综合边缘检测和区域生长的红外图像分割方法

针对红外图像的特点,提出了一种综合应用边缘检测和区域生长方法的图像分割方法。其思路为:先对图像进行边缘提取,得到边缘像素点集;然后利用该点集的平均灰度和目标区域的连通性作为生长判决条件,采用区域生长法实现图像分割。仿真结果表明,该方法能快速准确有效地实现红外图像分割,避免了单独使用边缘提取或区域生长法进行图像分割时的典型分割错误。

An infrared image segmentation method based on edge detection and region growing

An image segmentation method based on edge detection and region growing is proposed in view of the features of IR image. The idea is this: we first apply edge detection technique to obtain the set of edge pixels. And then, in the region-growing step, we take the average gray level of the edge pixel set and region connectivity as the criteria to determine whether the growing process should be terminated in this direction. The simulation results show that our proposed method is fast, reliable and accurate for infrared image segmentation. This combined method helps to avoid characteristic segmentation errors when edge detection or region growing technique is adopted separately.