1: \begin{abstract}

2:   Discrete geometric estimators aim at estimating geometric

3:   characteristics of a shape with only its digitization as input data.

4:   Such an estimator is multigrid convergent when its estimates tend

5:   toward the geometric characteristics of the shape as the

6:   digitization step $h$ tends toward 0.  This paper studies the

7:   multigrid convergence of tangent estimators based on maximal digital

8:   straight segment recognition. We show that such estimators are

9:   multigrid convergent for some family of convex shapes and that their

10:   speed of convergence is on average

11:   $\BigO{h^{\frac{2}{3}}}$. Experiments confirm this result and

12:   suggest that the bound is tight.

13: \end{abstract}

14: