基于DDA算法改进B-snake模型的曲线提取

Curve extraction based on B-snake model improved with DDA algorithm

在场景的三维重建中,Snake点的疏密程度决定了目标曲线能否被精确的表达。提出一种基于B-Snake的目标曲线半自动提取方法,它采用用户在目标曲线附近输入的点构成一条折线,然后经直线段光栅扫描转换的DDA算法快速得到折线经过的所有像素点,对这些像素点每隔固定间距进行采样,经采样后的像素点即为初始Snake点,最后用基于贪婪算法的B-Snake模型来提取目标曲线。与已有的Snake算法相比较,由于采用DDA算法提高了算法的自动化水平,避免了手工拾取大量初始蛇点;由于可以得到较密的Snake点,从而可以精确地表达目标曲线。实验结果表明,利用该方法能够精确地提取目标曲线。

In 3D scene reconstruction, the spacing of the ~nake points affects the accurate expression of the target curve. This paper offers a self-automatic B-snake based curve extraction algorithm, which uses the polyline constructed from the nearby points of the object curve inputted by the user, then all the pixel points passing through the polyline are obtained rapidly using the DDA algorithm and are sampled in a fixed interval and the object curve is extracted using the B-snake model based on Greedy algorithm. Compared with previous approaches, the proposed method improves the automation level of the algorithm because of using the DDA algorithm and does not involve manually picking U Plarge numbers of initial snake points. Because denser snake points can be obtained, the target curve can be expressed relatively accurately. Experiment results show that the proposed method can extract object curve accurately.