My Exploration of Building a Coin-Sorting Machine by Date

As a software engineer and lifelong coin collector, I’ve been tinkering with an idea: building a machine that sorts coins by date and mint mark. It’s a project born from my twin passions, one that could transform how we hunt for rare finds – though I’ve discovered plenty of hurdles along the way. Let me share what I’ve learned through this journey and whether it’s something that might help fellow collectors.

Technical Challenges and Approaches

From my workbench experiments, I found the whole system hinges on image recognition and database matching. I started by creating a solid database with multiple photos for each coin side – crucial because a worn coin looks completely different from a fresh one. To handle different orientations, I stored at least 360 rotated images per coin type, using edge detection to compare against known templates. When a coin doesn’t match within tolerance, it routes to a reject bin; matches get processed with OCR to read dates and mint marks. The toughest parts?

• Wear variations: A pristine coin won’t match one that’s seen decades of pocket wear, requiring dozens of reference images per condition
• Handling unknowns: The system needs smart sorting into accept/reject/unknown bins – simpler for users but complex to program
• Starting small: Working with a single denomination first prevents overwhelming the hardware

While today’s tech makes this possible, the coding demands are intense and high-speed cameras don’t come cheap – we’re talking thousands of dollars to build.

Financial Feasibility and Market Potential

When I ran the numbers, the economics got real sobering real fast. For personal use, you’d need years of coin roll hunting just to break even after counting gas and coin costs. Banks and services like Brinks typically won’t work with individuals, and they couldn’t care less about sorted dates. Selling the machines could be an option, but let’s be honest – it’s a niche market. Some dedicated collectors might pay up to $5,000 for a unit that spots errors or VAMs, but I don’t see it becoming mainstream. Why?

• Cost hurdles: The hardware and software expenses put it out of reach for most hobbyists
• Profit reality: Treating this as a passion project makes more sense than a business – patents and legal fees add up fast
• Market risks: Pre-orders sound nice but feel risky without a working prototype

While I still love the concept, it’s clearly more about the journey than any profit potential.

Practical Advice for Coin Collectors

If this sparks your curiosity, my advice is simple: start small. Focus on specific series like Morgan dollars where date and mint mark detection can reveal hidden treasures. For everyday hunting, such a machine could speed up finding pre-1982 coppers for melt value or spotting key dates. But keep these realities in mind:

• Grading limits: Automation helps flag coins needing closer inspection, but it can’t replace human judgment on wear or errors
• Budget-friendly options: Partner with local banks for bulk coins or stick with manual sorting – your wallet will thank you
• Market savvy: Watch metal prices – when copper spikes, sorting cents becomes more worthwhile

This whole adventure reminds me why we collect: build for the love of discovery, not just the bottom line.