摘要
在现有的基于稀疏表示的图像超分辨率算法的基础上,提出了一种新的基于模糊聚类的超分辨率重建算法,并使用L-曲线法确定正则化参数,有效降低了图像的边缘锯齿效应,提升了图像整体的平滑性,改善了基于稀疏约束算法的主客观重建质量。通过与线性插值法、Elad重建方法的仿真对比分析,基于模糊聚类的超分辨率重建方法可以显著提高果实自动化采摘图像的超分辨率重建效果。
Based on existing image sparse representation based super-resolution algorithm, a new super-resolution reconstruction algorithm based on fuzzy clustering and L-curve method was proposed to determine the regularization parameter, to enhance the overall smoothness of the image by effectively reducing the image jagged edge effects, effectively improve the algorithm based on subjective and objective constrained sparse reconstruction quality. By linear interpolation, Elad simulation method comparison analysis, super-resolution reconstruction method based on fuzzy clustering can significantly improve the effect of super-resolution reconstruction automated picking fruit image.
出处
《湖北农业科学》
北大核心
2014年第3期681-685,共5页
Hubei Agricultural Sciences
基金
国家科技支撑计划项目(2013BAD20B10)
关键词
超分辨率重建
稀疏表示
聚类
L-曲线
果实采摘
super-resolution reconstruction
sparse representation
clustering
L-curve
fruit harvesting
