基于图元的手绘几何图形识别

Geometric Sketching Recognition Based on Graphics Unit

对通过手写板、光笔、数字笔等输入设备绘制的几何图形,采用基于图元表示法进行识别。该方法首先把输入的笔画组分解为图元(直线、园、圆弧等),对分割出来的各个图元进行识别,进而对识别的图元进行排序及位置关系提取,最后对图形进行拟合规整、输出识别结果。识别过程简单,速度快,准确率高,与输入笔序无关。

The geometric graphics drawn through devices （Tablet,Light pen, Digital pen etc） are recognized by the method based on graphics unit. Firstly we divide the input strokes into several graphics unit （E. g. line circle etc）, and recognize every divided graphics unit. Then sort all the graphics units and pick up the relationship among graphics units, finally geometrics graphics approximation and print out the recognized result. The process of recognition is simply and fast speed, high accuracy rate, and the method doesn＇t require the fixed order of sketched strokes.