Visual Inspection
Capture repeatable images at a machine or workstation for condition checks, presence checks and inspection prototypes.
ESP32-S3 EDGE CAMERA
NORVI AI Optic
NORVI AI Optic
ESP32-S3 Camera
A compact edge imaging platform for embedded vision, machine monitoring and TinyML development. Combine a 5 MP OV5640 autofocus camera with ESP32-S3 processing, local display, microSD storage, dual LED illumination and external event triggering.
ESP32-S3
5 MP Autofocus
16 MB Flash
8 MB PSRAM
microSD
External Trigger
USB-C
ORDERING STATUS
Temporarily Not Accepting Orders
AI Optic is currently unavailable for new orders. Documentation
and technical information remain available.
NORVI HardwareDesigned and manufactured by NORVI
CE / EMCDocumented compliance standards
Open DevelopmentArduino and ESP32 workflows
AI OPTIC / PRODUCT VIEW
CORE CAPABILITIES
Everything needed to move from image capture to edge logic
AI Optic combines camera, processing, storage, local interaction and triggering in one enclosed ESP32-S3 platform, reducing the amount of supporting hardware needed for a vision prototype or embedded monitoring node.
5 MP AutofocusOV5640-AF image sensor for focused image capture.
ESP32-S3 Compute16 MB flash and 8 MB PSRAM for embedded vision workloads.
Local DisplayLCD configuration supports local preview and device UI.
microSD StorageStore captured images and application data locally.
External TriggerSynchronize image capture with an external event or sensor.
Dual LED LightingIntegrated white illumination for controlled close-range
capture.
ENGINEERING FIT
A camera that can be part of the machine workflow
Instead of treating the camera as a standalone imaging accessory, AI Optic can capture from a physical trigger, process locally on the ESP32-S3, display the result and retain data on microSD.
Event-driven capture
Trigger an image when something actually happens
Use the external trigger input to connect image capture with a button, dry-contact event or suitable interface circuit from another system.
Local processing
Keep lightweight vision logic close to the camera
ESP32-S3 processing and PSRAM make the platform suitable for embedded computer-vision experiments and TinyML workflows such as the NORVI Edge Impulse object-detection example.
01 / EVENT
Trigger or Schedule
Capture on a local button, external event or programmed interval.
02 / CAPTURE
5 MP Autofocus
Acquire an image with the OV5640-AF camera and integrated illumination.
03 / PROCESS
ESP32-S3 Edge Logic
Run application code, preprocessing or a compact ML inference workflow locally.
04 / USE
Display, Store or Send
Preview locally, save to microSD or connect the result to a Wi-Fi application.
HARDWARE AT A GLANCE
Compact vision hardware with the supporting interfaces built in
The product is more than a bare ESP32 camera board. It packages the camera, user controls, display option, storage, lighting, rechargeable backup power and mounting capability into a finished enclosure.
Imaging
OV5640-AF Camera
5 megapixel autofocus camera module for image capture and embedded vision development.
Operator Interface
2.1-inch TFT
ST7789 display on LCD-equipped configurations, with 240 × 320 pixel resolution.
Storage
microSD over SPI
Local removable storage for captured images, datasets and application files.
Control
Rotary / Toggle Control
Use rotation and push actions to navigate settings and build a compact local user interface.
Power & Programming
USB-C
USB-C is used for ESP32-S3 programming and battery charging.
Mounting
Tripod-style Mounting
Use the built-in screw-type holder with bracket or long-arm mounting accessories.
EDGE AI WORKFLOW
Train with your own images, then run inference on the ESP32-S3
NORVI documents a practical workflow using the AI Optic camera as a web server for collecting training images, training a model in Edge Impulse, exporting it as an Arduino library and deploying the model back to the ESP32-S3 for local classification.
Collect images directly from the deployed camera hardware.
Use autofocus routines with the OV5640 sensor.
Train compact vision models using Edge Impulse.
Export the model as an Arduino library for ESP32-S3
deployment.
Run classification locally without requiring a cloud inference
round trip.
APPLICATION IDEAS
Build visual monitoring and embedded machine-vision prototypes
Use AI Optic where a compact local camera, event trigger, storage and programmable edge processing are more useful than a general-purpose USB or IP camera.
Object Detection
Train compact image models and test local inference workflows with the ESP32-S3 and Edge Impulse.
Barcode & Label Workflows
Use controlled image capture as the imaging layer for barcode, label and identifier-reading applications.
Meter Reading
Capture meter faces on a schedule or trigger event for later processing or an application-specific recognition model.
People Counting
Prototype entry, exit and occupancy-counting logic using a locally deployed edge camera.
Time-lapse Monitoring
Capture images over time for plant growth, habitat observation, equipment state or long-running process review.
CONFIGURATION OPTIONS
Available display and mounting formats
AI Optic has been offered in multiple display and mounting formats so the same camera platform can support compact development setups and fixed imaging tasks. New orders are temporarily paused.
With LCD Display
Use the local display for captured-image preview, settings and application-specific user interfaces.
Ordering temporarily pausedWithout LCD Display
Use the camera-focused configuration when a local screen is not required by the deployment.
Ordering temporarily pausedMounting Accessories
Camera bracket and long-arm mounting options support different installation needs.
Ordering temporarily paused
TECHNICAL SPECIFICATIONS
Key specifications for engineering evaluation
The table below focuses on the specifications most useful for deciding whether AI Optic fits an application. Full GPIO mappings and peripheral details remain in the datasheet and user guide.
MCU / ModuleESP32-S3-WROOM-1U-N16R8
Flash Memory16 MB
PSRAM8 MB OSPI
CameraOV5640-AF, 5 MP, autofocus
LCD-equipped configurationST7789 TFT, 2.1 inch, 240 × 320 pixels
WirelessWi-Fi
microSDSPI interface
USBUSB-C for programming and charging
External Trigger1 digital input, dry-contact trigger function
Trigger Input Maximum Rating3.3 V DC / pull-down to ground
LightingDual integrated white LED flashlights
Local ControlRotary / toggle switch plus trigger and reset controls
EnclosureHH-042
Dimensions69 × 50 × 21 mm
MountingScrew-type female holder
IP Degree of ProtectionIP20
Operating Temperature-10 to +85 °C
Operating Altitude0 to 2000 m
DOCUMENTATION & EXAMPLES
Go from product evaluation to working firmware
Keep the landing page concise, then move developers into the exact technical resource they need for wiring, GPIO mapping, display setup or edge AI development.
COMPLIANCE & ENVIRONMENT
Documented for industrial electronic environments
EN 61131-2:2007Programmable controller equipment requirements listed in NORVI
documentation.
EN 61010-1:2010+A1:2019Safety requirements for electrical equipment for measurement,
control and laboratory use.
EN IEC 61010-2-201:2018Particular requirements for control equipment documented by
NORVI.
2014/30/EU EMCElectromagnetic compatibility directive listed in the AI Optic
datasheet.
FAQs
Common
Common
Questions
Can't find what you're looking for? Our team is available for a free technical consultation.
AI Optic uses the ESP32-S3-WROOM-1U-N16R8 module with 16 MB flash and 8 MB PSRAM, providing a practical base for camera handling, local UI, networking and compact edge-vision applications.
The product uses the OV5640-AF camera module with 5 megapixel resolution and autofocus support.
Yes. NORVI provides an object-detection workflow that collects camera images over Wi-Fi, trains a model with Edge Impulse, exports it as an Arduino library and deploys it back to the ESP32-S3 for local classification.
Yes. AI Optic includes one external trigger input so image capture can be synchronized with a suitable external event or dry-contact interface.
Yes. The product includes microSD card support using an SPI interface, allowing captured media and application data to be stored locally.
AI Optic has been offered with and without the LCD display. The LCD-equipped version uses a 2.1-inch ST7789 TFT with 240 × 320 pixel resolution.
USB-C is used for ESP32-S3 programming and battery charging. The device also contains a rechargeable backup battery.
PRODUCT STATUS
AI Optic orders are temporarily paused
We are currently not accepting new orders for NORVI AI Optic. You can still review the hardware documentation and example workflows, or contact NORVI for technical and product enquiries.
