OpenMV aims to make machine vision as easy as Arduino made microcontrollers. They supply a small camera and microcontroller with MV software already loaded. At last, you can make projects that react to events in the real world just by seeing things happening.

What CAN it do? You could…

  • Tell a motor to open your blinds only after the camera sees you sit up in bed in the morning.
  • Control a pan/tilt unit to keep the camera pointed at the nearest face.
  • Sound an alarm if something looks like it’s moving too fast.

The H7 is the fourth generation of machine vision microcontroller OpenMV has released in as many years. That rapid iteration shows: the IDE feels polished. The documentation holds your hand with a quick but complete guided tour of how to get set up and touring what this little smart camera can do.

As I write this, OpenMV has more than a dozen machine vision routines, including recognizing faces and QR codes. Perhaps more importantly, they provide more than 100 sample programs, several for each of the different vision routines the camera knows.

Though the OpenMV hasn’t quite hit their goal of Arduino-level ease of use, the trade-offs they made were sensible. The module is sold without the headers attached, which is both good and bad. Upside: You can solder directly to it, you’re not stuck with only breadboarding. Downside: You can’t just get started with a breadboard and jumper wires. You’re going to have to break out a soldering iron, sooner or later. Before the first time you use the module, you’re also encouraged to unscrew the lens off the sensor and wipe both down with alcohol. And you’ll want to focus the lens manually. All these choices let the small-and-scrappy OpenMV team keep costs reasonable while their sales volume hasn’t yet taken off.

Of all the boards I’ve gotten to sample this year, this one is a new favorite. OpenMV has brought a previously difficult technology into reach for many. Their focus on ease-of-use and their copious example programs should let any maker get started in vision-activated gizmos.