussian. The LE90 requires sorting the residual values, which can be a relatively slow operation for larger rasters.

See Also

paired_sample_t_test, wilcoxon_signed_rank_test

Function Signature

def root_mean_square_error(self, input: Raster, reference: Raster) -> str: ...

Project Links

WbW Homepage User Manual Support WbW