You are on page 1of 13

PRESENTATION

USAMA ARQEM AI-C2-20 MUHAMMAD KAMRAN AI21-C2-15

AHMED MUSTAFA AI21-C2-10 RUMASSA ISHAQ AI21-C2-07

KALEEM ULLAH AI-C2-24 WAQAR ALI AI21-C2-07-21


PROJECT TITLE
VEHICLE NUMBER PLATE RECOGNITION
OBJECTIVE OF THE PROJECT

• Monitoring
• Security
• Surveillance
• Road safety
DATA COLLECTION

• In order to design the code for number plate recognition we considered following models
• OpenCV and Pytesseract pip3 python package
• Edge Detection based on Plate Extraction
DATA COLLECTION

• To detect number plates using python code, following areas were studied thoroughly
• Optical Character Recognition to convert images into machine coded text.
• Techniques to Extract Features from Image Data using Python.
• Segmentation techniques.
• In order to test the model in real time, techniques to separate frames from video were
studied.
BLOCK DIAGRAM
HOW IT WORKS

• License Plate Detection


• First step was to resize the image to the required size and then grayscale it.
• Blurring is done using a bilateral filter which removes the unwanted details from an image.
• Edge detection is done using canny edge method from OpenCV.
• The next step is looking for contours on image and sorting them from big to small.
• It will give the number plate area.
• The next step is masking the area apart from number plate.
CONT.…

• Character Segmentation
• The next step is to segment the license plate out of the image by cropping it and saving it
as a new image.
• Character Recognition
• We used the Pytesseract package to read characters from image.
• The program detected it and printed the same value on jupyter notebook.
ACCOMPLISHED TASK

• The program is able to detect the license plate correctly and crop it.
• The program provides good image quality and correct orientation.
• The program was tested on real time video and it was able to detect number plate up to a
distance of three feet.
LIMITATIONS

• It can detect license plate up to a distance of three feet.


• The accuracy of the depends on the clarity of image, orientation, light exposure etc.
• The Tesseract library has sometimes unable to recognize the characters properly.
METHODS TO DEAL WITH THE LIMITATIONS

• To get better results Machine learning algorithms can be implemented along with this.
• Problems related to Tesseract library can be corrected by either using better orientation of
images or by configuring the Tesseract engine.
REFERENCES

• https://medium.com/programming-fever/license-plate-recognition-using-opencv-python-7
611f85cdd6c
• https://aip.scitation.org/doi/pdf/10.1063/1.3526208
• https://techvidvan.com/tutorials/python-project-license-number-plate-recognition/
• https://medium.com/@balaajip/optical-character-recognition-99aba2dad314

You might also like