Digital Signal Processing Reference
In-Depth Information
the image recognition processing color images converted to grayscale images. Selec-
tion criteria is the gray-scale transformation, the dynamic range of the pixel increases,
the contrast of the image is strengthened, so that the images become clearer easy to
identify. The image graying is the first step in image processing, and it is a very im-
portant step.
Fig. 2.
Graying processing of image
2.3
Wavelet De-noising
Work to eliminate image noise is called the smoothing or filtering of the image.
Smooth has two purposes: to improve the quality of the image and extract the object
features. Image filtering methods are the wavelet filtering, average filtering, morpho-
logical filtering and median filtering. The wavelet filter is a simple and better method,
it is the layers wavelet decomposition coefficient modulus is greater and smaller
than a certain threshold value of the coefficient, respectively for processing, the wave-
let coefficients reconstruction after processing a noise canceling after the image. The
use of wavelet filtering can filter out some of the camera sometimes will affect the
image produced by the effect of external factors, such as dust.
Wavelet analysis for image filter works as follows, a noisy image signal can be ex-
pressed as:
(2)
s
(
x
,
y
)
=
f
(
x
,
y
)
+
Ee
(
x
,
y
)
(
x
,
y
=
0
,...
n
−
1
Where s (x, y) is the noisy image signal; f (x, y) is the useful signal; e (x, y) is the
noise signal; E is the noise intensity.
The noise signals are usually high-frequency signal, the useful signal is often a low
frequency signal, De-noising to eliminate high-frequency signals while retaining the low-
frequency signals. First, the wavelet decomposition of the image, it is because of the
noise included in the higher frequency details; then use the threshold to handle the form
of the decomposition of the wavelet coefficients for processing; Finally, the signal wave-
let reconstruction can achieve the purpose of the image signal de-noising. Wavelet analy-
sis for image de-noising process, broadly divided into the following paragraphs:
First is wavelet decomposition of the image signal. Select a suitable wavelet and
the appropriate level of decomposition (denoted N), then treat the analysis of the im-
age signal X N-layer decomposition.
