Implicit surface fitting using directional constraints (Q5942335)

The authors develop an algorithm for fitting implicitly defined surfaces to measured data, using the least squares norm, which takes specific account of the probe directions used to generate the data. The process is a natural generalization of a method given for parametrically defined surfaces. However, as well as allowing greater generality in the kind of surfaces being fitted, it also has other advantages. For example, the absence of the extra position parameters defining the points can make the calculation of the Jacobian matrix simpler. In addition, the orientation of the surface does not need special treatment and there is evidence that this can lead to computational savings.