工程扫描图象骨架线提取的研究

骨架线提取是图象数字化的关键步骤,本文分去噪和细化两步来提取图象的骨架线。通过仔细研究工程扫描图象的特征,提出一种基于密度划分的去噪算法。基于密度划分的原理,自定义了密度函数,并分别讨论了灰度图象和黑白二值图象两种情况。在图象细化阶段,采用一种直接根据中心象素8-邻域情况进行细化的算法。经实验证明,上述方法对工程扫描图象处理速度较快,实现简单,可用于实际应用。

Study on Figure Traction of Scanned Project Graph

Figure traction of graph is the most important step in graph digitalizing,de-noising and thinning are used in this paper to tract the figure of graph.Through specific study of the characteristic of scanned project graph,a de-noise algorithm based on the pixels density is proposed.Under the principle of density partition,it defines the density function,and discusses the grayed graph and the two-value graph respectively.During thinning phase,an algorithm of thinning by considering its eight-neighborhood conditions directly is used.After experimentation,the methods proposed above are proved to be suitable for scanned project graph,fast processing speed and easy to realize make it can be put into practice use.