摘要
为了提高组织结构细节的显示,满足机器分析的要求,对图像进行多层的小波分解与重构,通过对多层分解后的低频逼近系数和各层分解的细节系数作处理,以达到对图像进行增强目的。引入的第二代曲波(ⅡCurvelet)增强算法,根据各子带的噪声水平分别进行分段非线性增强,最后进行反变换得到增强图像。增强对比实验表明:ⅡCurvelet增强算法在整体效果和局部细节上优于小波增强算法,更能达到抑制背景噪声、突出目标细节、提高对比度的效果,对检测图像增强更加有效。
Image-enhancement is in order to improve the organizational structure of the details and meet the requirements of machinery. After the wavelet decomposition and reconstruction on the images in a muhilayer method, and then the multistorey low-frequency approximation factors and the details of all levels of decomposition are dealt with, so as to achieve the objective of enhancing images. The second-generation Curvelet enhancement algorithm acts on the images for nonlinear enhancing, according to the sub-band' s noise levels, and the last step is to process anti-enhanced image transformation by the algorithm. The contrasting experiments show that the secondgeneration Curvelet enhancement algorithm has better effects on the overall and the local details than the wavelet enhancement algorithm, and is superior for image enhancement in suppressing the setting noise, highlighting the details of goals and improving the contrasting effectiveness.
出处
《传感器与微系统》
CSCD
北大核心
2008年第12期8-10,共3页
Transducer and Microsystem Technologies
基金
国家"十一五"科技支撑计划资助项目(061341015)
关键词
图像增强
小波变换
第二代曲波算法
异物检测
image enhancement
wavelet transform
Ⅱ Curvelet algorithm
foreign-matter examination