Scribd
Upload
41 views3 pages

Image Processing Technique

The document describes a noise removal algorithm for images. It begins with converting the original color image to grayscale to reduce processing power. Then it equalizes the histogram to improve overall contrast and balance intensity values. Next it applies blurring to smooth the image and reduce noise before edge detection. Edge detection compares pixel contrasts to identify object edges. Dilation is used to expand regions and shapes. Finally, noise removal identifies and removes blobs that are not the correct size, aspect ratio or location to be a license plate.

Uploaded by

pmadoh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
41 views3 pages

Image Processing Technique

The document describes a noise removal algorithm for images. It begins with converting the original color image to grayscale to reduce processing power. Then it equalizes the histogram to improve overall contrast and balance intensity values. Next it applies blurring to smooth the image and reduce noise before edge detection. Edge detection compares pixel contrasts to identify object edges. Dilation is used to expand regions and shapes. Finally, noise removal identifies and removes blobs that are not the correct size, aspect ratio or location to be a license plate.

Uploaded by

pmadoh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

4.

Noise Removal Algorithm

Noise in Image

Noise in image is a random variation of image quality such as film grains or improper
brightness and colour in a part of image. Noise removal algorithms are used to implement in
some corrupted image or images that are hardly visible. Noise removal algorithms are
techniques used to decreasing noise and increase the overall visibility of an image by using
mathematical calculation in pixels.

Greyscale conversion

Using original colour image convert into grey images which only contain brightness
information of the image. The algorithm is using the average value from the total sum of
three RGB value. For a computer to recognition, grey value is enough for process to get
useful information and saving the processing power of the computer. Grey scale image is
easier for computer to understand the contour or edges of an object and background on an
image.

Equalize Histogram

Equalize histogram are used to implement in an image to improve overall contrast with a
balance of intensity values. It spread out the most used intensity values in the overall length
of intensity range of an image (Sudhakar, 2017). This will made some changed on the lower
contrast part and higher contrast part to be balance, so it would not become over exposed and
under exposed.

This Photo by Unknown Author is licensed under CC BY-SA


Blur

Using blur operation before edge detection are tends to be smoothing the image by reducing
the noise of the image. The blurring algorithm sum up the total neighbours pixels value and
divide by the total number of pixels, this allow the image be smoothen and blur which made
easy to detect the edges.

Edge Detection

Edge detection is comparing the contrast of the pixel between two sides to get detect of the
part of an image is an edge of the object. Every pixel is comparing neighbour pixels and get
the difference, if the difference has reached a specific threshold that means the pixel is an
edge of an object. Edge detection is useful to getting the contour of object in an image, and
the contour are useful information for object recognition.

This Photo by Unknown Author is licensed under CC BY-SA

Dilation

The dilation operation is used to expand the region of the shapes contain in the input image.
Dilation algorithm used a mask to check neighbour pixels is either white or black, using ‘or’
operation the modify the pixels, if the one of the neighbour pixels is white, it should be white.
Noise removal

The width of the car plate should be longer than the height of the car plate, so using the logic
to remove all noise with over height noise blob. Moreover, calculating the size of the blob
divide by the area of its box of rectangle and multiple by 100 to get the percentage of the blob
occupy how many percent of the rectangle. If the percentage is low, that means must not be a
car plate and remove the noise. Furthermore, the car plate should be near over the central of
the image. Thus, which blob is near over the edges of the image should be removed as a
noise.

You might also like