AI insect detection camera
Close-up insect imaging with an embedded vision platform for local machine-learning inference.
The project
Combine macro imaging, optical alignment, and embedded inference for observing small moving subjects.
Product highlights
- IMX219 8 MP camera and MIPI-CSI image interface.
- RK3588-based processing and TensorFlow Lite in the documented inference scope.
- Low-distortion optics and mechanical alignment for close-up observation.
How we approached it
On-device AI projects like the AI insect detection camera succeed or fail on resource limits. We decide early which processing runs on the device and which, if any, runs elsewhere, then profile memory, timing and heat on the target hardware (RK3588, IMX219, MIPI-CSI).
Accuracy and speed are measured on the real implementation rather than assumed from processor specifications, and the interface is designed to stay responsive while models run. That keeps the product usable without depending on a cloud connection.
Technology at a glance
| Discipline | Edge AI |
|---|---|
| Key technology | RK3588, IMX219, MIPI-CSI |
| Type of work | Edge ai device |
Questions about this project
What technology did SixSol use for AI insect detection camera?
The project used RK3588, IMX219, MIPI-CSI. IMX219 8 MP camera and MIPI-CSI image interface.
Can SixSol build a similar edge AI device?
Yes. Share your goal, any existing hardware or code, and the constraints you know. We start with a scoped first milestone and a working build. Learn more about our embedded and edge AI service.
Inside the project.
Product, board, and interface visuals from the project portfolio.
