This slide is about introduction of blurred image recognition system using legendre's moment invariant algorithm and explain about blurred image will be recognized and converted into original image
2. OBJECTIVES:
- The main objective of this project is to recognize the
blurred image
- Blurred image recognition is used for restorage
purpose
- Applicable in automatic target recognition &
tracking, character recognition, 3D scene analysis &
reconstruction.
3. EXISTING SYSTEM:
- Blurred image recognition by complex moment invariants, this
is existing system , blurred image was recognized by using the
complex moments .
- Complex moments are with respect to centrally symmetric
blur, this does not provide the recognition accuracy & also it is
sensitive to noise ,this is due to the fact that the polynomials are
not orthogonal.
4. PROPOSED SYSTEM:
- The proposed system is blurred image recognition by using
orthogonal moments .
- The orthogonal moments are better than the other types of
moments in terms of information redundancy & are most robust
to noise.
- The performance of the proposed descriptors is evaluated with
various point spread functions and different image noises.
- The proposed descriptors are more robust to noise & have better
discriminative power than the methods based on complex
moments
5. INTRODUCTION:
- One of the most frequent tasks in image processing is the
recognition of an image (or, more frequently, of an object on
the image) against images stored in a database.
- Whereas the images in the database are supposed to be ideal,
the acquired image represents the scene mostly in an
unsatisfactory manner.
- Because real imaging systems as well as imaging conditions are
imperfect, an observed image represents only a degraded
version of the original scene.
6. CONT…
- Blur is introduced into the captured image during the imaging
process by such factors as diffraction, lens aberration, wrong
focus, and atmospheric turbulence.
- The widely accepted standard linear model describes the
imaging process by a convolution of an unknown original (or
ideal) image f ( z , y ) with a space-invariant point spread
function (PSF) h(x, Y)
- where g(z,y) represents the observed image. The PSF
h ( z , y )describes the imaging system, and in our case, it is
supposed to be unknown.
8. CONT..,
INPUT IMAGE:
- Image is captured through the camera , if that image is in
unsatisfactory manner means known as blurred image
- The images are affected because of the following factors,
1. Wrong focusing
2. Atmospheric turbulence
3. Lens aberration
9. Cont..,
- There are different types blurred images , some of
them are,
- Zoom Blur
- Motion Blur
- Atmospheric Blur
- Domain Shifting
- Threshold Blur
10. Cont..,
ZOOM BLUR:
- This type of image is created due to long
focusing of the camera lens i.e out of focusing the
image
15. ADD NOISE TO AN IMAGE:
- Varies noises are ,
- White Gaussian noise
- Salt & pepper noise
- Noises are added , because it only gives
recognization process.
- From that, define the filter co- efficient
17. Cont..,
LEGENDRE MOMENTS:
- The blurred image is recognized by using the legendre
moments invariants
- Orthogonal moments are mainly used to recognize the
blurred image
- Orthogonal moments cover the whole image during the
recognization process
18. CONT..,
BLUR INVARIANTS:
- The blurred image is compared with the database , by using
the orthogonal moments
- Blur are some type of noises( gaussian noise with standard
deviation and salt & pepper noise)
- Here , calculate the point spread function for deblurring
the image i.e calculate the blur invariants
19. EDGE DETECTION:
- It function is mainly detect the edges of an
image
- Edges are used to reconstruct the image
20. MASK CREACTION :
- Mask Creation is based upon the PSF values i.e filter
values
- Apply the convolution between the original image
with the image prior , from that deblur the image
21. Cont.,
RECONSTRUCTED IMAGE:
- Finally , the original image is reconstructed by using this
moments invariants method
- This will provide the greatest accuracy compared with
the previous method
24. START
Read an image from
workspace
Add noise to an image
Choose the noise to
be added
Choose the
noise
if = 1
Apply White Gaussian
noise
Display the image
A
FLOW CHART:
25. if = 2
Apply salt & pepper
noise
Display the image
If = 3
Noise free Display the image
If > 3
Terminate
B
A
26. Find the blur invariants
Perform the edge
detection
Load filter values
Create the mask
Apply convolution
between
unknown image
with blurred image
Reconstructed image
B