It should be noted that these results are not the last word in the modeling of centroid errors in undersampled, noisy data. They reflect the application of a particular algorithm under conditions of "well-behaved" noise. The algorithm has not yet been tested under field conditions.

The four graphs shown above are summarized in a single three-D interactive model.

Care must be taken in the interpretation of the standard deviation of the centroid error. Due to the very quantization being analyzed, outliers are introduced which may undermine the assumption of homogeneity of variance across the family of simulations presented here. See the Summaries below for statistical and graphical descriptions of typical and worst-case findings for each simulation.

Summaries for each combination of undersampling and noise are linked below:

		Pixels in quantized spot: 29x29, 17x17, 12x12, 9x9, and 7x7
Gaussian additive noise	0.00000	Get Page
	0.00125	Get Page
	0.00250	Get Page
	0.00375	Get Page
	0.00500	Get Page

Page 1

Page 2

Page 3

Page 4

Appendix