提出了基于限邻域经验模式分解(Neighborhood Limited Empirical Mode Decomposition,NLEMD)的图像增强新算法.二维NLEMD是在Huang等人EMD自适应特性基础上通过设定最大邻域(时宽)和采用邻域内局部自适应均值算法代替包络均值算法进行分解,克服以往EMD分解算法出现的灰度斑现象.本文通过NLEMD对图像细节信息的强挖掘能力来获取图像中的高频边缘信息,最后根据剩余量的整体亮度均值和整体亮度对比度自动调整剩余量来调整图像的整体亮度.实验结果证明,与以往传统增强算法相比,本文算法具有更强的细节获取能力和整体亮度可控性,增强效果优于以往传统算法.
One image enhancement algorithm based on neighborhood limited empirical mode decomposition (NLEMD) is proposed. NLEMD is one novel time-frequency analysis tool which has the adaptive features of Huang' s EMD and meanwhile adopt neighborhood limited ( max time-width) to overcome other EMD' s gray spots in images. The high frequency data is got using the ability of NLEMD, then we revise the remnants according to the illumination and the illumination contrast to adjust the whole illumination of the image. Experiments prove that the novel algorithm is efficient in image enhancement and better than current algorithms in detail achieving.