边界追踪及Freeman码在定量金相中的应用

根据不同情况,对多相合金中的颗粒状第二相进行二值化、颗粒分割、孔洞填充以及滤波等一种或多种图像预处理,得到离散的多目标黑白二值图像.为了统计分析多个颗粒,提出了多目标边界追踪算法,利用八连通边界追踪得到的每个颗粒的唯一标志点及其边界的Freeman码,由此可以获得体积分数、周长、面积、形状因子、平均自由程等第二相的特征参数.实验结果表明,在完成要求的预处理工作之后,该算法在离散多目标统计中实用有效.

Application of boundary-tracing and Freeman code in quantitative metallography

The grainy second phase of multi-phase alloy was processed by various image pretreatments according to different conditions to obtain binary image of the discreet multi-object. These image pretreatments include two-value transformation, grain segmentation, hole-filling and filtering. To analyze the multi-object, a boundary tracing algorithm for multi-object was presented. With the algorithm, the Freeman code of every grain's boundary was produced by 8 direction boundary tracing, thus the second phase characteristic parameters, such as volume fraction, perimeter, area, shape factor, mean free path etc, were obtained. The experiment results show that the algorithm is effective for discreet multi-object analysis by having done the above-mentioned pretreatments.