How to count the vehicles of non-uniform speed in a video without using computer vision toolbox?

    clc
clear all
n=0;
% Read movie 
SF = 1; %Starting Frame
MV = mmreader('a.avi'); %To read Movie
EF = MV.NumberofFrames ; %Ending Frame
img = read(MV,1); %reading particular frame 
A = double(rgb2gray(img)); %motion vector works on Gray Image formate
[height width] = size(A); %Movie size?
% motion estimation 
 h1 = figure(1);
for f = SF: EF %Frame from SF To EF
    B = A; 
    img = read(MV,f);
    A = double(rgb2gray(img)); 
    %Foreground Detection
    thresh=11;
    fr_diff = abs(A-B);
    for j = 1:width
        for k = 1:height
        if (fr_diff(k,j)>thresh)
            fg(k,j) = A(k,j);
        else
                    fg(k,j) = 0;
        end
        end
    end
    subplot(2,2,1) , imagesc(img), title (['Orignal Sequence Frame#',int2str(f)]);
    subplot(2,2,2) , imshow(mat2gray(A)), title ('Previous Frame');
    subplot(2,2,3) , imshow(mat2gray(B)), title ('Next Frame');
    sd=imadjust(fg);
    level=graythresh(sd);
    m=imnoise(sd,'gaussian',0,0.025);
    k=wiener2(m,[5,5]);
    bw=im2bw(k,level);
    [labeled,numObjects] = bwlabel(bw,4);
    n=n+numObjects;
    subplot(2,2,4) , imagesc(bw), title (['Foreground #',n]);
    hold off;
    pause (.1);
% saveas(h1,strcat('Result-Paris-',num2str(f)),'jpg');
end
disp(n);

The above is the code,which converts the frames into binary images and find the frame difference to detect the foreground objects. But it doesnt compute the correct count because I dont know to fix the correct interval between frames. Kindly help me in completing it.