Need help in collision detection

Question

0.00/5 (No votes)

See more:

i build a maze. it has walls of brown colour and floor is of white colour. player is a rectangular box controlled by arrows keys. on pressing arrow keys i want to first check tht the area my object will occupy is a floor or not, if it is a floor i allow it to move otherwise not.
i check this by
1. reading pixels of the next occupied area using glReadPixels
2. compare the readed pixels with white color, if it is white increment in the object direction otherwise not

the problem is this when i press left key, the object does not move even if it is on the floor

some of my global variables

float objectx, objecty, objectSize = 20;
RGBApixmap block(10,20);
bool isFloorColor = false;

objectx and objecty is initialzed in generateMaze()function which is in main

this is my object code

C#

void Object(float objX, float objY)
{
    glColor3f(64/255.0,0,64/255.0);
    glRectf(objX,objY,objX+objectSize,objY+objectSize);
}

arrow keys handler(i just mention left arrow key)

C#

void mySpecialKeyFn(int key, int mx, int my)
{
    switch(key)
    {

        case GLUT_KEY_LEFT:
            glReadPixels(objectx-10, objecty, block.nCols, block.nRows, GL_RGBA, GL_UNSIGNED_BYTE,block.getPixmap());
            isFloorColor = block.checkfloor();
            if(objectx-10>95 && isFloorColor==true)
                objectx-=10;
            glutPostRedisplay();
            break;
}

C#

void display(void)
{
....
           Object(objectx,objecty);
           glFlush();
}

function in headerfile RGBA.h

C#

bool checkfloor()
{
           int count = 0;
           for(int i =0 ; i<10; i++)
           {
               if(this->pixel[i].r ==(uchar)255.0 && this->pixel[i].g ==(uchar)255.0 && this->pixel[i].b ==(uchar)255.0)
                   count++;
           }
           if(count>=5)
               return true;
           else
               return false;
}

Posted 3-Oct-11 7:30am

Sweety Khan

Add a Solution

Comments

Sweety Khan 3-Oct-11 13:31pm

if i wrote any thing wrong n unclear, plz ask me inspite of closing my question

Manfred Rudolf Bihy 3-Oct-11 13:52pm

Repost! Deleting a question which was already answered is really bad style. I hope this will not be repeated by you in the future. I"ll monitor this question.

Sweety Khan 3-Oct-11 13:55pm

no one answered it. n secondly i didnt delete it was closed by Chris Maunder.

enhzflep 4-Oct-11 3:15am

To be fair - it was a rather moronic sounding post!

You asked how to do collision detection with the walls, while at the same time stating that you didn't know where the walls were! We're not mind-readers you know.

If you drew the maze, you know where all the walls are, simple as that. If you didn't draw the maze, then how on earth do you expect to be able to do collision detection with it?

An analogy would be drawing a string to the screen - if you've displayed the string in the first place, then of course you know what it says and where it's positioned. There's absolutely no need to perform Text Recognition on an image of the screen.

Now, if you've failed to maintain a variable with the position and another with the string, then that's an oversight on your part. However, you DO know what and where the text is already. The same thing applies for the maze walls...

Mohibur Rashid 3-Oct-11 23:42pm

I am no graphics programmer or designer but I have a question,
Do you think it will be best idea to check by pixel??? why don't you create a object array that will also hold the position, by checking the position you can figure out where your object is. ...

Sweety Khan 10-Oct-11 10:43am

actually i use binary tree algorthm to build maze. its the property of algo tht it does not store positions, every time we access cell, it is independent of others. it only has a 2D array whose elements represent walls n paths in maze, i hope im little clear. thts why i m using this approach for detecting floor.

GPUToaster™ 4-Oct-11 5:51am

Instead of using opengl inbuilt function glREctf, use your own geometry. Try to do ray tracing. Checking with pixel wont help you, and even if you get some result, wont be accurate and your program will become slower. After all its all math.

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Stefan_Lang · Accepted Answer · 2011-10-05T01:54:00

Solution 1

Change

C++

if(this->pixel[i].r ==(uchar)255.0 && this->pixel[i].g ==(uchar)255.0 && this->pixel[i].b ==(uchar)255.0)

to

C++

if(this->pixel[i].r ==255 && this->pixel[i].g ==255 && this->pixel[i].b ==255)

Your mistake:
You used a C-style typecast on constants of an incompatible type (float, in this case). Note that a cast does not convert the binary representation, it just reinterprets it - which in this case produces garbage.

Best practices to follow:
1. NEVER use C-style type casts in a C++ program. Most of the time a cast is just a sign of using the wrong types - change your variable types accordingly!

2. Even if you do really need a cast, use the C++ casts such as dynamic_cast or reinterpret_cast. And use them appropriately: see http://www.cplusplus.com/doc/tutorial/typecasting/[^]

Besides, it is inefficient and error prone to check your output and interpret it back into the model, when instead you could just check the model you used to produce that output.

Posted 5-Oct-11 1:54am

Stefan_Lang

Comments

Niklas L 6-Oct-11 7:30am

To OP: In addition, having the information in the model enables you to change your graphic background to something more interesting, like an image or similar, not having to rely on specific color information.

Sweety Khan 10-Oct-11 10:34am

sorry i didnt understand

Stefan_Lang 10-Oct-11 11:12am

The point Niklas was trying to make is that you should not look at the image, you should at the data instead that was used to generate it! Or more to the point, the positions of the walls. That way you are free to paint your walls over a random image (the background image) and do not need to care what color each pixel has. E. g. you could have an image that looks like some rough, rocky floor, or an earthen floor with roots springing up and patches of grass and moss everywhere, and then you paint your walls on top of that.

Sweety Khan 10-Oct-11 10:34am

oki. one of my mistake is objects x n y are left bottom pt and glread pixels need top left so i corrected it but still the same problem. to observe whts going on i print pixels which are read by glreadpixels and found tht it is reading too top from the point i gave. so i decrement y as required. it works fine in the first line of maze n detect floor well but below tht line it again reads from some where else :(

Stefan_Lang 10-Oct-11 11:07am

Have you fixed the error I pointed out above? An uchar is NOT a float, and you can NOT cast from float to uchar and expect meaningful values! 255.0 is a float! 255 is not.

But anyway, forget about checking pixels. It is complicated, inefficient and error prone, and will cause you more headaches than it's worth. You mentioned a function generateMaze() in your posting, so you do have the information where the walls are being drawn right in your program! When generateMaze() (randomly?) plazes the walls, store these positions! Then you can use that information to do your collision detection. You just have to iterate over all walls and see if they intersect with your object. Easy, and doesn't require any OpenGL calls at all!

Sweety Khan 14-Oct-11 1:59am

its dOne. thanx for forcing me to do collision detection by storing positons :)

Stefan_Lang 14-Oct-11 4:27am

Grats, and you're welcome *swings whip for emphasis* ;-)

Sweety Khan 10-Oct-11 12:50pm

yeh checking pixels has waste my time and my marks :(
the way u r saying seems vry difficult to me but now i know there is no other way. i will try it