Turning edge vision intelligence into scalable, production-ready delivery
Rockchip RV1106B G2, 0.5 TOPS NPU, 128MB RAM, camera FPC—edge AI at scale for IP cameras and doorbells
SigmaStar SSC308QE, dual Cortex-A32, integrated IPU/IVE, 4K encoding—for HD AI IP cameras
OV sensors + ISP/SoC, optional on-board NPU—mass production for smart home and security IP cameras
Line inspection and industrial vision capture with industrial I/O and production bring-up
Multi-core CPU + FPGA imaging platform with extensible AI post-processing for medical and industrial endoscopy
HD, low-power portable form factor—lower deployment cost for industrial inspection
Fingerprint / PIN + on-device face recognition—production-ready smart entry devices
4K stabilized action-camera form factor—custom delivery for outdoor live streaming and field capture
AWB / AE / HDR / noise reduction tuning for OV and other sensors on mainstream SoCs
Capture, encode, record, multi-interface output, and streaming—embeddable in doorbells and IP cameras
Freeze, white balance, digital zoom, dual-display—paired with rigid endoscope camera systems
Intrinsic/extrinsic calibration and timestamp sync for stereo, TOF, and optical-flow fusion
Real-time detection and segmentation pipelines—custom training and INT8 deployment
Pre-trained models for doorbells and IP cameras—fine-tune with your own data
On-device plate detection and recognition for smart IP cameras and parking—production bring-up included
Face detection and matching for doorbells, access control, and IP cameras—with quantization and fine-tuning
From imaging, modules, and PCBA to endoscopy and edge algorithms—explore full solution pages for details
Sensors, ISP, and edge models together—for doorbells, IP cameras, and service robots.
Learn more >
Optics, stereo/TOF, calibration, and pilot builds—production-ready imaging front ends.
Learn more >
Selection, layout, bring-up, SMT, and test in one flow—stable mass production.
Learn more >
4K / FPGA imaging platform for medical and industrial inspection.
Learn more >
YOLO / MobileNet from training and quantization to on-board deployment—faster time to production.
Learn more >AI R&D · Specialized SME credentials · Integrated vision hardware and algorithms
FPGA, embedded vision, OmniVision driver stacks, and edge AI—deep integration with SigmaStar and Rockchip platforms.
Up to 16-layer HDI, SI/PI simulation, MIPI high-speed optimization; SMT + AOI/X-Ray for stable volume delivery.
AWB, AE, noise reduction, and HDR tuned per OV sensor and scene—OEM-grade imaging baselines.
Shenzhen and Singapore coordination—typical engineering inquiries answered within 24 hours, on-site bring-up available.
Latest on AI vision and edge intelligence
For doorbells, IP cameras, and industrial inspection—closing the train–quantize–deploy loop to shorten customer time to production.
Doorbells, IP cameras, and industrial inspection move inference to on-board NPUs—imaging plus algorithms as core differentiator.
Robotics and line inspection adopt global shutter and parallel cameras—MIPI and sync engineering as deployment keys.
Shunter’s Shenzhen R&D team owns sensor integration, ISP, edge AI, and board-level engineering; HOVISION in Singapore handles overseas business and global customer service—one vision engineering capability for worldwide delivery.
Hardware, algorithms, and production engineering move on one chain—shorter cycles from prototype to mass production.
Schematics → PCB layout (up to 16-layer HDI) → SMT assembly → AOI/X-Ray → imaging tuning → final test. One stop—less handoff risk across vendors.
Long-term collaboration with industry partners to deliver AI vision and embedded systems at scale
