Free AI Motion Detection with Reolink, AI Tool & Home Assistant

Enhance your home security with an AI-powered motion detection system using Reolink cameras and Home Assistant. This guide provides a comprehensive, step-by-step approach to setting up a free and effective system using readily available tools. By integrating Reolink cameras with iSpy, AI-Tool, and Deepstack AI, you can achieve intelligent motion detection and receive timely notifications through Home Assistant.

To implement this AI motion detection system, you will need the following components: * **Reolink Cameras:** (e.g., RLC-520, RLC-810A) - These cameras provide the video feed for motion detection. * **iSpy v7.2.1.0 64-bit (Freeware):** This software manages the camera feeds and captures images upon motion detection. * **AI-Tool:** This tool processes the captured images using AI to identify specific objects. * **Deepstack AI:** This AI engine performs the object detection. * **Home Assistant:** This home automation platform integrates all components and provides notifications.

1. **Install iSpy:** Download and install iSpy on your Windows machine. 2. **Add Cameras:** * Open iSpy and use the “Add IP Cameras with Wizard” option. * For Reolink RLC-520 and RLC-810A, select “Not Listed” and manually add the Make and Model. * Tick the “Scan RTSP Addresses” box and click Next. * Enter your camera's Username/Password. * If you know the camera’s IP address, input it manually or click “Scan Local Network”. * Choose the stream. Using the sub stream (e.g., FFMPEG:410:rtsp://user:password@IP Address:554/h264Preview_01_sub) is recommended to reduce network traffic. 3. **Name the Camera:** Once added, rename the camera using the “Edit Camera” option.

1. **Motion Detection Tab:** * Go to the “Edit Camera” screen and select the “Motion Detection” tab. * Draw detection zones over the areas where you want motion to be detected. Covering the entire area is acceptable, as the AI will refine the detection. 2. **Images Tab:** * Enable “Local Saving Enabled” and set “Save When” to “Motion Detection”. * Remove the “Overlay Text” if desired. 3. **Storage Tab:** * Set the “Media Location” to a folder you can easily remember (e.g., C:\aiinput\video\CameraName\Grabs). * Enable “Storage Management” to automatically delete images after a set period (e.g., 24 hours).

1. **Install Deepstack AI:** Download and install Deepstack AI for Windows. 2. **Download AI-Tool:** Get the AI-Tool .zip from GitHub. 3. **Extract AI-Tool:** Extract the contents of the .zip folder to a location like Program Files.

1. **DeepStack Tab:** * Select your Deepstack installation folder. * Set API to “Detection API”, Mode to “High”, and choose a Port (e.g., 83). * Tick “Automatically Start Deepstack”. 2. **Settings Tab:** * Set the “Default Input Path” to the folder used for Media Location in iSpy. * Set the “Deepstack URL” to localhost:port# (e.g., localhost:83). * Tick “Start with user login”. 3. **Cameras Tab:** * Click “Add” and give the camera a name (e.g., FrontDoor). * Set the “Input Folder” to the iSpy image folder (e.g., C:\aiinput\video\FrontDoor\Grabs). * Select the relevant objects you want detected (e.g., Person, Car). 4. **MQTT Settings:** * Under Actions, click “Settings” next to MQTT. * Input your Home Assistant MQTT broker Server:Port (e.g., 192.168.0.72:1883) and credentials. * The Topic will be ai/[Camera]/motion [Detection], which passes a payload like ai/FrontDoor/motion Person. * Set a “Cooldown Time” to avoid excessive notifications (e.g., one minute).

Update your Home Assistant `Configuration.yaml` file to listen for the MQTT payloads from AI-Tool. Here’s an example configuration: ```yaml binary_sensor: - platform: mqtt name: “Motion - Front Door - Vehicle” device_class: motion state_topic: “ai/FrontDoor/motion” payload_on: “Car” off_delay: 60 - platform: mqtt name: “Motion - Front Door - Person” device_class: motion state_topic: “ai/FrontDoor/motion” payload_on: “Person” off_delay: 60 ```

Create automations in Home Assistant to trigger actions based on the MQTT messages. For example, to send a notification when a vehicle is detected: 1. **Name:** Motion Detection - Vehicle Front Door 2. **Mode:** Single 3. **Triggers:** * Type: MQTT * Topic: ai/FrontDoor/Motion * Payload: Vehicle 4. **Actions:** * Type: Call service * Service: Notify.mobile_app_iphone * Service Data: ```yaml message: Motion Detection - Front Door - Vehicle data: attachment: content-type: jpeg push: category: camera entity_id: camera.front_door ```

* **False Positives:** Adjust detection zones and object detection settings in AI-Tool to minimize false positives. * **Network Traffic:** Using the sub stream from the Reolink camera helps reduce network load. * **Cooldown Time:** Setting an appropriate cooldown time in AI-Tool prevents notification spam. * **Deepstack Performance:** Ensure Deepstack AI is running smoothly and has sufficient resources.

By following this guide, you can set up a robust and free AI-powered motion detection system using Reolink cameras, iSpy, AI-Tool, and Deepstack AI integrated with Home Assistant. This system provides intelligent object detection and timely notifications, significantly enhancing your home security.