Uncertainty-driven viewpoint planning for 3D object measurements

In this paper we present an uncertainty-driven viewpoint planning approach for measurement and digitalization of free form 3D objects. The object surface is first decomposed into a number of cross section curves, and each curve is reconstructed by a closed B-spline curve. Then, we analyze the uncertainty of the B-spline model. Based on this, we predict the information gain for each cross section curve for candidate views. After predicting the information again of all the B-spline models into the view space. The viewpoint that contains maximal information gain about the object is selected as the Next Best View. Experimental results are given in the digitization and reconstruction of freeform objects using our view planning method.