This summary of the video was created by an AI. It might contain some inaccuracies.
00:00:00 – 00:25:37
The video provides a comprehensive guide on integrating free local AI computer vision with Blue Iris, focusing on optimal setup and use for security surveillance. Key points include installing Blue Iris software alongside the open-source CodeProject.AI, configuring both for motion, facial, and license plate recognition. The guide explains best practices for hardware, user management, storage optimization, and leveraging Intel Quick Sync for efficient video processing. Specific setup instructions encompass various brands of IP cameras (e.g., Dawa, Hikvision, Reolink), with detailed configuration for AI modules, motion detection, and home automation integration. Special attention is given to setting up Reolink cameras, customizing motion detection parameters, and adjusting AI settings for object validation. For facial and license plate recognition, the video covers camera positioning, day/night settings, and using specific models like Laurita’s 64mm for LPR and Anki CZ 500 for facial detection. Essential recommendations include splitting profiles for different lighting conditions and utilizing MQTT for automated alerts. The video concludes with legal considerations for AI use and offers additional resources through ongoing updates and affiliate links.
00:00:00
In this part of the video, the presenter discusses the improvements made to adding free local AI computer vision to Blue Iris over the past two years and provides a comprehensive setup walkthrough. Key actions include installing Blue Iris and CodeProject.AI, configuring best practices and general settings, adding cameras from various manufacturers, setting up AI-based motion detection, training the AI to recognize familiar faces, and setting up free local license plate recognition using CodeProject.AI. The video emphasizes that while Blue Iris requires a one-time license fee, CodeProject.AI is free and open-source. The presenter advises on hardware requirements, and walks through the installation and setup process for Blue Iris and CodeProject.AI, mentioning important settings like using Blue Iris with a VPN and configuring storage options on a solid state drive for optimal performance.
00:03:00
In this segment of the video, the creator sets up a Blue Iris server for optimal storage and user management. They create multiple folders on a secondary hard drive, allocating specific sizes and setting parameters for deletion and protection of files. For example, a new storage folder is set to 1,000 gigabytes to store footage until it fills up, at which point the oldest files will be deleted. They also discuss user account configuration, creating a secure admin profile, and disabling the default local console account. The web server tab is highlighted as a key feature, enabling remote access through a web browser and VPN setup for secure local network access. Lastly, the hardware acceleration profile is configured to use Intel’s Quick Sync video decoding for efficient processing, and the AI tab is adjusted to use the Code Project AI server for computer vision analysis.
00:06:00
In this part of the video, the speaker walks through the setup of the AI and camera systems using Blue Iris and Code Project AI. They explain how to enable and install AI modules for facial and license plate recognition, and how to configure Blue Iris to start and stop these AI services. They also mention configuring custom models for object recognition and setting up the cameras to work with Amazon Echo and home automation platforms.
The segment covers the detailed process of adding different brands of IP cameras to Blue Iris, including Dawa, Hikvision, and Reolink. Instructions include entering IP addresses, login credentials, enabling specific camera settings like motion detection and direct-to-disk recording, and ensuring compatibility by enabling necessary protocols (e.g., ONVIF) and configuring network settings. The guide emphasizes different setup steps for each camera brand, ensuring they can properly integrate with the AI services and Blue Iris software.
00:09:00
In this segment of the video, the instructions detail how to configure various settings for a Reolink camera. Key actions include deselecting the “send rtsp keeper lives” option to prevent frequent disconnections, setting the maximum frame rate slightly higher than your camera’s frame rate, and disabling live overlays to reduce system resource usage. Audio settings can be adjusted if the camera has a microphone.
For motion detection, it is recommended to use Blue Iris’s built-in motion sensor, adjusting the minimum object size and contrast to refine sensitivity. AI object detection can validate motion events, and zones can be defined for specific behaviors, such as detecting movement in a particular direction. If the camera supports on-device motion detection, it can be configured to use the camera’s digital input.
When the camera is triggered, Artificial Intelligence (AI) validates the motion by recognizing specified objects. The AI settings include selecting the object types to identify, setting the confidence level for object recognition, and defining how many images to analyze for confirming detections. The focus is on optimizing these settings for effective and efficient motion detection and validation.
00:12:00
In this part of the video, the speaker sets up a camera system for motion detection and recording using Blue Iris software. The key actions include configuring image analysis at 500 millisecond increments, setting firing actions only upon trigger, and saving computer vision details while ignoring static images. A break time of 10 seconds and a maximum trigger duration of 120 seconds are set.
In the recording tab, options for different recording conditions are explained, with the recommendation to use “continuous plus alerts” for efficient high-resolution recording upon AI-validated motion. A 2-second buffer is used to accommodate the analysis time.
The alerts tab is configured to tailor actions based on specific AI detections, allowing for different responses to detected objects like people, cars, and dogs. Key configurations are made for exporting clips and monitoring camera status.
Testing includes verifying motion and AI detection by walking or driving in front of the camera and checking the Blue Iris status window. The AI’s effectiveness is reviewed by confirming detected objects and analyzing triggers. Finally, the speaker clones the camera settings for additional cameras, updating necessary details like IP addresses and stream URLs.
00:15:00
In this segment of the video, the speaker sets up facial detection on a front door camera. They outline the steps to enable AI facial recognition, create a storage folder named “doorbell faces,” and allocate space for storing face images. The process includes walking in front of the camera to capture various angles, registering faces using Code Project AI, and verifying detections in Blue Iris. The video emphasizes the importance of close-up, eye-level cameras for effective facial detection and discusses the limitations of capturing license plates, noting the need for precise camera positioning. The speaker also mentions using specific cameras like Anki CZ 500 and Amcrest 1063ew PTZ for license plate recognition and highlights the challenges and best practices for camera placement to achieve clear plate captures.
00:18:00
In this part of the video, the focus is on optimizing camera settings for license plate recognition (LPR) both during the day and at night. The LPR camera recommended is the Laurita’s 64 millimeter focal length with a 2 megapixel sensor. For daytime use, it’s crucial to zoom in so the car fills the frame, use a high shutter speed (at least 1/1000), and set manual fixed focus to prevent the plate from being out of frame. At night, the exposure should be minimized to highlight the reflective license plates, with infrared LEDs illuminating the image. The video also involves configuring Blue Iris software for AI-based plate recognition, advising on specific settings to accurately read license plates. Emphasis is placed on recording at full resolution continuously, adjusting the AI settings accurately, and testing to mitigate misreads and errors, which includes reviewing AI alerts to ensure accurate plate recognition.
00:21:00
In this part of the video, the creator addresses how to tackle issues with nighttime and daytime license plate recognition using Blue Iris. The solution involves setting up two different profiles: one for daytime with high confidence levels and larger detection zones, and another for nighttime with lower confidence levels and specific object sizes. They also explain how to configure Blue Iris to automatically switch profiles at sunrise and sunset by inputting location data for accurate timing. Additionally, they show how to adjust the focus and zoom settings for both day and night using web commands and Blue Iris PTZ control presets. They outline the importance of custom commands and the use of an ALPR model to optimize performance and avoid unnecessary GPU usage. Lastly, they suggest integrating the information with home automation platforms like Home Assistant using MQTT for enhanced functionality.
00:24:00
In this part of the video, the speaker explains how to set up immediate actions for a license plate camera using MQTT by adding a web request and configuring the payload with Json for alerts. They stress the importance of researching local laws regarding the legality of plate recognition and facial detection, noting that these practices are legal in their city but may not be elsewhere. The speaker mentions that while instructions are current as of April 2023, changes might occur, and updates will be provided on their website and in pinned comments. They also provide affiliate links and thank their patrons for support, encouraging viewers to like and subscribe.