摘要
为了实现灵活、简洁和自适应地表示交通图像,该文将图像稀疏分解新方法引入到交通图像处理中,提出基于量子进化算法的交通图像稀疏分解方法,以加快对交通图像稀疏表示的处理速度,从而为进一步提取交通参数奠定良好基础。采用非对称图像原子构建交通图像原子库,用寻优能力强和收敛速度快的量子进化算法,实现在过完备图像原子库中搜索最佳匹配交通图像结构的原子,有效地实现对交通图像的稀疏表示。仿真实验结果表明,该方法能对交通图像进行快速、有效地稀疏分解,证实了所提出方法的可行性。
To represent a traffic image in a flexible, sparse and adaptive way, this paper introduces the image sparse decomposition method into traffic image processing and presents a novel traffic image sparse decomposition approach based on quantum-inspired evolutionary algorithms to accelerate the processing speed of traffic image representation, which is helpful to extract traffic feature parameters from images. The sparse representation for traffic images is fulfilled by using nonsymmetrical image atoms to construct a traffic image atom dictionary, and employing the quantum-inspired evolutionary algorithm with strong search capability and rapid convergence to search the best image atom from an over-complete image atom dictionary to match the local structures of traffic images. Simulation experiments show that the introduced method can obtain a sparse decomposition of traffic images in a fast and effective way, which validates the approach presented here.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2010年第1期40-45,共6页
Journal of Nanjing University of Science and Technology
关键词
图像处理
交通图像
稀疏分解
量子进化算法
image processing
traffic images
sparse decomposition
quantum-inspired evolutionary algorithms
