Full HD Sony® Starvis™ IMX662 Smart AI GigE Camera

Overview

SMARTOdin22 is a 2MP GigE camera with an in-build NPU capable of running various AI models. This AI-enabled camera is built around the Sony STARVIS IMX662 sensor. It boasts advanced AI features, including comprehensive object detection such as Person, Bicycle, Motorcycle, Car, Bus, and Truck recognition, along with robust facial detection capabilities. It is embedded with a dedicated NPU SDK that enhances the module's AI capabilities, empowering your applications with cutting-edge intelligence.

This is an ONVIF-supported ethernet Gigabit camera that can reliably send images and control data over a wired LAN network. This camera has HDR (High Dynamic Range) functionality making it possible to run AI algorithms even in challenging lighting conditions. Its superior low-light and NIR sensitivity further enhance imaging quality, particularly in extremely low-light conditions. The GigE interface of this camera enables the transfer of video data with a maximum cable length of up to 100m.

Specifications

Key features:
- Frame Rate: HD @ 60 fps & Full HD @ 60 fps
- Output format: Compressed MJPEG, H.265 and H.264 formats
- Shielded Cat6 M12 to RJ45 cable for transmission of both power and data for long distance
- Cloud connect - FCAPS, Firmware Upgrade, Video streaming
- Unique ID for each camera
- Lightweight, versatile, and portable design
Interface:
- Interface: GigE
- Connector: M12
- Cable length: 3m
Camera features :
- Sensor: Sony® STARVIS 2 IMX662
- Focus Type: Fixed focus
- Sensor Resolution: Full HD
- Chroma: Color
- Shutter type: Electronic Rolling Shutter
- Optical Format: 1/2.8"
- Output Format: Compressed MJPEG, H.265 and H.264 formats
- Pixel size: 2.9 μm x 2.9 μm
- Sensor Active Area: 1937H x 1097V
- Dynamic Range: 88 dB
- Holder: M12 (S-Mount)
- FOV: TBD (with the lens provided by e-con)
Electrical & Mechanical:
- Power-up Options:
  - IEEE802.3af standard PoE Input through M12 Connector
  - 12V supply through M8 connector for non-POE applications
- Operating Temperature Range: -30°C to 70°C
- Power requirements: 2.4W
- Size in mm (l x b x h) :
  - Without lens: 46mm x 46mm x 53.95mm
  - With lens: 46mm x 46mm x 63.06mm
Miscellaneous:
- Compliance: RoHS

* - Not available with the off-the-shelf version. Please contact camerasolutions@e-consystems.com for more details.

Software

- Supported OS - Windows, Linux
- Supported Profiles - ONVIF supported Profiles S & T
- Supported Streaming formats - MJPEG, H.265 and H.264
- Ethernet IP Configuration - DHCP
- Features of RouteCAM Viewer Application
  - Automated IP Discovery & Streaming
  - Support for Camera Controls
  - Provision for changing resolution and format available in Linux.
- Linux Socket Application for extended Camera features
- OTA (Over-The-Air) Upgrade *
- Cloud connect - FCAPS, Firmware Upgrade, Video streaming
Artificial Intelligence(AI) Models:
- Dedicated NPU SDK
  - Object detection includes Persons, Bicycles, Motorcycles, Cars, Buses, Trucks
  - Face detection
- Supported Model Architecture
  - mobilenetV2_SSDliteV1
  - mobilenetV2_SSDliteV2
  - mobilenetV2_SSDliteV1_full_rfb**
  - Yolov5s**
  - Yolov7-tiny**
- To explore Artificial Intelligence(AI) Models, refer here » refer here»
Software controls:
- Brightness
- Contrast
- Saturation
- White Balance
- Gamma
- Gain
- H Flip
- V Flip
- Sharpness
- Exposure
- Focus (only for liquid lens)
- ROI size
- Denoise
- Exposure Compensation

Please contact camerasolutions@e-consystems.com for more details.

** Deformation of Original architecture

AI capabilities

Model	Description		Explanation
Face Detection	Detects faces in an image	Number of Faces Detected Time of Capture Capture Image on Detection	Indicates the count of faces identified within the image. Specifies the timestamp when the image containing faces was captured. Triggers the system to capture an image whenever a face is detected.
Vehicle Detection	Detects vehicles in an image	Total Vehicle Detections Time of Capture Breakdown by Category Capture Image on Detection	Displays the total count of vehicles detected in the image. Records the timestamp when the image was taken. Classifies detected vehicles into categories such as cars, trucks, or motorcycles. Initiates image capture whenever a vehicle is detected.
Front Vehicle Distance estimation	Estimates distance to the front vehicle	Vehicle Presence (yes/no) Proximity Alert Record Video on Close Proximity	Indicates whether there is a vehicle in front (yes) or (no). Alerts if the distance to the front vehicle is too close, indicating a potential collision risk. Activates video recording when the distance to the front vehicle becomes too close.
Perimeter Security	Detects person intrusion	Proximity Alert Capture Image on Detection	Triggers an alert when a person intrudes into a secured area, indicating a possible security breach. Captures an image when a person intrusion is detected, allowing for visual documentation of the event.

Customization

SMARTOdin22 - Customization:

Are you looking for different AI models?
Do you need a different lens to meet your FOV and other requirements?
Do you need a lens with motorized control?
Do you want a different form factor?
Do you need LED flash?
Complete customization of firmware and hardware as per customer requirements. Please contact us for any customization and additional features.

SMARTOdin22_PoE_IP67 - Cloud-connected Smart AI Camera

Overview

Specifications

Software

AI capabilities

Customization

SMARTOdin22 - Customization:

Related Videos

Targeted Applications

Related Articles