Рет қаралды 4,529
In this episode I speak with Blake Blackshear, the author of Frigate, a local NVR designed for Home Assistant with AI object detection. It uses artificial intelligence to perform real-time object detection locally for IP cameras.
Frigate has been one of my favorite open-source projects since it’s completely local, always introducing new features from user feedback, and has a huge community for support. I’ve been running Frigate in my house for over 2 years to provide alerts and record footage from all my security cameras. It’s deep integration with HomeAssistant for triggering automations and sending alerts when objects such as people are detected has been vital to it’s adoption rate with home automation enthusiasts.
With version 13 of Frigate currently in beta, I thought this would be a great time to talk with Blake to get a behind the scenes look of Frigate’s history and where it’s going in the future. Version 13’s most anticipated feature is the introduction of custom models through the Frigate+ service. While Frigate itself is open-source and free, Frigate+ is a paid service that allows you to train custom object detection models for your security cameras.
My Full Frigate 13 (Beta) Review: rogerstechtalk.com/my-frigate...
Frigate NVR: frigate.video/
Frigate Github: github.com/blakeblackshear/fr...
TIMESTAMPS:
00:00 Intro - How Frigate Got Started
00:09:12 - Blake's personal Frigate Setup
00:11:58 - Engaging with the Frigate Community
00:25:13 - Blake's False Positives
00:27:08 - New features in Frigate Version 13
00:34:35 - Creating the Frigate+ Custom Model
00:49:34 - The Frigate+ Backend Architecture
00:54:38 - Annotating Images for Frigate+ Custom Models
01:02:46 - Privacy Concerns with Frigate+
01:07:30 - What's next for Frigate