Frigate NVR with Odyssey Blue and Coral USB AI Accelerator
Frigate is a completely open source and local NVR designed for Home Assistant with AI-powered object detection. With stylish Odyssey Blue mini PC powered by Intel Celeron J4125, Coral TPU, and IP cameras, you are all set local NVR system with local AI processing!
Designed for Home Assistant, seamless integration into your smart home system. Make your home eyesight by integrating object detection into Home Assistant, OpenHab, NodeRed, or anything with MQTT support.
Today we wanna share a super cool open-source project to help you build up an NVR system with AI detection, and everything will be processed locally! Frigate is a completely open source and local NVR designed for Home Assistant with AI-powered object detection(Kudos sent to Frigate team!). It uses OpenCV and Tensorflow to perform real-time object detection locally for IP cameras. It brings a rich set of features including video recording, re-streaming, motion detection, and supports multiprocessing.
Traditional NVRs have higher false positive rates and can require hours of fine-tuning to reduce false positive rates because they rely on simple motion detection. Leveraging Google Coral TPU USB accelerator for 100+ FPS object detection, you can run advanced AI analysis with an Odyssey Blue, an Intel Celeron J4125 powered powerful Linux mini PC, to determine if the motion is actually a person, car, or any object you want to detect. With Frigate’s local processing, there is no need cloud fee anymore for storing your personal camera footage and the analysis record.
All processing performed locally: No need to pay for sending your personal camera footage to the cloud for analysis.
Never miss a frame of detection: Frigate can run over 100 object detections per second, so it never misses a frame. Frigate tracks objects in real-time and can determine the exact time a person starts walking up your front steps or a car enters your driveway.
Fine-tune your events and alerts with zones: Frigate tracks objects in real-time and can determine the exact moment a person starts walking up your front steps or when a car enters your driveway. Refine your notifications based on precise locations.
Designed for Home Assistant, seamless integration into your smart home system
Give your home eyesight by integrating object detection into Home Assistant, OpenHab, NodeRed, or anything with MQTT support. Frigate integrates directly into Home Assistant’s media browser, provides a low-latency camera entity, and exposes real-time sensors and switches to power automation and notifications to your heart’s content.
Easily monitor cameras with activity: Dynamically renders cameras with active detection with Birdseye view.
The best hardware combo: Odyssey Blue X86J4125+ Coral TPU + IP cameras
Use of a Google Coral Accelerator is highly recommended if you are aiming to outperform even the best CPUs and can process 100+ FPS with very little overhead. We were only able to achieve 5-10 FPS with a CPU, so that's more than a 10x speed improvement.
For the Server, we found a discussion on Reddit ”What’s the best hardware for Frigate using a Coral m.2 instead of the USB version?” Seeed Odyssey Blue J4125 wins! “I use the Odyssey X86 Blue by Seeed Studios. It has an Intel Celeron J4125, 8GB RAM and two M.2 slots (1xSATA and 1xPCIe). I run Frigate in addition to 9 other docker containers (e.g. Home Assistant, Pihole, Samba, Unifi Controller, etc.) and sit around 25% CPU usage running 4x 720p detect streams at 15 fps (5 fps set within Frigate).” -- shred86
Odyssey Blue J4125 128 GB is a powerful mini PC with super low power consumption, it can work with Frigate stable and can also extend to other applications like NAS, 4K media theater, Home Automation, gaming PC, and many more! It's also a perfect device for personal, commercial, and industrial applications. Are you a Linux user? Install whatever Linux OS you like. We also have abundant keep updating tutorial resources to help everyone no matter you are a mini PC beginner or a developer!
For cameras, Frigate documentation recommends Dahua, Hikvision, and Amcrest in that order. "Dahua edges out Hikvision because they are easier to find, not because they are better cameras. I personally use Dahua cameras because they are easier to purchase directly."
Demo of using Odyssey Blue only and 2 MP Fixed Bullet Network Camera.
- Tight integration with Home Assistant via a custom component
- Designed to minimize resource use and maximize performance by only looking for objects when and where it is necessary
- Leverages multiprocessing heavily with an emphasis on real-time over processing every frame
- Uses a very low overhead motion detection to determine where to run object detection
- Object detection with TensorFlow runs in separate processes for maximum 100+ FPS
- Communicates over MQTT for easy integration into other systems
- 24/7 recording: records video with retention settings based on detected objects
- Re-streaming via RTMP to reduce the number of connections to your camera
What is Home Assistant?
Home Assistant is a free and open-source software for home automation that is designed to be the central control system for smart home devices. It can be accessed via a web-based user interface, via companion apps for Android and iOS, or using voice commands via a supported virtual assistant like Google Assistant or Amazon Alexa.
Home assistant enables the local control of smart home devices, offering better security, greater reliability, and more flexibility than many cloud-connected devices.
More Home Assistant projects working with Seeed hardware：
- Smart Home with Home Assistant on OpenWrt
- How to use Home Assistant with a Seeed Mini Router (Installation, Configuration, and Control)
Get started easily
You can also follow up on this step-by-step guide and install Home Assistant and Frigate with the Odyssey X86J4125 series.
Check out Everything Smart Home YouTube video explaining how to use Frigate NVR with Home Assistant!