Jul 13, 2012 for the code, pic parameters here. Licence Plate Localization and letters extraction using matlab. Car License Plate Reader using MATLAB. Plate by implementing that in MATLAB. Developed A License Plate Localization using Morphology and Recognition. NUMBER PLATE LOCALIZATION SYSTEM.
Active3 years, 6 months ago
Proceedings of the Federated Conference on Abstract—License plate localization is the most important part of the license plate recognition system. Ability to correctly. Localization of License Plate Using Morphological Operations ` V.Karthikeyan and V.J.Vijayalakshmi Abstract-It is believed that there are currently millions of.
I want to develop an matlab program that can recognise the vehicle plate.
But before it, i have to do localization of the plate in the 'CAR' image.
HOW TO DO IT?
I have read many paper that said methods like modified Hough Transform, Bottom hat, vertical smearing, Gabor Filtering.
I Have 2 example codes, but the result image is wrong.
Here is the my First code but the result is wrong.
And here is my Second Code and the result is wrong too..
Here is examples of my images and this
Please help me. Any suggestionn ??
Farhat MannFarhat Mann
19311 gold badge44 silver badges1212 bronze badges
2 Answers
The Viola-Jones algorithm based on Haar-features is widely used for object detection. It is especially popular for face detection.
You can find an overview about the algorithm on wikipedia and get the original 2001 Viola and Jones paper here: (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.10.6807)
It is implemented in Matlab's computer vision system toolbox, more specifically the
vision.CascadeObjectDetector System object does the job.
Sample code can be found here: http://www.mathworks.com/matlabcentral/fileexchange/29437-viola-jones-object-detection
So far, so good. But here is the catch: afaik, Matlab only provides classifiers for faces, eyes, upper body, and so on, but not for license plates. Thus, you need to train the algorithm yourself. Therefore, you will need annotated training samples. These are image for which the license plate location is known. The
trainCascadeObjectDetector function does that training.
Sebastian StegerSebastian Steger
Matlab Code Examples
While I agree with @Sebastian's approach. I do believe that the Viola-Jones algorithms doesn't perform well with affine and projective distortions to the image. This is true for faces and should be the same for any other objects.
If the system you are trying to implement can ensure that the number plate is always facing straight at the camera most of the time. This would work well.
Another approach that might be worth exploring is basic text detection. There are multiple algorithms for the same. Since we can safely assume that number plates always contain text on them. You can use the MSER feature algorithm to find this text . You can find the MATLAB implementation here. . Once Text is detected. Using spacial locality it should be easy enough to extract the location of the number plate within the image. The only case where this would probably fail is when there are damaged number plates.
Midhun HarikumarMidhun Harikumar
Not the answer you're looking for? Browse other questions tagged matlabimage-processinglocalizationimage-recognition or ask your own question.
Active3 years, 6 months ago
I want to develop an matlab program that can recognise the vehicle plate. Lexmark x5470 driver for windows 8.
But before it, i have to do localization of the plate in the 'CAR' image.
HOW TO DO IT?
I have read many paper that said methods like modified Hough Transform, Bottom hat, vertical smearing, Gabor Filtering.
I Have 2 example codes, but the result image is wrong.
Here is the my First code but the result is wrong.
And here is my Second Code and the result is wrong too.. Microsoft rdp 6.0 download windows 7.
Here is examples of my images and this
Please help me. Any suggestionn ??
Farhat MannFarhat Mann
19311 gold badge44 silver badges1212 bronze badges
2 Answers
The Viola-Jones algorithm based on Haar-features is widely used for object detection. It is especially popular for face detection.
You can find an overview about the algorithm on wikipedia and get the original 2001 Viola and Jones paper here: (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.10.6807)
It is implemented in Matlab's computer vision system toolbox, more specifically the
vision.CascadeObjectDetector System object does the job.
![]()
Sample code can be found here: http://www.mathworks.com/matlabcentral/fileexchange/29437-viola-jones-object-detection
So far, so good. But here is the catch: afaik, Matlab only provides classifiers for faces, eyes, upper body, and so on, but not for license plates. Thus, you need to train the algorithm yourself. Therefore, you will need annotated training samples. These are image for which the license plate location is known. The
trainCascadeObjectDetector function does that training.
Sebastian StegerSebastian Steger
While I agree with @Sebastian's approach. I do believe that the Viola-Jones algorithms doesn't perform well with affine and projective distortions to the image. This is true for faces and should be the same for any other objects.
If the system you are trying to implement can ensure that the number plate is always facing straight at the camera most of the time. This would work well.
Another approach that might be worth exploring is basic text detection. There are multiple algorithms for the same. Since we can safely assume that number plates always contain text on them. You can use the MSER feature algorithm to find this text . You can find the MATLAB implementation here. . Once Text is detected. Using spacial locality it should be easy enough to extract the location of the number plate within the image. The only case where this would probably fail is when there are damaged number plates.
License Plate Localization Matlab Codes For Fske
Midhun HarikumarMidhun Harikumar
Not the answer you're looking for? Browse other questions tagged matlabimage-processinglocalizationimage-recognition or ask your own question.Comments are closed.
|
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |