Welcome, Guest
Username: Password: Remember me

TOPIC: Geometry and Kinematics of Guided-Rod Sharpeners

Geometry and Kinematics of Guided-Rod Sharpeners 6 months 3 days ago #15643

  • AnthonyYan
  • AnthonyYan's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 100
  • Thank you received: 64
  • Karma: 16
Thanks, mark76!

Right now, there is no user interface; it's all command line driven. The current version is written in Matlab mostly because Matlab has tons of math and graphics routines. Matlab's contour and graphics commands were used to create each frame of the animation. Each frame was then saved as a PDF file and later converted to PNG. Apple's QuickTime Pro was then used to assemble the PNG files into a .MP4 video.
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15811

  • AnthonyYan
  • AnthonyYan's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 100
  • Thank you received: 64
  • Karma: 16
Here is a new version. This will be the last version for quite awhile, I think.

Changes:
Expanded some of the discussion about gimbals, universal joints, and spherical joints in Chapter 3.
Re-compressed the animated contour plots with better (?) anti-aliasing settings.
Minor formatting improvements.

Geometry and Kinematics of Guided-Rod Sharpeners
Version 1.0beta17

Download:
drive.google.com/file/d/0B8rQYhU8N9ZGSENqc2Q2MlRFbTA/

Alternate Download:
www.mediafire.com/download/2flrqn7po9um3...DRAFT_1.0beta17).zip
The administrator has disabled public write access.
The following user(s) said Thank You: wickededge, mark76, Mikedoh

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15846

  • AnthonyYan
  • AnthonyYan's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 100
  • Thank you received: 64
  • Karma: 16
wickededge wrote:
Lagrangian,
That is fantastic! Your explanation works very well for and then to see it represented so well visually really sends the concept home. Of course I'm now thinking of how we might calculate the correct position using this method for lots of knives...

Hi Clay, everyone,

I'm also thinking about how to apply this to many knives as possible. I would like, eventually, to make a program that people could use. But for now, it might be helpful if you and/or users could submit photos of their knives which follow the suggestions below. With a small collection of such photos, we could learn quite a bit from test runs.

I would suggest that photos of knives be created by either a digital camera, or a flatbed scanner.

Digital Camera:
(1) Camera should be pointed perpendicular to the plane of the knife and centered on the knife.
(2) The camera should be zoomed in as much as reasonably possible. Longer focal length reduces perspective distortion.
(3) The camera should be physically as far away as reasonably possible from the knife. Increased distance reduces perspective distortion.
(4) Graph paper should be placed behind the knife and as close to the knife as reasonably possible. This will allow some of the residual distortion to be detected and then removed by tweaking in Photoshop, etc.
(5) To show the size of the knife accurately, place a ruler next to the knife. With inch and/or centimeter marks, we can scale the image accordingly.
(6) Please identify the knife. Next to the knife, say on a Post-It note, write the knife manufacturer, model, and year (if known).

Flatbed Scanner:
Same as digital camera, except there is no camera placement of course. Like a digital camera, use graph paper as a background, include a ruler in the scan, and write the knife maker and model on the background. Remember to be careful and not to scratch the glass of your scanner with your knife edge, or bolts, or metal clips! It might be worth using some plastic film or a cloth to protect your flatbed scanner.

If any of us could post a few images as described, then we could collect maybe a dozen images. With that, we can start analyzing them and also start thinking about how to create a program for everyone to use. I'm very busy these days, but it is something I would like to do, even if slowly.

So please post some images! :)

Sincerely,
--Anthony "Lagrangian" Yan
The administrator has disabled public write access.
The following user(s) said Thank You: wickededge

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15847

  • wickededge
  • wickededge's Avatar
  • OFFLINE
  • Administrator
  • Posts: 1762
  • Thank you received: 610
  • Karma: 91
Will do!
--Clay Allison
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15849

  • AnthonyYan
  • AnthonyYan's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 100
  • Thank you received: 64
  • Karma: 16
Hi Clay, Everyone,

Hmm... I did an experiment, and made two photos of my Benchmade Griptillian. First, I used my tablet to take a photo as described above, but it didn't come out very well. Then I tried my flatbed scanner, and the image was much better. Discussion below.

Using my tablet, I took a photo. I used a carpenter's level to make sure the desk was level with respect to the ground. Then I put the carpenter's level on my tablet, to make sure the tablet was also level with respect to the ground. This would ensure that my tablet camera is pointed perpendicular to the plane of the knife. Unfortunately for me, my tablet's camera isn't very good. I'm not too happy with this image.


Next, I tried my Epson V33 scanner (a low end flatbed scanner). This worked much better. I placed a clear transparency between my knife and the glass top to prevent scratching. I have heard that CCD flatbed scanners tend to have a larger depth of focus, so they are better for scanning shallow 3d objects (like leaves, feathers, and for us, knives). The other common type of flatbed scanner is CIS (if I remember right), and usually as a much shallower depth of focus, so I don't know if it would scan knives well. But it is worth a try.


One problem I should have anticipated, but did not: The knife edge is silver, and I used a paper white background. That makes the edge rather difficult to see when zoomed into the photo. Instead, we should use something with more contrast, maybe a black background or something.

So I now recommend the following:
(1) Use a flatbed scanner. The image quality is much higher, plus there is no image distortion from the camera or 3d perspective (so long as the knife will lay flat and parallel to the scanning bed).
(2) With a flatbed scanner, keep the ruler in the image. But graph-paper is unnecessary. Instead, choose a background which has high contrast with the knife edge.
(3) Scan the entire knife, including the handle. This will help us understand and visualize the results.
(4) I think scanning at 600 dpi is optimal for small knives. That sounds really high, but suppose we want to accurately trace a 3-inch blade on a 1080p monitor. At 1080p, we have 1920 pixels across the monitor to cover three inches in real life. So that is (1920 pixels)/(3 inches) = 640 dpi. Here is the blade only at 600 dpi (right-click to download or open the full sized image in a new tab).


Here is a test scan with a darker background (just carboard).


Just remember, some modern steels are very hard. (ZDP-189 can be super hard at 65 HRC.) I don't know if they are hard enough to scratch glass, but I say better safe than sorry.

P.S. If you are curious about the CCD vs CIS difference in flatbed scanners:
en.wikipedia.org/wiki/Contact_image_sensor
Last Edit: 5 months 3 weeks ago by AnthonyYan.
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15863

  • wickededge
  • wickededge's Avatar
  • OFFLINE
  • Administrator
  • Posts: 1762
  • Thank you received: 610
  • Karma: 91
I'm not having a lot of luck using the scanner, the images are coming out very washed out. I wonder if I'm missing something in the way I'm doing it or if my scanner just isn't up to it. I'll try with a camera instead.
--Clay Allison
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15865

  • mark76
  • mark76's Avatar
  • OFFLINE
  • Administrator
  • Posts: 1195
  • Thank you received: 361
  • Karma: 86
AnthonyYan wrote:
I would like, eventually, to make a program that people could use.

Hi Anthony,

I was thinking about the same. That is not to say I'll make a program for that tomorrow, but who knows... What library/software do you use to convert scanned images into line drawings? I suppose Mathlab doesn't have tracing features, or does it?
Last Edit: 5 months 3 weeks ago by mark76.
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15875

  • AnthonyYan
  • AnthonyYan's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 100
  • Thank you received: 64
  • Karma: 16
wickededge wrote:
I'm not having a lot of luck using the scanner, the images are coming out very washed out. I wonder if I'm missing something in the way I'm doing it or if my scanner just isn't up to it. I'll try with a camera instead.

Hi Clay,

Hmm... A couple of questions?
(1) Which flatbed scanner are you using?
(2) Actually, my flatbed scans are also very washed out. But you can try "auto levels" or "adjust curves" in Photoshop to make it much better (or other image processing software). Maybe try that?
(3) If you like, send me a private message with your scanned image, and I'll see if it is easily fixable or not.
Last Edit: 5 months 3 weeks ago by AnthonyYan.
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15876

  • AnthonyYan
  • AnthonyYan's Avatar
  • OFFLINE
  • Expert Boarder
  • Posts: 100
  • Thank you received: 64
  • Karma: 16
Hi mark76,

Oh, that's an interesting idea. I may try to see if some open-source computer vision library (such as OpenCV) can help us convert scanned images into line drawings.

For my current program, I had to write a small utility function to help me manually trace the photo of the knife
The administrator has disabled public write access.

Geometry and Kinematics of Guided-Rod Sharpeners 5 months 3 weeks ago #15879

  • tcmeyer
  • tcmeyer's Avatar
  • NOW ONLINE
  • Gold Boarder
  • Posts: 208
  • Thank you received: 144
  • Karma: 21
Here's the photo (scan) of your Benchmade blade, which I cropped using Corel's Paintshop, and modified using their "edge effects" function. Does something like this help you get closer to what you need? The edge of the blade seems very clear. Maybe you could selectively delete everything but the actual edge.

I cropped out the ruler, but I think it could be included.


benchmadeblade.jpg
The administrator has disabled public write access.
The following user(s) said Thank You: AnthonyYan
Time to create page: 0.154 seconds