Add Custom Vision to Your Turtlebot

See what's new at Remyx.ai

Customizing AI models for your project can be difficult if you don't have a curated dataset or the compute/time to train a model from scratch. This is why we build remyx.ai - an easy to use AutoML service that doesn't require data to get started.

Using our engine, you can generate custom image classifiers after describing class labels of interest. In this demo, we chose a simple scene classifier which can support decisions for how a robot will move in different settings.

Using the turtlebot3, we could mount an OAK-FFC camera for vision. We found it best to step down the voltage coming from the OpenCR board to correctly supplement the power directly from the battery.

We also found using the kit’s controller easier to use than keyboard navigation while bypassing additional compute for the Pi by directly communicating to the OpenCR module instead.

The camera feed helps in teleoperating our bot, so we modified the rgb_publisher node from the depthai-ros-examples repo.

The next video shows our scene classifier inference as the turtlebot follows into the bedroom.

To see more of our turtlebot vision experiments, check out smellslike.ml and try out remyx.ai to power your next AI application with a custom image classifier!