● 100% Non-Red Supply Chain · Made in Taiwan · CTE Group

Drone vision that sees in the dark — from FPGA today to a dedicated AI ISP chip.

AI Angel builds a non-red AI image-signal-processing (ISP) and edge-AI platform for drones — real-time HDR, denoising and scene AI on the edge, with secure over-the-air model updates. Proven on FPGA today, advancing to a dedicated 28 nm AI ISP chip. With partner JANMAN Precision, the rugged airframe and the AI brain arrive as one turnkey package.

120 dB HDR (AI ISP target)
2–4 TOPS edge AI @ INT8
Federated Learning — data stays on-device
AES-256 + Root of Trust
Lattice FPGA → TSMC 28 nm chip roadmap Non-Red · zero Chinese components Edge AI · offline inference JANMAN ISO 13485 / GMP · 56 patents ODM / OEM · open SDK
The challenge

Drone imaging is two problems at once.

The camera must survive the field — and it must see through haze, glare and darkness and decide in real time, at flight speed, without your imagery ever leaving the aircraft.

Extreme conditions

High-altitude haze, backlight, 0.1 lux nights and 150 km/h motion blur defeat ordinary cameras. You need adaptive HDR, dehaze and motion compensation — computed live.

Latency, power & security

Cloud round-trips add delay and leak data; drones are power- and weight-limited. You need low-latency inference on the edge, at low power, with the data kept on board.

One product, two disciplines

The AI brain and the rugged airframe — from one partnership.

AI Angel (CTE Group) provides the intelligence; JANMAN Precision provides the structure. Together, one complete, fully non-Chinese package.

The intelligence — AI Angel / CTE Group

Sees in the dark

  • AI ISP: real-time HDR, denoise, dehaze, scene recognition on the edge
  • Secure OTA — signed, rollback-protected model updates
  • Federated Learning — raw imagery never leaves the aircraft
  • Built non-red: Lattice FPGA today → dedicated 28 nm AI ISP chip
The structure — JANMAN Precision (建名精密)

Survives the field

  • Aerospace-grade PEEK / CFRTP rugged housings; anti-vibration, IP-rated
  • One-stop: mould → injection → metal insert → assembly
  • Precision to 0.0005 mm · 56 patents · founded 1985
  • ISO 9001 / 14001 / 13485 · MoHW GMP · LEED green factory
Cinematic film

See the vision in motion.

A short cinematic film — non-red drone AI, from the edge to the sky.

Play the AI Angel cinematic film
Click to load · best viewed with sound on
The technology

A full AI ISP engine — algorithm and silicon, co-designed.

Eight tightly-integrated capability areas turn a raw, noisy sensor feed into clean, decision-ready imagery — efficient enough to run on the edge.

🧩

Architecture & AI accelerator

AI accelerator + ISP pipeline + memory subsystem co-designed; tensor & convolution engines, low-latency data paths, configurable pipeline.

🌅

ISP core algorithms

Multi-exposure HDR, motion compensation, spatial/temporal denoise, dehaze, local tone mapping, colour processing.

🧠

AI models & optimisation

Lightweight CNNs, INT8 quantisation, knowledge distillation and hardware-aware training — big-model accuracy at edge cost.

⚙️

Hardware acceleration

Dedicated ISP accelerators — feature detection, motion-estimation and filter arrays, colour-transform engines.

🔋

Memory & low power

Hierarchical cache, DMA and bandwidth optimisation; DVFS, clock-gating and power-domain design for < 2.5 W on 28 nm.

🛡️

Security & firmware

Root-of-Trust secure boot, AES-256 hardware encryption, TEE, device drivers and control logic — offline by design.

System architecture

One integrated stack — from the sensor to the cloud.

Image quality, edge AI and secure updates are co-designed as a single system. The drone stays fully autonomous in the field; the cloud keeps the whole fleet improving over the air.

100% Non-Red Lattice FPGA On-fabric AI inference Federated learning · data stays local OTA A/B switch < 50 ms On-chain provenance
① Cloud & OTA platform
Autonomous agent

🤖 Self-managing OTA

An AI agent continuously watches live quality metrics — PSNR, SSIM, latency, accuracy — and decides on its own when to push, upgrade or roll back a model. No truck-roll, no downtime.

Platform

☁️ Model & fleet management

Version control, A/B deployment, knowledge-distillation training and a live dashboard across the entire fleet.

Data & trust

🔐 Federated & on-chain

Only de-identified gradients ever leave a device — raw imagery never does. Every deployment is written to a blockchain ledger for a tamper-proof audit trail.

Signed, encrypted gateway between cloud and fleet — OTA channel over 4G, zero raw imagery in transit ▼
② Edge device · on the drone
Image input

📷 Sensors → FPGA

Multi-camera RAW feed streamed straight into the FPGA over MIPI CSI-2 — no host PC in the loop.

Image engine · core IP

🧩 Lattice FPGA image engine

A full 9-stage ISP and a 3-tier AI inference stack run entirely in FPGA fabric — the proprietary core that becomes our dedicated 28 nm AI ISP chip.

9-stage ISP pipeline
1Black level
2Digital gain
3Defect fix
4Demosaic
5HDR
6Colour matrix
7Colour space
8Gamma
9AI enhance
3-tier edge AI inference
Tier 1 · Scene controlAdapts the ISP to the scene in real time · < 1 ms
Tier 2 · EnhanceSuper-resolution + denoise + HDR fusion · < 3.5 ms
Tier 3 · PerceptionObject detection + depth, run asynchronously · ~12 ms
Real-time controller · secure OTA

⚙️ On-device flight brain

A dedicated real-time controller drives the FPGA and manages A/B firmware: it evaluates each new model against the quality metrics, switches atomically in < 50 ms, and rolls back automatically if anything regresses — the drone keeps flying through every update.

Three independent radio links
2.4 GHz · ELRS
Flight control (C2)Low-latency command & control, 10 km+ range.
4G LTE
Backup + OTAModel updates, telemetry and image upload — kept off the control link.
5.8 GHz
FPV videoPilot's live feed on its own link, so nothing competes for bandwidth.
Performance

Proven on FPGA today — and the chip we're building next.

We keep demonstrated results and design targets clearly separate.

Proven today · on FPGA
AI Angel edge platform
Real-time edge denoise16 fps
Noise reduction−42%
Image qualityPSNR ≥ 30 dB · SSIM ≥ 0.92
OTA A/B switch + rollback< 50 ms
Supply chain100% non-red
Raw data leaving devicenone
Measured on-site — the same figures we accept a system against.
Roadmap

From FPGA validation to a dedicated chip — and beyond.

2026 ● NOW

Validated & going to market

FPGA platform proven; JANMAN partnership in place; first Sample-Kit pilots with drone makers, and a full SDK for partners.

2027

Chip & first integrations

AI ISP tape-out and engineering samples; first ODM/OEM integrations and Beta-site validation.

2028

Scale

Volume production; broaden from inspection into mobile-vehicle vision and infrastructure.

2029+

Next generation

12 nm process, 4–8 TOPS; expansion into ADAS, smart-city and Industry-4.0 vision.

Where it's used

Built for the toughest inspection & field missions.

🏭

Indoor facility inspection

  • Narrow-space navigation (30 cm)
  • Low-light imaging (0.1 lux)
  • 24/7 autonomous patrol
  • Real-time anomaly detection
Est. addressable market ~ NT$680 M (2026)
🚙

Mobile-vehicle vision

  • Industrial anti-vibration (< 0.5 px)
  • Harsh environment (IP67)
  • Multi-sensor fusion
  • High reliability (MTBF > 50,000 h)
Est. addressable market ~ NT$380 M (2026)
🗼

Critical-infrastructure monitoring

  • End-to-end encrypted video
  • Edge-AI real-time processing
  • Long-term stable operation
  • Early-warning integration
Est. addressable market ~ NT$100 M (2026)
Trust & sovereignty

Non-red by design. Data stays yours.

For critical infrastructure and national programmes, the supply chain and the data both matter.

100% Non-Red supply chainZero Chinese components across silicon and hardware — suitable for government procurement.
Data sovereigntyFederated Learning & offline inference — raw imagery never leaves the aircraft or local servers.
Hardware securityAES-256 hardware encryption, Root-of-Trust secure boot and a Trusted Execution Environment.
Signed, auditable OTAEvery model update is signed, rollback-protected and logged.
The company

Two decades of imaging silicon — now aimed at drones.

AI Angel · CTE Group

Part of CTE Group (founded 2010, Taipei) with deep heritage in image-signal-processing and camera-module design across mobile, wearable, automotive and security — and ASIC & FPGA imaging experience. AI Angel focuses that expertise on non-red drone edge-AI, OTA and a dedicated AI ISP chip, delivered ODM/OEM with an open SDK to empower local manufacturing.

JANMAN Precision · 建名精密

An aerospace-grade precision manufacturer since 1985 — rugged PEEK/CFRTP housings, precision mechanics and one-stop mould-to-assembly production, with 56 patents and ISO 13485 / GMP certification. JANMAN provides the structure that keeps the optics stable in the field.

Get in touch

Let's talk about your drone programme.

Tell us your use case and we'll come back with a concrete next step — a demo, a Sample-Kit pilot, or an ODM/OEM discussion.

Email

Benison.pin@aiangel-ctegroup.com.tw
benison.pin@gmail.com

Phone

+886-902298875

Company

AI Angel · CTE Group (晶芯電子), Taiwan
in partnership with JANMAN Precision (建名精密)

Contact

Benison Pin (賓振呈) — CTO

Website

www.aiangel-ctegroup.com.tw