基于分水线算法的二维直方图阈值化分割在箱包DR-CT检测系统中的应用

Thresholding using two-dimensional histogram and watershed algorithm in the luggage inspection system

DR图像的区域分割是箱包DR-CT检测系统中的一个重要环节。采用二维直方图阈值化方法,在一维直方图灰度信息的基础上,引入局部空间的信息,利用分水线算法在二维直方图垂直投影上求取分割阈值,取得了较好图像分割效果。

In this paper we present a new DR image segmentation method based on two-dimensional histogram and watershed algorithm. We use watershed algorithm to locate threshold on the vertical projection plane of two-dimensional histogram. This method is applied to the segmentation of DR images produced by luggage inspection system with DR-CT. The advantage of this method is also analyzed.