摘要
颜色空间内的突变信息感知是实现深度背景下的隐藏目标识别的重要技术。传统方法采用差分进化局部对比方法进行突变信息感知,当图像分块信息不均匀时,信息提取不全。提出一种基于瑞利差分的颜色空间突变信息内容感知方法,进行颜色空间背景全局性信息建模,计算非显著性突变信息。基于瑞利分布的非均匀稀有度特性,得到背景颜色空间和突变信息目标区域分离结果和差异性特征,实现突变信息内容感知。实验结果表明,算法能能使突变信息在深度颜色空间背景下突出显示,较好地检测出图像的显著性特征。在非显著性微小目标探测领域应用前景较好。
Color space mutation information perception is an important technology to realize the hidden depth target recogni.tion under the background. Traditional methods uses differential evolution mutation information perception, when the imageblock information is asymmetry, information extraction is not complete. A color space mutation information perception meth.od is proposed based on Rayleigh difference calculation, background of global information modeling is obtained, and nonsignificant abrupt information is computed. Based on the characteristic of non uniform rarity of Rayleigh distribution, obtainthe background color space and mutation information of target regional separation results, mutation information contentaware is achieved. The experimental results show that the algorithm can make the mutation information displayed in thedepth of color space, it is highlighted in the context, and the significant feature of the image is detected better, so it has agood application prospect in the non significant and small target detection field.
出处
《科技通报》
北大核心
2014年第12期118-120,共3页
Bulletin of Science and Technology
关键词
瑞利分布
差分
颜色空间
突变信息
rayleigh distribution
difference
color space
mutation information