Stabilizing Heteroscedastic Noise With the Generalized Anscombe Transform. Application to Accurate Prediction of the Resolution in Displacement and Strain Maps Obtained With the Grid Method.

Michel Grediac (1) , Frédéric Sur (2)

1 IP - Institut Pascal
2 MAGRIT - Visual Augmentation of Complex Environments

The objective of this paper is to show that it is possible to predict the noise level in displacement and strain maps obtained with the grid method, but that actual noise of camera sensors being heteroscedastic, it is necessary to stabilize this noise in grid images prior to employing the predicting formulas. The procedure used for this purpose relies on the Generalized Anscombe Transform. This transform is first described. It is then shown that experimental and theoretical resolutions in strain maps obtained with the grid method are in good agreement when this transform is employed.

Domaines: Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]; Mécanique des matériaux [physics.class-ph]; Mécanique des matériaux [physics.class-ph]; Vision par ordinateur et reconnaissance de formes [cs.CV]

Michel Grediac, Frédéric Sur. Stabilizing Heteroscedastic Noise With the Generalized Anscombe Transform. Application to Accurate Prediction of the Resolution in Displacement and Strain Maps Obtained With the Grid Method.. SEM Annual Conference & Exposition on Experimental and Applied Mechanics - 2014, Jun 2014, Greenville, SC, United States. pp.225-230, ⟨10.1007/978-3-319-06986-9_25⟩. ⟨hal-01001869⟩