Raspberry Pi just updated its camera add-on and HAT kits
Raspberry Pi has just announced software updates for its growing suite of AI products. This comes with full AI HAT+ support on the Trixie version of the Raspberry Pi OS and a brand new debug feature for the AI camera.
The Raspberry Pi AI HAT+ and corresponding AI kit, both of which rely on Hailo AI accelerators, are now fully compatible with the recently launched Trixie version of the Raspberry Pi OS. You can get all the necessary software packages directly from the standard apt repository and quickly start developing your machine learning projects.
The hardware itself is pretty powerful for a Pi add-on, especially when looking at the specs. The AI HAT+ is rated at 26 TOPS, which immediately tells you that this card is designed for heavy-duty tasks like real-time image processing and serious neural network acceleration.
There’s a big technical change here that I think is actually a huge win for people who frequently tinker with these modules. Raspberry Pi has decided to remove the Hailo device driver from major kernel releases. Instead, the team will now use Dynamic Kernel Module Support (DKMS) to create and install this kernel driver during the package installation process.
This decoupling is a very smart decision because it gives the development team more flexibility when releasing future software versions. Importantly, this change allows users to downgrade the device driver without being forced to downgrade the entire kernel as well.
If you have already generated custom models using an older version of the Hailo Dataflow compiler, the ability to roll back just the driver is essential to maintaining compatibility with your existing work. The installation steps are almost exactly the same as before, but now you need to make sure you have installed the DKMS framework first before installing the main ‘hailo-all’ package.
The Raspberry Pi AI Camera also received a major update regarding a feature that developers have apparently been asking for since the camera’s launch. The team implemented an input tensor injection functionality on the AI camera. This tool allows you to easily debug custom or specially designed neural networks running directly on the device.
I would say the biggest problem when deploying custom AI models is verifying that they actually work correctly under various conditions after they are compiled. Input tensor injection solves this problem by providing a reliable testing methodology. This feature allows you to validate the quality and performance of the network by repeatedly feeding it images from an existing dataset.
Whether you use a standard, well-known data set like COCO or something completely customized and tailored to your specific application, the process ensures you have reliable, repeatable testing. This is the kind of feature that dramatically improves the development experience for anyone seriously interested in edge computing.
To take advantage of this new debugging feature for your AI camera, simply run the standard update commands. You need to run “sudo apt update” and then continue with “sudo apt full-upgrade -y” to make sure your system is completely up to date. There’s even a sample input tensor injection script available if you want to try the feature right away.
Source: Raspberry Pi
