I am a Lincoln cent collector and for several years have been using TurboCad for variety attribution. It certainly does speed up the process - but returns a false positive or negative on about 1 out of every 14 coins. Does anyone else use a CAD system as a attribution tool?
Hi, Im not sure everyone knows exactly what Turbocad is, and how it is incorporated into 'attribution' I can say I have used Autocad before, but I'm struggling with how warehouse design and layout has to do with coins. Can you elaborate?
Same here. I have used AutoCAD but I don't understand what you are doing with TurboCAD to attribute coins. Please elaborate.
Me too. One place I worked had several CAD types. AutoCad and Integraph were the most used. I think somewhere along the line AutoCad purchased most other sytems. These are all used for Engineering type projects though. How do you use it for coins? Can't even imagine.
It is a simple concept but somewhat complicated to explain. 1. Import a coin photo into TurboCad 2. establish a (x) and (y) axis based on established known benchmark from fixed devices in the coin's design 3. plot the (xy) coordinates of the mint mark 4. Compare the coordinates to a database of known attributed varieties. Sorta like a triangulation or GPS system for Lincoln cent varieties. I'll explain in greater detail if you would like.
Thats sounds like a really interesting and novel idea, if i got it right. Correct me if I got the jist of it wrong. Use a starting point associated with the coin's specific attribute specific to its variety, (like canadas' 1926 Far 6-use the top of the 6), and plot horizonal or verticle lines to another point, like a denticle or such, in order to accuratly determine if it is an authentic variety. Its kind of like the same method surveyors would use. If I've nailed it, i would wonder if the professional coin grader would use a tool of this nature. Ya know, it would be nice if they did, and they could return a soft or hard copy of the plotted information, with the grade coin. They do indicate the variety. Im sure alot of people agree, that some varieity of coins are very hard to determine even when using a strong loupe. This may make it much more cut and dry. Now, Id love to see this process taken one step further, and have a standard software/scanning program that can accuratley grade coins. No more human intervention.
Are you importing the photo on a 1:1 scale? Or, how do you know what the scale of the photo is? Do you caliper each coin, and then scale it up or down once you've imported the photo? This is very interesting, I may try this in AutoCAD, if I can figure out the scaling.
Scaling is not a problem - I take all my photos through my stereoscope using the same digital camera setting. As far as using a CAD system with pictures of different sizes - theorectically - couldn't you measure angles instead of distances - angles shouldn't change with respect to size of photo.
Excellent point! I'm going to think out loud, I could draw a template for each coin say a box around the coin, then draw lines to or through certain attributes on the coin, and scale the template up and down based on the perimeter of the coin so long as the box touched, say, four nodes on the perimeter, so that size of the coin wouldn't matter.
i am an avid ADt user, and never ever thought of using it for this purpose... i give you the most original idea award for the day Very original... and if it works, as you say..... very cool!
Jon Lusk has been doing this for years (over ten years) with his Numistudy software. It's only used for half cents and large cents right now but Conder tokens are being added. It autoscales your images to the same size as the ones in it's reference files, you can specify the same points on the subject and reference coins and it can compare lengths or it can compare it to each variety to give a list of possibilities. You can specify three points and it can compare angles and do overlays so you can see how the angles compare. You can also do fade in and out on the overlays to compare positioning. And alhough it is only made for those series you can build your own reference files so you could adapt it to other series as well. Oh and since errors in the placement of the points could result in mistakes you select an area on the image and it blows it up to full screen so you can place the point very precisely. Makes it very easy to get the points as closely as possible to the exact same place on each coin. The downside, although the software is reasonable, just a couple hundred dollars, if you want the data base reference files you are looking at $2K to $6K depending on which or how many of them you want. The reason the reference files are so expensive is because they are professionally done images of the condition census coins for every variety. Not just the CC1 coin but the top six to ten coins
Ineresting. A company called Hubble had a lighting program where you could see things in 3D. You would import all the data for, say a restaurant, then input all statistics and POOF, you would have a walkthrough of that restaurant with the lighting as detailed. A similar program is used by Architects for homes and other buildings. I wonder how those would work for coins. You would input all data and end up with a 3D, turnable, enlargable, coin.
Last post should have started with INTERESTING. Left out the T. Also, just contacted a person that has that Hubble program called CALA. I asked him if it was possible to enter coins into that program. I think he is still laughing. Got to go see him.
where i work, we have looked into Leica for 3d scanning, but so far, it hasnt been to our liking. We are going to wait for the tech to improve to where we would use it more to our advantage rather then only as eye candy. Soon, they may be able to incooperate BIM into it as well .. but who knows really
I have posted a picture for illustration purposes along with a short explanation of my process for using CAD as a variety attribution tool. The picture may be hard to see – when formatting a TCW into a JPG file the image tends to degrade. Step 1 - Establish the Y axis with a straight line along the west side of the numeral 1 of the coin’s date Step 2 - Establish the X axis by pulling a perpendicular line off the Y axis at the first tangent point on the south side of the last numeral of the coin’s date. Step 3 - Establish Line A by pulling a perpendicular line off the Y axis at the first tangent point on the North side of the mint mark Step 4 - Establish Line B by pulling a perpendicular line off the Y axis at the first tangent point on the South side of the mint mark Step 5 - Establish Line C by pulling a perpendicular line off the X axis at the first tangent point on the west side of the mint mark Step 6 - Establish Line D by pulling a perpendicular line off the X axis at the first tangent point on the East side of the mint mark. Step 7 - Find the XY coordinates of the line intersections – points (AC, AD, BC, BD) Step 8 - I query my coin variety database for examples of the same date and mint mark with the same 4 coordinates based on a tolerance of +/- 1%
Sorry I don't however I have used MicroStation for about six years now as a highway designer. Never heard of TurboCad.
Thanks Farmer....visuals speak volumes. Thats exactly how i envisioned the whole process...well done. For those people that utilize inventory software, i would also save this pic within the dbase as a supplimentary record.
I purchased TurboCad off the shelf from one of the Big Box Electronics Retailers. If I remember correctly it cost around $50.00. Didn't know how to use CAD when I bought it and to be honest - still don't know much about it now.
