期刊论文

Zhou, L.(profzh@126.com)

中文

华中科技大学学报(自然科学版)

1671-4512

2010

38

1

85-87

10.13245/j.hust.2010.01.025

(2003ABA053):湖北省自然科学基金 (20035002016_09):湖北省武汉市青年科技晨光计划 (Q200618002):湖北省教育厅青年项目

本校为第一且通讯机构

针对储粮害虫图像复原中点扩展函数矩阵的逆矩阵不存在的情况,提出了一种用奇异值分解和Moor-Penrose广义逆来实现图像复原的方法,通过将矩阵作奇异值分解再求其Moor-Penrose广义逆.计算结果表明,与传统方法相比,该方法在复原由运动造成的图像模糊方面能取得较好的效果,复原的图像比较清晰.

There is no general inverse of the point spread function matrix when restoring the images from pests in stored grain. Thus, a method was proposed to implement image restoration of pests in stored grain by using singular value decomposition (SVD) and Moor-Penrose general inverse. Comparison with singular value decomposition and normal image compression, the image restoration were calculated. The results show that t...