OpenNCC V2 industrial target detection camera

Product Overview
OpenNCC V2 is a professional, edge-side AI vision processing platform built for intelligent perception, real-time inference, and rapid deployment. It integrates high-performance hardware processing, YOLOv8 deep learning framework support, and flexible infrared night vision expansion, providing a complete solution from model training, conversion to embedded deployment. It is widely used in intelligent monitoring, industrial inspection, security early warning, and IoT vision scenarios.
 
Key Features
1. Full-Link AI Model Support
Compatible with YOLOv8 series models, supporting custom dataset labeling, training, and optimization
Built-in AnyLabeling tool for fast image annotation and dataset preparation
Supports model conversion to TFLite/ONNX formats with INT8 quantization for efficient edge deployment
One-click model validation, testing, and export for NXP i.MX8MP platform
2. Stable & Reliable Hardware Performance
Professional vision processing architecture with strong real-time computing
Compatible with dedicated Night Vision IR LED Board for clear imaging in total darkness
IR LED specifications: 850nm wavelength, 120° wide angle, 4 high-power LEDs
Operating temperature: -20°C to +65°C, adapting to harsh industrial environments
3. Easy Web Management & Configuration
Intuitive web-based UI for quick device access and parameter setup
Supports DHCP/static IP networking and UDP video stream output
Real-time video preview via standard players (VLC)
Visual configuration for inference models, confidence thresholds, and NMS parameters
Event-triggered recording with pre/post event frame storage
4. Flexible Perception & Alarm Output
Customizable detection classes and confidence thresholds
GPIO trigger input and alarm signal output for external linkage
Supports HTTP Push with JSON structured alarm data
Silent frame mechanism to avoid repeated alarms
Remote firmware upgrade and factory reset
5. Rapid Development & Deployment
Complete toolchain for environment setup, training, conversion, and deployment
Compatible with Ubuntu 20.04 and Conda virtual environment
Simplified commands for model training, export, and testing
SMB network sharing for accessing inference results (videos/images)
Seamless connection with supporting infrared light board components
 
Technical Specifications
Item
Specification
AI Framework
YOLOv8, support custom training
Model Formats
TFLite (INT8 quantized), ONNX
Deployment Platform
NXP i.MX8MP
IR Light Board
4 × 850nm IR LEDs, 120° beam angle
Input Voltage
12V DC
Operating Current
450±10 mA
Operating Temperature
-20°C ~ +65°C
Storage Temperature
-40°C ~ +80°C
Network
Ethernet, DHCP / Static IP
Streaming Protocol
UDP, compatible with VLC
Management
Web GUI, default admin/admin1234
Inference Result Access
SMB network share
 
Typical Applications
Intelligent security monitoring & perimeter protection
Industrial automated inspection & defect detection
Night vision monitoring with IR auxiliary lighting
IoT edge vision perception & early warning
Custom target detection & behavior analysis