FPV AI Camera Module VisionCube S Kit for Drones, 450m Vehicle/170m Human Recognition, CRSF, PiP, 1 TOPS, 1080p@30Hz

Overview

The FPV AI Camera Module VisionCube S kit is a drone accessories solution designed for FPV integration. This FPV AI Camera Module combines a visible-light camera with an AI image processing board to deliver automatic guidance, tracking, target identification, lock-on, and strike assistance for FPV drones. It supports human and vehicle recognition, multi-target tracking, and picture-in-picture (PiP) viewing.

Key Features

• Default target types: human and vehicle recognition.
• Target detection range: Vehicle 450m; Human 170m.
• Multi-target recognition: identifies up to 50 targets within visual range.
• Minimum tracking pixel: 16×16 pixels.
• Dynamic memory locking: briefly occluded targets can be recaptured within 3 seconds and tracking continues.
• Adaptive adsorption (lock-on) tracking to reduce operation difficulty.
• Picture-in-picture architecture with an 8× electronic magnification sub-display.
• AI image processing board computing power: 1 TOPS.
• Protocol: CRSF; supported firmware: BetaFlight/ArduPilot.
• Visible-light camera: 1/2.8-inch sensor, 4 mm focal length, 1920×1080@30Hz input.
• Maximum tracking speed of dynamic targets: 60 km/h.

Specifications

What’s Included

• AI VisionCube S × 1
• SH1.0–12PIN 15 cm cable × 1 (for connecting flight controller and AI VisionCube kit)
• Cross recessed pan head screws M2×4 × 2
• Hexagon socket head cap screws M2×6 × 4

Applications

• FPV drones requiring automatic guidance, tracking, and target recognition.
• Human/vehicle detection and multi-target tracking with PiP viewing.

Details

AI VisionCube S 2025 Global Debut Next-Gen Hyper-Sensing AI Module

FPV AI Camera offers ultra-long-range, multi-target recognition, PiP, adaptive and continuous tracking, lightweight design. (18 words)

Lightweight Integration Design presents a modular concept that integrates BetaFlight/ArduPilot open-source flight controllers. The design supports deployment within 5 minutes via plug-and-play installation, suitable for industrial drones to FPV. The High-Definition AI Camera module and Processing Board are also featured.

450M over-distance recognition, tracks 50 targets, AI camera drone

Multi-target recognition with YOLOv7-optimized architecture tracks over 50 dynamic objects simultaneously.

Dynamic memory locking tracks moving targets, recaptures within 3 seconds after brief occlusion using trajectory memory.

Adaptive adsorption tracking reduces operation difficulty, featuring real-time vehicle monitoring and smart path adjustment.

The Picture-in-Picture Architecture displays the big picture on the main screen and zooms in on details with an 8x electronic magnification.

FPV AI Camera specifications: Four models (S, D, ST, DT) with target recognition for humans and vehicles, detection ranges up to 1200m, 16x16 pixel tracking, 60km/h speed support, crosshair fuzzy locking, trajectory prediction, memory tracking, up to 50 targets, and PIP support.

AI Image Processing Board with 1 or 6 TOPS computing power, CRSF protocol, supports BetaFlight/ArduPilot. Features two visible cameras with varying sensors, focal lengths, FOV, low light performance, and video input formats. Dimensions and mounting holes specified for each model.

FPV AI Camera: 12mm focal length, 26°×15° FOV, 9650mV/lux·s sensitivity, 1080p@60Hz, 40.8×25×26mm. Thermal camera: 12μm pitch, 9.1mm lens, 20.3°×15.2° FOV, 8–14μm range, 384×288@25Hz, USB, 26×26×42.63mm.

Packing list includes AI VisionCube S Kit, 2 M24 cross slot pan head screws, 4 M26 hexagon socket screws, and 1 SH1.0-12PIN 15cm ribbon cable for flight control connection.