Piecewise-Linear Curve Approximation for a Set of Digital Points

이진 영상점들에 대한 구역별 직선 근사화

Typical application of detecting piecewise-linear curves includes vectorizing of scanned drawings whirh is a vital step in installing any geographic information system. This paper considers the problem of optimally approximating a piecewise linear curve to a set of digital points while satisfying given intersection angles between each pair of neighboring lines. The criterion for optimality is to minimize the sum of squared deviations. The problem is formulated as an unconstrained nonlinear programming model. An algorithm which guarantees an optimal solution is then proposed and its validity is tested with both a synthetically generated image and a real image. The test results illustrate the excellent performance of the proposed algorithm.