SportVision

Real-time sports analytics toolkit built on the Roboflow ecosystem. Detects players and ball with any COCO-compatible model (YOLOv8, RF-DETR, etc.), tracks with ByteTrack, classifies teams by jersey color, and computes possession, speed, distance, and heatmaps.

Features

• Detection — Works with any COCO-compatible detector (YOLOv8, RF-DETR, etc.)
• Tracking — ByteTrack via supervision with fallback sequential IDs
• Team Classification — KMeans on HSV jersey histograms
• Homography — Pixel→field coordinate mapping via cv2.findHomography
• Analytics — Possession, speed, distance, and heatmap generation
• Annotation — Team-colored bounding boxes, stats overlay, player trails

Quick Start

pip install sportvision

from sportvision.pipeline import SportVisionPipeline

pipeline = SportVisionPipeline()
result = pipeline.process_frame(frame)

Detection needs a backend: pip install "sportvision[inference]" for RF-DETR, or install ultralytics for YOLO. Without one the pipeline returns empty detections and logs a warning once.

Two knobs make the pipeline actually real-time on a GPU. detect_every=N runs detection every Nth frame and reuses the last result in between. infer_size caps the detector's input size (default 640) so the model sees fewer pixels; detection quality barely moves and boxes are mapped back to full frame coordinates.

pipeline = SportVisionPipeline(device="auto", detect_every=3, infer_size=640)

Roboflow Workflows Plugin

SportVision ships as a Roboflow Workflows plugin. Install with inference and activate:

pip install "sportvision[workflows]"
export WORKFLOWS_PLUGINS="sportvision.workflows"

This registers 4 blocks you can use in any Roboflow Workflow:

Example: Using blocks directly in Python

import cv2
import numpy as np
import supervision as sv

from sportvision.workflows.team_classifier.v1 import TeamClassifierBlockV1
from sportvision.workflows.possession_tracker.v1 import PossessionTrackerBlockV1
from sportvision.workflows.distance_calculator.v1 import DistanceCalculatorBlockV1
from sportvision.workflows.sports_detection_filter.v1 import SportsDetectionFilterBlockV1

# --- Filter COCO detections to sports classes ---
det_filter = SportsDetectionFilterBlockV1()
# Assume `raw_detections` comes from a COCO model (person=0, sports_ball=32)
result = det_filter.run(detections=raw_detections)
detections = result["detections"]  # now player=0, ball=1

# --- Classify players into teams ---
team_block = TeamClassifierBlockV1()
# `image` must have a .numpy_image attribute (or use WorkflowImageData)
# refit_every=10 refits KMeans every 10 frames (0 = fit once, default)
result = team_block.run(image=image, detections=detections, n_teams=2, refit_every=10)
detections = result["detections"]  # detections.data["team_id"] is now set

# --- Track possession ---
possession_block = PossessionTrackerBlockV1()
result = possession_block.run(
    detections=detections,
    ball_class_id=1,
    ball_proximity_threshold=100.0,
)
print(result["possession_stats"])   # {0: 0.6, 1: 0.4}
print(result["possessing_team"])    # 0
print(result["warning"])            # "" or warning if team_id missing

# --- Compute distances ---
distance_block = DistanceCalculatorBlockV1()
# Optional: pass a 3x3 homography matrix for field-unit distances (e.g. meters)
result = distance_block.run(detections=detections, homography_matrix=[[0.01,0,0],[0,0.01,0],[0,0,1]])
print(result["detections"].data["distance"])  # cumulative distance per tracker

Example: Workflow JSON definition

{
  "steps": [
    {
      "type": "sportvision/sports_detection_filter@v1",
      "name": "filter",
      "detections": "$steps.model.predictions"
    },
    {
      "type": "sportvision/team_classifier@v1",
      "name": "teams",
      "image": "$inputs.image",
      "detections": "$steps.filter.detections",
      "n_teams": 2,
      "refit_every": 10
    },
    {
      "type": "sportvision/possession_tracker@v1",
      "name": "possession",
      "detections": "$steps.teams.detections",
      "ball_proximity_threshold": 100.0
    },
    {
      "type": "sportvision/distance_calculator@v1",
      "name": "distance",
      "detections": "$steps.teams.detections",
      "homography_matrix": [[0.01,0,0],[0,0.01,0],[0,0,1]]
    }
  ]
}

Try it on Colab

Architecture

src/sportvision/
├── detection.py      # SportsDetector — wraps COCO detectors, maps to sports classes
├── tracking.py       # SportsTracker — ByteTrack via supervision
├── teams.py          # TeamClassifier — KMeans on HSV jersey histograms
├── homography.py     # FieldHomography — pixel→field coords
├── analytics/
│   ├── possession.py # PossessionTracker — nearest-player-to-ball per frame
│   ├── speed.py      # SpeedEstimator — displacement/time → km/h
│   ├── distance.py   # DistanceCalculator — cumulative path length
│   └── heatmap.py    # HeatmapGenerator — 2D histogram + gaussian blur
├── annotators.py     # TeamColorAnnotator, StatsOverlayAnnotator, TrailAnnotator
├── pipeline.py       # SportVisionPipeline — orchestrates all modules
└── workflows/            # Roboflow Workflows plugin
    ├── _compat.py        # Inference compatibility shim
    ├── kinds.py          # Custom kind definitions
    ├── team_classifier/  # Team classification block
    ├── possession_tracker/   # Possession tracking block
    ├── distance_calculator/  # Distance calculation block
    └── sports_detection_filter/  # COCO→sports filter block

Sports Class IDs

Dependencies

Development

git clone https://github.com/MohibShaikh/sportvision.git
cd sportvision
pip install -e ".[all]"

# Tests
pytest tests/ -v

# Lint
ruff check src/ tests/ && ruff format --check src/ tests/

License

Apache-2.0