中文摘要：

通过CCD摄像技术，采集中密度纤维板（MDF）铣削过程中切屑流的高清晰图像。利用图像处理的最大类间方差法，确定所采集图像的阈值；利用该阈值对所采集的图像进行分割，得到切屑流的二值图像；基于八连通准则，对该二值图像中切屑流的边界进行搜索检测，求得切屑流边界线像素的位置坐标，确定切屑流边界线的位置；采用最小二乘法直线拟合边界线，确定切屑流的扩散角λ。结果表明：当铣削速度达到60m·s^-1后，其扩散角λ存在明显增大趋势。

英文摘要：

The velocities of the chip flow could be as high as 60 m · s^-1 during MDF milling process, which made chip flow boundary and diffusion angle very difficult to be estimated. Meanwhile, the efficiency of dust absorbing could also be directly affected by characteristics of chip flow. In this research, clear images of chip flow were taken by CCD cameras from MDF milling process. The thresholds of these images were computed by applying Otsu＇s method. Based on those calculated thresholds, the images obtained from CCD cameras were converted to the binary images. Principle of eightconnectivity was adopted to collect the coordinates of chip flow boundary pixels. The boundary lines of the chip flow in the images could be detected and further traced by linking suitable boundary pixels based on their coordinates. Furthermore, diffusion angles of chip flow were determined when the least square method was applied. The experimental results showed that the diffusion angle of the chip flow dramatically increased when milling velocity was higher than 60 m · s^-1 Therefore, image processing method presented in this paper could provide the insightful understandings of characteristics of chip flow. Also this study could help further researches on both of high-efficiency dust removal systems and dust collection hood design.