regarding fake currency detection through image processing

I am working on a fake currency detection project and i want to get the security thread or the silver bromide thread.how can i be able to get the security thread as i am new to matlab and image processing? please help and few suggestion on the project. i am basically working on the indian currency.please help me out.

Answer 1

[Walter Roberson]

It appears that you will need a UV light. The security measures are described in this story in an inset box.

[ABHAYASH]

what next can i do in this regard beside edge detection??

please help?

[Walter Roberson]

Locate the note in the field of view. Detect its angle and do an image rotation to bring it in to alignment. Find the bounding box. The silver bromide thread must then be in one of two positions relative to the bounding box, depending whether the note is upside down or not. You can calculate the size the line _should_ appear as based upon the size of the bounding box. If the line is too big or too small or appears in the wrong place or there are multiple lines, then there is a problem with the note.

[Savita Hanchinal]

plaese tell me code how to create dataset of image ,converting it to the gray scale image ,and edge detection ,segmentation,faeature extraction and image comparison

[Image Analyst]

See imageDatastore(), rgb2gray(), edge(), imbinarize(), and ssim().

Also see my Image Segmentation Tutuorial:

Image Analyst's File Exchange

and the FAQ:

https://matlab.fandom.com/wiki/FAQ#How_can_I_process_a_sequence_of_files.3F

Post your image and question in a new question if you still need help.

Answer 2

[Htet Yamin Ko Ko]

I'm also working on it. First, I divided the image into four blocks and then one block is selected based on the denomination of the input image which I have recognized earlier. Then grayscale conversion, binary conversion, canny edge detection and morphology thinning operations are performed and then Hough Line Transform is applied to detect the location of the security thread. Once I have the position and then I cropped the security thread feature and then number of white pixels in the cropped image is calculated. The problem is the extracted feature is not sufficient to verify the input image. I also tried maximum intensity, minimum intensity, area, perimeter and solidity features. But the extracted features are not very different from the real currency and fake currencies. I just want to share my experiments. The steps are varied based on the type of currency you are using to verify.

Answer 3

[tarun kumar]

please can i have the code for fake currency detection using matlab

[Walter Roberson]

Yes. Unless you are prohibitted by law from using such software:

1. You can write it yourself, possibly after reading previous postings in MATLAB Answers that discuss techniques
2. You can google for completed or partly completed projects along these lines
3. You can google for videos discussing the topics and showing code
4. You can ask for volunteers to write the code for you (but I would advise you that it is very rare for anyone to volunteer to write projects like these)
5. You can hire someone to write the code for you
6. You can hire one of the sites that specialize in plagiarism to supply you with the software