图像处理-图像复原算法-20110536-周延文

逆滤波

;

维纳滤波

;

有约束的最小二乘方滤波

Research of Algorithms for Image Restoration

With

the

development

of

society's

productivity,

image

processing

technology has already far broken through the early application, widely applied

to

a

lot

of

fields,

such

as

scientific

research,

industrial

and

agricultural

production, military technology, government department, health care, etc. The

image

processing

technology

includes

various

kinds

of

processing,

for

example,

carrying

on

the

digitization,

coding,

analyzing

to

the

image

etc.

,Certainly

the

degraded

image

processing

is

a

very

important

research

field in digital image processing, its research results are being widely applied

1

to each field stated above.

In

this

paper,

the

author

carried

on

research

to

some

commonly

used

classical image restoration algorithms. In the case of knowing the model of the

image

degradation

the

observed

images

are

restored

using

inverse

filtering.

Wiener filtering and constrained least squares filtering algorithm .A wealth of

empirical data on die parameter selection of the above algorithms is obtained,

and die experimental results are analyzed and summarized

Key words in age restoration inverse filtering .Wiener filtering constrained least

squares filtering

1.

图像复原技术

1.1

图像退化模型

图像复原，是指去除或减少在获得观测图像过程中产生的降质影响，因为使

图像 模糊的原因很多，所以通常用统一的数学模型对图像的模糊过程进行描述。

如果一幅原始 图像

f(x,

y)

，在一个退化函数 和一个加性噪声项的作用下，生成

观测图像

g(x,

y)

，一般的，退化函可以认为是线性、位置不变的，噪声也与位

置和当前像素值无关，退化过程可以被模型化为

(1)

其中

h(x,

y)

< br>刀是表示退化函数的空间描述，也称为

PSF,

即点扩散 函数，

.*

表示空间卷积

n(x, y)

为加性噪声。可以用向量矩阵的形式将式

( 1)

的退化模型

表

g=H

f +n

(2)

在式

( 2)

中，

1Y

是观测图像，假设其大小是

N xN, f