// Technology

2 cameras. 0 sensors. All the game.

Optical tracking is the gold standard in professional sport. Futnow makes it accessible: fixed installation, AI models trained on futsal, data delivered in real time.

// 01 — The pipeline

From light to decision.

Futnow CV Engine Detection & identification Output: x, y pitch positions
  1. Synchronized capture

    Two fixed cameras, calibrated to the centimetre, cover the entire 40 × 20 m court. Their clocks are synchronized to fuse both viewpoints.

  2. Detection & identification

    On every frame, our neural networks detect players and the ball, then maintain each player's identity through occlusions, clustering and futsal's constant rotation.

  3. Pitch reconstruction

    Detections from both cameras are triangulated into pitch coordinates (x, y). Trajectories are smoothed and physically validated—no teleportation.

  4. Metrics & broadcasting

    Distances, speeds, accelerations, collective structures: everything is calculated continuously and pushed to the live dashboard and API.

// 02 — The installation

Discreet. Fixed. Forgotten.

01

2 cameras per court

Mounted high, out of reach of play and stray balls. Built for artificial lighting and the contrast of futsal.

02

Installed in 2 hours

No court modifications. Power, network, calibration—the system is live.

03

Zero player load

No vest, no sensor, no kit to hand out. Players play.

04

Every court

Full-size or reduced court, multi-sport complex: models calibrate to the court's actual dimensions.

05

European hosting

Data hosted in Europe, GDPR-compliant, data ownership with the club.

06

API & exports

REST API, CSV and PDF exports. Your analysts build their own models on our raw data.

// Fixed cameras, zero sensors

The players play. The system does the rest.

Field-tested

Built for futsal, trained on futsal.

Futsal is computer vision's most demanding environment: players clustered tight, execution tempo, constant line switches, irregular artificial lighting, reflective flooring. That's where our models were built—and why they hold up.

  • Models trained on thousands of futsal matches, from semi-pro to elite
  • Robust to occlusion, clustering and rapid rotation
  • Continuous improvement—each match makes the system better
Meet the team →
Player silhouettes detected by Futnow's computer vision engine

Test it on your matches.

Free feasibility study, from a layout plan or photos.