VIDEO LINK: https://youtu.be/DwdHsuaTzb4
•The project idea is to implement an image processing algorithm that aims to detect the vehicle plate’s type and its governorate in a given image and detect the characters and numbers in it , It also detects the car’s owner information ( Name and Age ) and it detects the car’s model and traffic violations .
GENERAL IDEA: The idea of the project: is to take the image to plate enhancement function which does all the pre-processing then split plate takes it to divide the image into two parts letters and characters to give it to nums and chars function to get the number of chars and numbers to give it to mohfza function to know the name of Governorate then function split splits the chars and numbers images into each number and each letter alone in a picture so function OCR takes it to get the numbers and letters in parallel to that function plate types is working to know the color of the plate to know the type of the plate then connects the letters and numbers with the database to get the info we needed for this car we access all of this through GUI you can check source file to know the type of each plate and it's Governorate
⚫ We convert the input image to grayscale image then black&white to apply Image processing techniques to it.
⚫ We extract label image using CONNECTED COMPONENT then color its objects using mycolor matrix.
⚫ We remove all objects except numbers and characters.
⚫ We convert the processed image to Black&White again.
⚫ We remove noise pixels(undesired small objects) using EROSION, we use this image later to recognize the characters and digits.
⚫ We make every arabic letter with dots to be one single object using DILATION to use this image later to extract the real number of characters and digits in numANDchar function.
⚫ This function takes the processed image after dilation then splits it into Characters image and numbers image then return the number of objects in each image using connected component.
⚫ This function takes the plate image then resize it and crop the image to get the upper part and apply some functions to get the pixels of the image and its values then get the mean of the color and check its range within which range to know the plate type.
⚫ Each Governorate has its own plate so we check the number of letters and numbers and also check the first one or two letters to know the type of each one.
⚫ This function is implemented to extract numbers and characters easily from pltae, as it takes the output image from plate_enhancement function and split the plate into two parts, first part is the characters in the plate and the second part is the numbers.
⚫ This is our main function, where all the other functions are being called from. The main function takes the path of the image, reads it, send it to the plate enhancement function, then the output of the plate enhancemet fuction is given to split plate function to split the plate into numbers and characters, then we pass all the info to our functions to finally get the output (GovName, Plate Colour, etc..)
⚫ This is our GUI function, where we call the main function there, and with the use of browse button we can specify the image needed and its path is given to a text box then it's given to the main function to start doing its function, then we take the output of the main function and using some labels we show the output on the screen for the user. Also there is a restart button where the user can restart the application without having to exit and re-open the application.
1- Browse to choose the image
2- press submit , the info will appear
3- you can exit if you don't want to browse another plate or you can press refresh to browse another plate