Go Back   EcoModder Forum > EcoModding > DIY / How-to
Register Now
 Register Now
 

Reply  Post New Thread
 
Submit Tools LinkBack Thread Tools
Old 12-30-2007, 06:35 AM   #1 (permalink)
EcoModding Apprentice
 
Join Date: Dec 2007
Location: Melbourne
Posts: 152

The Mighty Mira - '92 Daihatsu Mira
90 day: 61.32 mpg (US)
Thanks: 0
Thanked 7 Times in 7 Posts
DIY Frontal Area Measurement

Apologies if this is somewhere else, but this way is fast and very accurate.

1. Find yourself a time somewhere between morning and afternoon (with good light).
2. Measure how high the car is, roughly.
3. Take a picture from directly in front of the car, at a height a half of your car's height, at least about 10m away from your car.
4. Download the Free and Open Source Software program, "The GIMP". GIMP stands for GNU Image Manipulation Program. If you use Ubuntu, it is in the repositories (synaptic or add programs).
5. Open up the picture in GIMP. Notice the point on the car for maximum height. Go out and measure from the ground to this point with a tape measure, taking care to eliminate parallax error from your measurement. Note this. Then measure the distance from the outside of the right front wheel to the left front wheel. Go back inside and to the GIMP again.
6. You will want to select around the car, so that nothing sticks out beyond your selection. Leave a little border around the car. Use the rectangle select.
7. Go to Edit -> Copy, then Edit -> Paste as New. This will have the car zoomed appropriately, if not, use the zoom commands in the View menu.
8. Save the new image.
9. Press "b" for "Create and edit paths". Or it's the upper right icon in the toolbox.
10. Select around the outline of the car as best you can, for curves, you will need more points than straighter sections. Try and be as accurate as possible. When you are done with the car and your dots are almost at the beginning, select "Create Selection From Path". You should get a nice little flashing outline of your car.
11. Go to Layer-> New Layer, and create a new layer with fill type transparency.
12. Use the bucket fill to fill the selection.
13. Go to Dialogues, Pixel Count. That is the number of pixels in your selection, which corresponds to the frontal area of your car in pixels. Note this figure. Now you just have to convert pixels to real area (m^2 or feet^2).
14. Measure the height in pixels of the highest point of the car, to the ground, using the measure tool (looks like a drawing compass in the toolbox). Or Tools -> Measure.
15. Measure the width in pixels between the outsides of the two front wheels (same span measured in 5.)
15. Once you have this, get height/pixel = Height obtained in step 5/ Pixels measured in Step 14. Should be a small number, much less than 1.
16. Get width/pixel = width obtained in step 5/ Pixels measured in Step 15.
17. Multiply the two number obtained in step 15 and 16. Should be a smaller number yet, even closer to zero. This gives you area/pixels.
17. Multiply the number obtained in 17 by the pixel count obtained in 13. This will give you the frontal area of your car in whatever units you were using.

Now that you have a very, very accurate frontal area of the car, you can compute your drag coefficient more accurately, once you have CdA (which the coast down test will give you).

__________________
"Every body perseveres in its state of being at rest or of moving uniformly straight forward, except insofar as it is compelled to change its state by force impressed." - Isaac Newton
  Reply With Quote
Alt Today
Popular topics

Other popular topics in this forum...

   
Old 11-30-2008, 06:31 PM   #2 (permalink)
EcoModding Lurker
 
fanamingo's Avatar
 
Join Date: Jan 2008
Location: Virginia
Posts: 46
Thanks: 0
Thanked 1 Time in 1 Post
Bump for a useful thread I just stumbled across in my searches. A couple suggestions for anyone else trying this:
  • Instead of taking a picture of your car (including camera distortion), you may find more accurate renderings at The-Blueprints.com - 26874 visits today.
  • Many manufacturers publish height and width dimensions. Keep in mind that width dimensions may not include the side mirrors.
  • If you can find a simple black and white linedrawing for your car like I did, you can use the Select by Color Tool for a quicker way to determine your car's outline.
  • In step 13, I went to Dialogs -> Histogram to find the pixel count.
I calculate the frontal area of my 8th gen Civic to be 22.8 square feet. I've been trying to find that number for a while now. Thanks, newtonsfirstlaw!
__________________
  Reply With Quote
Old 12-01-2008, 12:18 PM   #3 (permalink)
EcoModding Apprentice
 
dichotomous's Avatar
 
Join Date: Nov 2008
Location: vermont
Posts: 142

CivySi - '03 honda civic SI
90 day: 31.99 mpg (US)
Thanks: 0
Thanked 1 Time in 1 Post
this technique works better in CAD software. I can pull in a photo, scale it to real world dimensions, trace the exact outline and it will tell me the frontal area in whatever area units I feel like..

I might be inspired to do this for someone with a good picture and measurements, now and then...
__________________
  Reply With Quote
Old 12-01-2008, 01:24 PM   #4 (permalink)
Batman Junior
 
MetroMPG's Avatar
 
Join Date: Nov 2007
Location: 1000 Islands, Ontario, Canada
Posts: 22,528

Blackfly - '98 Geo Metro
Team Metro
Last 3: 70.09 mpg (US)

MPGiata - '90 Mazda Miata
90 day: 54.46 mpg (US)

Even Fancier Metro - '14 Mitsubishi Mirage top spec
90 day: 70.75 mpg (US)

Appliance car Mirage - '14 Mitsubishi Mirage ES (base)
90 day: 62.14 mpg (US)
Thanks: 4,078
Thanked 6,977 Times in 3,613 Posts
As long as you know one actual dimension of any part of the car shown in the photo/diagram, the math is pretty easy for calculating area with Photoshop.

See also: http://ecomodder.com/forum/showthrea...l-area-82.html
__________________
Project MPGiata! Mods for getting 50+ MPG from a 1990 Miata
Honda mods: Ecomodding my $800 Honda Fit 5-speed beater
Mitsu mods: 70 MPG in my ecomodded, dirt cheap, 3-cylinder Mirage.
Ecodriving test: Manual vs. automatic transmission MPG showdown



EcoModder
has launched a forum for the efficient new Mitsubishi Mirage
www.MetroMPG.com - fuel efficiency info for Geo Metro owners
www.ForkenSwift.com - electric car conversion on a beer budget
  Reply With Quote
Old 09-14-2013, 03:38 PM   #5 (permalink)
Master EcoModder
 
101Volts's Avatar
 
Join Date: Dec 2012
Location: Pennsylvania
Posts: 506

Woody - '90 Mercury Grand Marquis Wagon LS
Last 3: 19.57 mpg (US)

Brick - '99 Chevrolet K2500 Suburban LS
Last 3: 12.94 mpg (US)

M. C. - '01 Chevrolet Impala Base
90 day: 18.73 mpg (US)

R. J. - '05 Ford Explorer 4wd
90 day: 16.66 mpg (US)
Thanks: 936
Thanked 34 Times in 28 Posts
Quote:
Originally Posted by newtonsfirstlaw View Post
17. Multiply the two number obtained in step 15 and 16. Should be a smaller number yet, even closer to zero. This gives you area/pixels.
17. Multiply the number obtained in 17 by the pixel count obtained in 13. This will give you the frontal area of your car in whatever units you were using.

Now that you have a very, very accurate frontal area of the car, you can compute your drag coefficient more accurately, once you have CdA (which the coast down test will give you).
Perhaps fix the original post, #18 is marked as #17. Thanks for the guide.
  Reply With Quote
Old 11-22-2015, 01:46 PM   #6 (permalink)
EcoModding Lurker
 
Join Date: Nov 2015
Location: Berlin, Germany
Posts: 1
Thanks: 0
Thanked 0 Times in 0 Posts
Comparison between factory values and the photographer's method

Dear all,

As you all know, a precise frontal area is one key to a precise Cd determination. A precision of 0.01mē is desireable to reduce the uncertainty in Cd to around +/-0.005.

The following applies to Volkswagen Polo Bluemotion model year 2011:
For precise frontal area determination I decided to compare the factory given data with data obtained by the photography method descibed in this thread.

sources
drawings of VW:
volkswagen.de/content/medialib/vwd4/de/dialog/pdf/polo/polo_tup_110127/_jcr_content/renditions/rendition.file/polo_tup_110127.pdf
photograph:
taken with a 12x optical zoom camera (fully zoomed), distance of 160ft

data processing
GIMP Pixel counting method, pixels of drawings due to their high contrast can be easiliy counted using the "fuzzy select tool", the images are outlined manually using the "paths tool", only the left half image was outlined assuming the cars front face is symmetrical
scales were directly taken from drawings
wheel-to-wheel distance measured with folding ruler (uncertainty +/-1cm) for the image scale

results
ecomodder.com/forum/member-turbodiesel-albums-frontal+area-picture7105-polo-bluemotion-aus-prospekt.png
ecomodder.com/forum/member-turbodiesel-albums-frontal+area-picture7103-bild-polo-rotiert.png
ecomodder.com/forum/member-turbodiesel-albums-frontal+area-picture7104-bild-polo-rotiert-overlay-zeichnung.png

The first picture shows the drawings of VW. Note that height and width are not scaled 1:1, which had to be considered when calculating the drag area.
Interestingly the frontal area of the front view differs from the back view, the ground clearance is larger in the back view.
The Pixel-counted frontal area (front view) including the thick surrounding red outline is 2.15 mē, excluding the red line it's only 2.04mē.

The second image was rotated to align the car horizontally.
The third picture is an overlay of pic 1 and 2.
First sight suggests a good agreement of both pictures. Nonetheless, some significant differences remain:
a) Obviously the red line needs to be excluded from frontal area.
b) The underbody of the car is in the photo somewhat lower and contibutes frontal area that is not included in the drawing.
c) The tyres of the drawing are wider than stock as seen in the right side of the picture.
d) The photograph is not perfectly aligned as seen in the not colinear tyres of left side of the picture. Also the roof of the photo is lower than the drawing, which indicates that the distance the photo was taken is still to low as explained elsewhere in this forum.

The results of the photo evaluation are 1.97mē +/- 0.02mē. The error is the pure scale error due to length measurement and does not include any errors in image acquisition.

conclusion
The drawing's frontal area is 2.04mē what reduces by another 0.01mē when the tyre width is adjusted to meet the tyre width of the image.
The frontal area with the photographer's method most likely underestimates the true frontal area.
As a weighted mean value between these two areas a value of 2.02mē is suggested.

Some additional results from photo pixel counting:
Both tyres contribute a frontal area of 0.07mē. This could be reduced to almost zero by lowering the ride height of the car.
Both mirrors contribute a frontal area of 0.04mē.

Happy modding!
turbodiesel


Last edited by turbodiesel; 11-22-2015 at 01:55 PM..
  Reply With Quote
Reply  Post New Thread




Similar Threads
Thread Thread Starter Forum Replies Last Post
Different ways to measure/calculate frontal area (A) MetroMPG Aerodynamics 27 02-08-2015 02:00 AM



Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Content Relevant URLs by vBSEO 3.5.2
All content copyright EcoModder.com