My Adventure in Designing a Coin Sorting Machine by Date

As a software engineer and coin collector, I’ve been playing with the idea of building a machine that automatically sorts coins by date and mint mark. It began with a simple question: could I blend my tech skills with my numismatic passion to create something useful? But before jumping in, I needed to consider the real-world costs and practicality. Here’s where my journey took me.

The Technical Blueprint: How It Would Work

I imagined a setup using cameras and image recognition to identify coins. The core approach? Build a database of coin images—both sides—with date and mint areas masked. The system would digitally rotate these reference images through all possible angles, using edge detection to match real coins. Close matches would trigger OCR to read dates and mint marks. Non-matches go to a reject bin, while questionable coins get a “needs human eyes” slot.

• You’d need tons of reference images per coin type to handle wear differences—an uncirculated beauty looks nothing like a worn-down workhorse.
• Accounting for wear is critical since a pristine coin won’t match its circulated cousin, so multiple condition examples are essential.
• I’d simplify sorting to three bins: accept for target dates/mints, reject for definite mismatches, and unknown for coins requiring manual inspection.

Financial Reality: Crunching the Numbers

When I ran the numbers, the economics got tricky fast. Prototyping would cost thousands due to high-speed cameras and complex software. Even buying coins in bulk through services like Brinks doesn’t make the math work. Most collectors only want pre-1982 copper pennies for their metal value, and banks rarely need rolls sorted by specific dates.

Trying to sell these machines? The market feels limited. While some fellow collectors mentioned paying up to $5,000, recouping costs could take years after factoring shipping, returns, and patents. Honestly? View this as a passion project like woodworking or model building. The real reward is the creation process itself, not profit.

Practical Tips for Fellow Tinkerers

If you’re tempted to try something similar, start modestly. Focus on one denomination or era first. Sorting Morgan dollars for VAM varieties might be more rewarding, but requires serious image analysis skills. Some hard-won advice from my exploration:

• Build error detection into your design—spotting doubled dies or off-center strikes could make your machine truly special.
• Handle wear variations by including multiple condition images, mirroring how we examine coins by hand.
• Pre-orders might fund development, but prove your concept with a working prototype before making promises.

Parting Thoughts for the Curious

Despite the hurdles, I still love the concept. A well-built sorter could revolutionize coin roll hunting, helping us spot key dates faster. It might even change what survives in circulation. I’m still deciding whether to build mine, but if I do, I’ll share the journey. If this idea excites you and you’ve got the skills? Dive in as a passion project. Our hobby grows through experiments like these—yours could be the breakthrough we’re all waiting for!