Product
Ball Track / Cric DRS
Frame-accurate cricket ball tracking that reconstructs a delivery from a standard broadcast feed.
In deployment
The problem
What's broken today
Broadcast-grade ball tracking has been proprietary, tied to a dedicated multi-camera rig, and priced for the top tier of the sport. Every production below that tier settles close calls with a slow-motion replay and an opinion.
What it does
Ball Track / Cric DRS, in practice
Ball Track detects the ball in every frame of a standard broadcast feed, fits those detections into one continuous flight path, derives the pitch point and the point of impact, and renders DRS-style review output while the delivery is still live. The whole pipeline runs on a single GPU node at the ground.
Watch it work
See Ball Track in motion
Illustrative placeholder
Illustrative only, not live broadcast footage. The real Ball Track demo replaces this frame at launch.
How it works
The pipeline behind Ball Track / Cric DRS
Reads the broadcast feed
Ball Track takes the camera feed a production already generates, rather than requiring a dedicated multi-camera rig installed at the ground.
Detects the ball in every frame
A GPU inference pipeline locates the ball frame by frame, including the frames where it is small, motion-blurred, or partly hidden behind a player.
Reconstructs the flight path
Per-frame detections are fitted into one continuous trajectory, from which the pitch point and the point of impact are derived.
Renders the review
The reconstructed path and predicted continuation are rendered as DRS-style review output while the delivery is still live.
Features
What Ball Track / Cric DRS does
Frame-level ball detection
Locates a small, fast-moving object in each individual frame, which is a harder detection problem than tracking a person or a vehicle.
Trajectory reconstruction
Joins per-frame detections into a single flight path rather than a scatter of independent positions.
Pitch and impact points
Derives where the ball landed and where it struck pad, bat, or stumps from the reconstructed path.
Predicted continuation
Projects the path forward past the point of impact, which is the output an umpire review actually turns on.
Standard-feed input
Works from ordinary broadcast camera angles, removing the specialist camera installation that has kept this technology to the top tier of the sport.
Real-time output
Produces review graphics inside the natural break in play rather than as a post-match render.
GPU inference pipeline
Detection, reconstruction, and rendering run as one accelerated pipeline on a single GPU node.
Deployment options
Where Ball Track / Cric DRS runs
Every option below is the same software on a different surface. The deployment model changes where it runs, not what it does.
At the ground
Runs on a GPU node in the production truck or ground control room, next to the feed it reads. No frame leaves the venue.
Private cloud tenant
Deploys into a broadcaster's own cloud subscription for productions that centralise their video pipeline rather than running it per venue.
Edge node
A single self-contained appliance for grounds with limited uplink, running the full pipeline locally and sending only the rendered output onward.
Where it stands
Where Ball Track / Cric DRS stands today
In deployment
Ball Track is in deployment. Its detection and reconstruction pipeline is built on GPU-accelerated inference developed with NVIDIA Inception program support, engineered for frame-by-frame precision under real broadcast conditions rather than on clean test footage.
FAQ
Questions we get about Ball Track / Cric DRS
What is AI ball tracking for cricket?
AI ball tracking uses computer vision to follow the ball frame by frame across a delivery, reconstructing its full flight path from standard broadcast camera angles - the technology behind ball-by-ball trajectory graphics and DRS-style review.
How accurate does DRS-grade ball tracking need to be?
DRS-grade tracking needs frame-level precision on a small, fast-moving object, often from partial or obstructed camera views. Ball Track is built specifically for that problem, which is a much harder computer vision task than tracking a person or a vehicle.
Can Ball Track work with standard broadcast camera feeds?
Yes. Ball Track is designed to run on standard broadcast feeds rather than requiring a dedicated multi-camera rig, which is what has historically made broadcast-grade ball tracking expensive and hard to access.
Is Ball Track used for umpire decision review?
Ball Track reconstructs ball trajectory and delivers DRS-style output in real time - the same category of system used to support on-field umpire decision review in broadcast cricket.
Why did FalcRise start with cricket ball tracking?
Tracking a small, fast object this precisely is one of the harder problems in computer vision, and the pipeline that solves it is the same GPU inference layer the rest of the platform runs on. Ball Track is the hardest test of that layer, which is why it was built first.
Ready to see Ball Track / Cric DRS for yourself?
Frame-accurate cricket ball tracking that reconstructs a delivery from a standard broadcast feed.