convert pixels to centimeters

yes i know, still do you know how to do it?

take a look at this code, it's not useful but it'll give you a good idea of what can be done

http://www.delphipages.com/threads/thread.cfm?ID=115788&G=115786

or

http://delphi.about.com/od/adptips2005/qt/pixel2mm.htm (but the code is a mess and i can't seem to be able to fix it)

Amigo is right.

You would be able to calculate the position of a mouse using the CursorPos position I think to get the start and end points. This would give you the number of pixels travelling in the X and  axis, but you dont have a reference point due to screen resolution.

You would need to work from a standard resolution, and calculate a base pixel distance to cover a centimetre. You could then extrapolate for other resolutions from this base figure.

So say you worked on a 800x600 screen and you found that 1cm = 20 pixels on the X axis and 1cm = 15 pixels on the Y axis.

Therefore a 1200x1024 screen, X axis (1200/800)x20 = 30 pixels, Y axis (1024/600)x15 pixels.

I think my maths is correct.

I like that second example you showed. This is pretty much using the same maths calculation as I described above.

Rsolution / Size

Can you not get it to work?

seems like the example has a few typos. try this (havent tested tho)


procedure PixelsPerMM(
   canvas: TCanvas;
   var x, y: single) ;
var
    H:HDC;
    hres,vres,
    hsiz,vsiz:integer;
begin
    H:=canvas.handle;
    hres := GetDeviceCaps(H,HORZRES) ; {display width in pixels}
    vres := GetDeviceCaps(H,VERTRES) ; {display height in pixels}
    hsiz := GetDeviceCaps(H,HORZSIZE) ; {display width in mm}
    vsiz := GetDeviceCaps(H,VERTSIZE) ; {display height in mm}
    x := hres/hsiz;
    y := vres/vsiz;
end;
~~~~~~~~~~~~~~~~~~~~~~~~~

var
   cx, cy : single;
   mmx, mmy : integer;
begin
   PixelsPerMM(Handle,cx,cy) ;

   mmx := Trunc(468 / cx) ;
   mmy := Trunc(60 / cy) ;
end;

i have the same problem here, handle of what!?

> PixelsPerMM(Handle,cx,cy) ;

the error is Incompatible types: TCanvas  and HWND

well, try form1.canvas or canvas or the canvas of sth else ;-)

i'd a[[reciate it if you can show me how this work?
it doesn't seem to work, not for form1.canvas, not passing the desktop as canvas, etc. i ask my question again:

is there any way to convert pixels to cm or mm or m, or whatever and know how much the mouse traveled?

by the way if you want to know why i need this code visit my page at
www.geocities.com/uf_new
i'm writing a mouse application with ergonomics in mind, since i developed an RSI case. part of it is how much mouse movement a user produces on each session

Please note, converting to CM (or even to MM) should probably be done as double result, due to the resolution difference.

Eg;

1024 px screen width = 320 mm screen width, so it takes 3.2 pxls = 1 mm, or 1pxl = 0.3125 mm

Russell

ok, let me try it,
a question tho, why do you want to return a tpoint in function PixelPointToMM?

I was only trying to be complete; X coord px conversion, Y coord px conversion, and TPoint (which would be useful if you were doing an (X px, Y px) -> (X mm, Y mm) conversion. If you don't need it, then don't use it.

Russell

excuse my ignorance, i just don't see where the mm are in a tpoint...
i understand that the 1st 2 functions return integer and that they return mm, the last function returns a tpoint so i am having trouble understanding the whole idea behind it.

When dealing with the mouse, one of the most common structures is the TPoint, which is a structure to hold both the horiz/vert position. The X, Y values are integer data types, and the resolution is assumed to be in pixels.
All I did was provide you a helper function that takes both a horiz and vert pixel measurement and converts it to millimeter in one call. Eg:

var  ptCursorInMM:  TPoint;
begin

  // Convert mouse position in pixels to point structure in mm
  ptCursorInMM:=PixelPointToMM(Mouse.CursorPos);

  // ptCursorInMM.X is the horiz measurement now in MM
  // ptCursorInMM.Y is the vert measurement now in MM

end;

Or, for example, you are trying to convert the width/height of something in pixels -> mm.  Its nothing more than a conversion helper function, thats all. So I will say again, if you don't need it, then don't use it.

Russell

ok, every time the mouse move i call this function (quick and dirty):
posx and posy are integers that are updated every time the mouse move (by the mouse hook) so they are always updated with the latest position of the mouse.

procedure TForm1.Mousetravel;
var
 traveled, currpos:tpoint;
 total:integer;
begin
    currpos.X:=posx;
    currpos.Y:=posy;
    traveled:=PixelPointToMM(currpos);
    total:=traveled.X+traveled.Y;
form2.StaticText2.Caption:=inttostr(strtoint(form2.StaticText2.Caption)+total);
end;

even tho i move the mouse the values are being updated randomly and without any logic (seemingly).
i will give you the points, since the answer seems to work, but again what i need in the total (in mm, cm or whatever in metric) that the mouse traveled until it stops. so, any ideas?

thanks

this is how the function looks at the moment (to km)

procedure TForm1.Mousetravel;
var
 traveled, currpos:tpoint;
 total:integer;
 km:double;
begin
    currpos.X:=posx;
    currpos.Y:=posy;
    traveled:=PixelPointToMM(currpos);
    total:=traveled.X+traveled.Y;
    km:=(total*0.001)*0.001;
    form2.StaticText2.Caption:=FloatToStrF(StrToFloat
                               (form2.StaticText2.Caption)+
                               km, ffNumber, 8, 4);
end;

urif,

Two things stand out. First, I already mention this; you will find the resolution between pixels/mm to be such that a double value should be used. Eg, on my system

3.2 pxls = 1 mm, or 1pxl = 0.3125 mm

So moving one pixel would yield 0 mm in integer format. (It takes 3.2 pixels to yield 1 mm).

Second, your math for determining distance is incorrect. Take for example the following. You start at a position of (0,0) and then move to (1,1). Using your math, you will have calculated a move of 2 units.

total:=traveled.X+traveled.Y

This is incorrect, as the distance traveled is actually 1 unit.

You didn't mention this before, only that you required a conversion from pixels to mm. The conversion routines ARE correct, but again, should probably be scaled to double for greater accuracy. If you would like, I will provide these as well as the correct algo for distancing.

Regards.
Russell