segmentise the BW image for car plate number.

hi all, Im having difficulties in segmentise the BW of car plate number. need help. I did run the blobdemo by imageanalysist but it doesnt work. this is my coding and the image of my project. i need to finish this project by the middle of may. thanks.

clc
clear all
close all
I=imread('149.jpg');
figure, subplot (2,2,1),imshow(I)
 Igray = rgb2gray(I);
 subplot (2,2,2), imshow(Igray)
start_row = 450;
start_col = 900;
cropIbw = Igray(start_row:500, start_col:990, :);
subplot (2,2,3),imshow(cropIbw)
Ibw = im2bw(cropIbw,graythresh(cropIbw));
% Ibw = im2bw(Ibw,0.4);
subplot (2,2,4),imshow(Ibw)

% offsetX = start_col-1;
% offsetY = start_row-1;

Answer 1

[Image Analyst]

BlobsDemo does work. You must mean that you didn't adapt it properly for your images. (If that's not the case, let me know.) For example, there is a low probability that graythresh() will select the best threshold for your image. But you really don't say why you're not happy with the code you provided. All you're doing is converting to grayscale, cropping, and thresholding. But you're not asking a question so we don't know what else to say.

Answer 2

[UMP]

Hello image analyst,

Im not saying that im not happy with my own coding. I do happy with it. however, im having difficulties to do the segmentation to the cropped BW image. In the cropped image it has the car plate number. So i want to segementise it before proceed to the template matching or ANN to produce the recognition step. so this my question, how to do the segmentation to the alphanumerical value containt in the cropped BW image. I thought u understood before. seems u didnt understood, so i ask it clearly now. the blobdemo coding didnt help me much because i didnt know how to suit it with my data. its too long tho. and im not expert in matlab by the way.