Smart vision, smarter automation

Bring intelligent vision to your projects with HomeGenie Mini firmware for ESP32 Camera. This open-source firmware transforms your ESP32 Camera module into a powerful, adaptable smart camera, seamlessly integrating with HomeGenie's advanced A.I. capabilities for unparalleled automation across diverse applications.

Features

With HomeGenie Mini firmware, your ESP32 smart camera gains the following superpowers:

• Secure remote control: Control and monitor from anywhere via MQTT with end-to-end encryption applied to the video stream and all communications. Your data stays private.
• Scheduled recording & time-lapse: Schedule picture saving to SD-Card^(*) or create captivating time-lapse sequences using the built-in scheduler.
• In-house AI with HomeGenie Server: Unleash powerful real-time and A.I.-based actions by integrating with HomeGenie Server. Keep your data local, with no need for Internet connectivity or external service providers.
• Simple API for easy integration: Integrate and use it with any other application or services.
• Flexible automation: Automate and integrate with other projects using pre-built actions or leverage the built-in JavaScript engine to create custom automated tasks, such as sending notifications or controlling other devices.

Tested hardware

The following hardware configurations have been tested with HomeGenie Mini Smart Camera firmware. Tests were performed using maximum picture quality settings and JPEG output. Both tested boards utilized the same camera module, an OV3660 (3MP).
Test results are shown in the tables below.

ESP32-CAM

This board embed a ESP32S single core 160MHz chip with 4MB of PSRAM, enough to handle up to 5MP resolutions.

Frame size	FPS max	Resolution
480x320	14	HVGA
640x480	12	VGA
800x600	10	SVGA
1024x768	9	XGA
1280x720	8	HD
1280x1024	7	SXGA
1600x1200	5	UXGA
2048x1536	4	QXGA

Given these performance characteristics, the ESP32-CAM is now better suited for applications where moderate frame rates are acceptable. While smooth, real-time video is still limited, it's capable of event-triggered image capture, periodic video updates, and monitoring moderately dynamic environments. The ESP32-CAM remains an excellent choice for budget-constrained projects.

Specific application examples for ESP32-CAM

• Simple security camera: Detect motion and send high-resolution image to a server
• Bird feeder monitor: Take more frequent pictures to capture more bird activity.
• Pet or baby monitor (basic): Provide infrequent video updates for checking on a baby or a pet.
• Time-lapse photography: Capture time-lapse sequences at a low to medium resolution.

ESP32-S3-CAM

This board embeds an ESP32-S3 dual-core 240MHz chip with 8MB of PSRAM. At almost the same cost as the ESP32S-based model, this ESP32-S3-based board represents a significant upgrade in terms of processing power (dual-core vs single-core) and PSRAM (8MB vs 4MB).

Frame size	FPS max	Resolution
480x320	22	HVGA
640x480	20	VGA
800x600	18	SVGA
1024x768	15	XGA
1280x720	15	HD
1280x1024	12	SXGA
1600x1200	10	UXGA
2048x1536	7	QXGA

The ESP32-S3-CAM, with its significantly improved performance, opens up a wider range of possibilities.

Specific application examples for ESP32-S3-CAM

• Advanced security camera: perform person detection, object tracking, and real-time alerts with higher accuracy.
• Smart home automation: integrate with smart home systems for more responsive and intelligent control based on visual data.
• Industrial monitoring: monitor equipment, detect anomalies, and perform quality control checks on production lines.
• Retail analytics: track customer movement, identify popular products, and optimize store layouts.
• Robotics: provide robots with better visual perception for navigation, object manipulation, and human interaction.
• Higher quality time-lapse: capture time-lapse at a better resolution

In short, the ESP32-CAM is suitable for simple, low-bandwidth tasks, while the ESP32-S3-CAM is suitable for advanced use-cases that require better quality images or real-time analysis.

Like other ESP SoCs, both ESP CAM includes 2.4GHz Wi-Fi connectivity and also Bluetooth though only Wi-Fi is enabled in this firmware.

Unleash the power of AI vision

The HomeGenie Mini firmware leverages the ESP32 Camera's capabilities and unlocks its potential with HomeGenie's cutting-edge A.I. engine. By combining the camera's visual input with the processing power of HomeGenie, you can create truly intelligent solutions. This integration uses the latest advancements in A.I. and Machine Learning, including:

• Object detection: recognize and identify objects within the camera's field of view (e.g., tools, components, people, vehicles, products, packages). The versatility allows for applications far beyond the home.
• Motion analysis: go beyond simple motion detection by understanding the type of movement. Is someone walking, running, interacting with equipment, or is it just environmental changes? This nuanced analysis reduces false positives and enables more context-aware automation in any setting.
• Scene understanding: analyze the overall scene to infer activities and situations. Is a room occupied, is a machine operating, is a space cluttered? This allows for intelligent decision making in a range of scenarios.

These A.I. functionalities are seamlessly integrated into HomeGenie's familiar control panel, allowing you to create complex automation rules with simple drag-and-drop interfaces, readily adaptable for any project.

Getting started

Uploading the firmware

To upload the HomeGenie firmware to your ESP device, connect the device to a USB port on your PC. Open the HomeGenie website and navigate to the firmware upload page. In the firmware upload form, click the "Create device" button. On the subsequent screen, click the "Continue" button, select the appropriate USB port from the list, and confirm your selection. Wait for the firmware to be uploaded to your device.
After the upload completes, reset the device; the LED will start blinking rapidly. Click the 'Initialize firmware' button on the upload form to finalize the firmware upload.

Open media

Connecting to Wi-Fi

After firmware installation, the ESP device LED will blink rapidly, indicating it's not connected to Wi-Fi. To configure Wi-Fi, press and hold the WPS button for 4 seconds until the LED is steady ON. Enable WPS mode on your router just long enough for device connection to take place, this may take several seconds. Once the device is successfully connected the LED will start to blink slowly signifying normal device operation.
HomeGenie Mini devices can be alternatively connected also via Bluetooth or serial terminal. The Bluetooth option is not available for ESP32-CAM firmware.

Open media

Using with HomeGenie Panel app

HomeGenie Panel app allows you to control the smart camera from your smartphone. To get started, open the HomeGenie Panel app, start the device discovery and then find and select your smart camera device from the list of available devices. Confirm device location and time settings and select all modules. The smart camera is now added to the panel dashboard.
By clicking the camera widget preview will open the full screen view of the smart camera. From there you can control the integrated RGB LED or take pictures. The widget also allows you to change video resolution and control various image settings including brightness, contrast, saturation and basic effects.

Open media

Clicking the status button on the toolbar allows you to check connected devices and enable MQTT Network for remote control and monitoring of the smart camera and other devices, giving you freedom to choose to use your own MQTT service or one of the public MQTT services, some of which are free and require no registration.
End-to-end encryption with offline key exchange is enabled by default for enhanced privacy and security of your data including video streams.

Using with HomeGenie Server

HomeGenie Server lets you unlock the power of machine learning and AI with your ESP32 smart camera. Supporting YOLOv8, YOLOv10, and YOLOv11, it provides state-of-the-art, real-time object detection, even on resource-constrained ARM64 SBCs like the Raspberry Pi.
These versatile models enable a wide range of computer vision tasks, including object detection (identifying and locating objects), instance segmentation (detecting object boundaries), image classification (categorizing images), pose estimation (detecting and tracking keypoints on human bodies), and oriented object detection (OBB).
Best of all, HomeGenie Server includes the latest YOLOv11 pre-trained models for object detection and pose estimation, so you can start experimenting immediately!

Open media

Application scenarios

• Smart security: receive alerts only when a person is detected on your property or in a restricted area, minimizing false alarms.
• Elderly care: monitor the activity levels of elderly relatives in their homes, or patients in a care facility. Detect falls or lack of movement and automatically notify caregivers.
• Pet monitoring: track your pet's activity, detect unusual behaviors, and automatically dispense food or water as needed. Or, in an agricultural setting, monitor livestock behavior.
• Energy management: automatically turn off lights and appliances when a room is unoccupied based on object and scene recognition. Extend this to optimize energy usage in commercial buildings or factories.
• Package delivery Notifications: receive instant alerts when a package is delivered and placed within the camera's view. This could be used for inventory tracking in warehouses.
• Industrial monitoring: monitor production lines for anomalies, track tool usage, or ensure safety protocols are followed.

Create this device now! 🪄

Connect your ESP32/ESP8266 microcontroller to your computer via USB, select your firmware version, and click "Create device" to upload the HomeGenie Mini firmware.

See the Device setup page for further information about configuring a HomeGenie Mini device.

Modules and API

In addition to the common Device API, this device implements the following modules and API.

CAM module

Controls camera sensors.

Domain / Address

HomeAutomation.HomeGenie/CAM

Properties

• Sensor.ObjectDetect
• Sensor.ObjectDetect.Subject

Commands

• Camera.GetPicture