基于最小像素误差控制的曲线矢量数据自适应数学描述  

Self-Adaptive Mathematic Description Base on Minimum Pixel Error Control for Vector Data of a Curve

作　　者：张蓉生[1] 李立[1] 魏学锋[1] 李娜[1] 章胜玲[1] 

机构地区：[1]河海大学水利水电工程学院,江苏南京210098

出　　处：《电子学报》2008年第11期2160-2164,共5页Acta Electronica Sinica

摘　　要：矢量数据的B样条曲线拟合算法被广泛地应用于数据的可逆压缩和矢量曲线数据的数学表达.但现有算法的误差评估大都是针对各坐标轴具有相同尺度单位的,而对各坐标轴尺度单位不同的工程曲线图则无能为力.据此提出了一种不大于1个像素最小栅格误差的矢量数据自适应序号参变量样条曲线拟合算法.该算法首先从矢量数据中以近乎相等的曲线弧长选出若干型值节点,并通过盈亏修正获得节点误差限为0.001的序号参变量样条曲线拟合.然后遍历拟合曲线对应的全部栅格点,检测它所在3×3邻域内是否有矢量数据栅格点或栅格图像曲线点位于其中,若有则该点满足于1个像素的误差限,反之将该曲线型值节点段进行标记,并转到曲线的下一型值节点段继续进行检测.对不满足要求的曲线型值节点段,通过添加一个型值节点或曲线序号数的移位等使之满足误差要求.实验结果表明该算法具有更高的准确性和使用的方便性.The B-spline curve-fitting algorithm is widely used in reversible compression of vector curve data and its mathe- matic 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 firing 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.

关 键 词：矢量曲线数据 可逆压缩 最小像素误差控制 自适应 数学描述 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]