Self-Adaptive Mathematic Description Base on Minimum Pixel Error Control for Vector Data of a Curve
Zhang Sheng-ling · Dianzi xuebao · 2008
The B-spline curve-fitting algorithm is widely used in reversible compression of vector curve data and its mathematic expression.But the error evaluation of the present algorithms mostly aims at the coordinate axes with same scale unit;it was helpless for the engineering drawing whose coordinate scale units were different.So we proposed a spline-curve fitting algorithm with parametric variable of sequence number,self-adaptive,whose least error is not more than one pixel.At the first,in algorithm,a few curve nodes was chosen from vector curve data by the approximately equal arc length,and the fitting curve was obtained by profit or loss correction that the error limit of the nodes was 0.001.Then,all raster points of the fitting curve were gone through,and their 3×3 neighborhood were examined whether raster point or raster image curve point is among them.If there was,it satisfied the error margin of one pixel.Contrarily,the distinctive mark was made on this curve segment between nodes and the next curve segment was gone on checking.For the marked curve segments,the error margin of one pixel was satisfies by means of increasing one curve node or shifting node order-numbers.The test results show that proposing algorithm is more exact and convenient.