摘要
针对迷彩设计中常用的彩色图像分割方法计算量过大、分割效果不明显等问题,提出一种改进方法。先确定背景中与主色接近的军标色,然后结合区域分割与聚类分析的思想,得到不同特征的分割效果。实验以图像色度为分割特征,在选取不同阈值的情况下,列举并分析了所得结果。实验结果表明,改进后的方法能够较好地分割背景图像并保留其细节纹理特征,得到较为理想的设计效果。
In order to resolve the problems of huge calculation workload and unobvious segmentation effects in the process of camouflage design,proposes an improving method of color-image segmentation.It firstly confirms the nearest military standard colors,following that combines the method of region-dependant segmentation with the method of cluster analysis,finally gains the segmentation results of various characters.The experiment uses the hue of the image as the segmentation character.It enumerates and analyses the results in the case of the different thresholds.The experimental results demonstrate that the improved method is able to segment the background-images better and preserve more texture details,and thus attain ideal effect on design.
出处
《兵工自动化》
2010年第10期64-65,75,共3页
Ordnance Industry Automation
基金
总后基建营房部科研项目(No.营080709)
关键词
迷彩设计
彩色图像分割
聚类分析
区域分割
camouflage design
color image segmentation
cluster analysis
region-dependant segmentation
