摘要
在深入分析典型图割算法的基础上,提出了一种基于粗精交互融合和迭代图割的舰船可见光图像分割方法,主要包括矩形包围盒式的粗交互、基于高斯混合模型的迭代图割、多边形编辑式的精交互和窄带区域内的图割优化。仿真结果表明,本文方法基本达到了预期目标,交互少、分割效果好,可满足后续特征提取和目标识别的需求。
Image segmentation is one of the difficult problems in the field of computer vision. Under the conditions of automatic segmentation being difficult,it may be practical to use interactive image segmentation method. Based on the deep analysis of classic graph cut method, the segmentation method for ship visible image based on combining coarse and precise interaction and iterative graph cut is proposed, which mainly includes four steps: coarse interaction by rectangle bounding box, iterative graph cut based on Gaussian mixture model, precise interaction by polygon border editing, and graph cut refinement in the strip zone. The simulation results show that the proposed method has arrived at the expectation, i. e. , small interaction and good segmentation effects, and it is able to meet the needs of feature extraction and target recoanition.
出处
《光电子.激光》
EI
CAS
CSCD
北大核心
2012年第8期1609-1615,共7页
Journal of Optoelectronics·Laser
基金
中国博士后科学基金(20100471451)
水下测控技术国家级重点实验室基金(9140C2603051003)资助项目
关键词
交互式图像分割
图割
粗精交互融合
高斯混合模型
interactive image segmentation
graph cut
combining coarse and precise interaction
Gaussi-an mixture model