Curve-Fitting Algorithm for the Inspection of Subtle Feature Lines on Automotive Outer Panels.

Abstract

Feature lines on automotive outer panels are difficult to make, measure, and inspect. This paper proposes an algorithm to fit a line-arc-line curve to the measured data points of a subtle feature line. First, based on an iterative method and random sampling consensus, the algorithm separates points corresponding to a circular arc. Further, the lines tangent to the circle are estimated based on the other data points. The algorithm repeats the procedure until the error is lower than the specified tolerance. The sensitivity and applicability of the algorithm were analyzed by applying it to simulated data points, experimental specimens, and actual automotive panels. The results demonstrated that the algorithm is robust to noise and surface waves and can be applied to the automotive panel-manufacturing process.