Photo Finish Camera Product

Overview

A photo finish camera is a specialized imaging system used in competitive athletics to resolve extremely close race finishes. While human timekeepers can determine winners to the nearest 0.1 second, elite sprints often feature simultaneous finishes to within 1–10 milliseconds. A photo finish camera captures a continuous line-scan image perpendicular to the finish line, recording which runner's torso crosses the line first at microsecond precision.

The system comprises a high-speed line-scan sensor, precision timing synchronization (typically GPS-disciplined), real-time frame capture, and GPU-accelerated edge detection software that automatically ranks finishers by comparing their leading edges. Photo finish is now standard in Olympic Games, World Championships, and professional track meets.

How it works

Line-Scan Imaging

Unlike conventional cameras that capture full 2D frames, the [[photo-finish-camera-sensor-assembly|sensor assembly]] uses a linear array (4096 pixels) that scans a 1D slice perpendicular to the finish line at extremely high speed. The [[image-sensor|CMOS line-scan sensor]] reads pixels at 8 kHz, meaning it captures a complete horizontal "slice" across the finish line every 125 microseconds.

Each slice is spatially precise: the [[camera-lens|lens assembly]] projects the finish line onto the 4096-pixel array, giving a spatial resolution of approximately 1 mm per pixel (depending on focal length and finish-line width). Temporally, the 8 kHz line rate means that a runner's crossing is recorded with 125 μs precision in time.

As the race event unfolds (30–100 seconds), thousands of sequential line scans are captured, building a 2D image showing the finish line over time. The vertical axis (built up from sequential lines) represents elapsed time, and the horizontal axis represents position across the finish line.

Timing Synchronization

The critical challenge is synchronizing the camera's line-scan readout with an accurate global time reference. The [[photo-finish-camera-timer-sync|timer sync module]] receives a 1 PPS (pulse per second) signal from a [[photo-finish-camera-gps-receiver|GPS receiver]]. This GPS timing is accurate to within 100 nanoseconds, allowing multiple cameras and timing systems at a venue to share a common time reference.

A [[photo-finish-camera-reference-oscillator|disciplined oscillator]] is locked to the GPS 1 PPS signal, producing a stable 25 MHz clock that drives the [[photo-finish-camera-trigger-unit|trigger unit]]. This clock generates precise line-start pulses that trigger each line-scan readout. The [[mcu|microcontroller]] in the trigger unit also timestamps each line with GPS-synchronized time (e.g., 14:35:42.003847 UTC).

Image Capture and Processing

The [[photo-finish-camera-capture-station|capture station]] uses a [[photo-finish-camera-frame-grabber|high-speed PCIe frame grabber]] to pull raw pixel data from the sensor at 32 MB/s (for 8-bit capture of 4096 × 1 lines at 8 kHz). The data is written directly to the [[photo-finish-camera-ssd-storage|NVMe SSD]] for real-time archival. This approach guarantees no frames are lost even during peak activity (multiple simultaneous finishes).

After the race (or in real-time), the [[photo-finish-camera-processing-unit|processing unit]] analyzes the finish image using GPU-accelerated algorithms:

1. Edge Detection: A [[photo-finish-camera-gpu-accelerator|CUDA-capable GPU]] scans each column (runner position) to find the leading edge (torso) of each participant's image.
2. Time Assignment: For each runner's leading edge pixel, the GPU reads the timestamp from the corresponding line.
3. Ranking: Runners are ranked by comparing the time at which their edge crossed the finish line (within milliseconds).
4. Manual Review: Officials review the image on a high-resolution display, confirming the automatic ranking and making appeals decisions if needed.

Lens and Exposure Control

The [[photo-finish-camera-lens-module|lens module]] uses a variable [[camera-lens|6–25 mm focal length lens]] or a fixed 8 mm, mounted with a precision kinematic [[photo-finish-camera-lens-bracket|bracket]] for X/Y/Z adjustment. The lens is positioned perpendicular to the finish line at approximately 3–5 meters away, viewing a 10–20 meter wide section that includes the finish line and approach zone.

A motorized [[photo-finish-camera-aperture|iris aperture]] (f/1.4–f/16) allows the operator to control exposure without adjusting shutter speed (the line-scan sensor has no conventional shutter). A [[photo-finish-camera-filter-mount|rotating filter turret]] can hold neutral-density or polarizer filters to manage bright outdoor sunlight or glare off the track surface.

Multi-Camera Synchronization

For larger events (relays, team sports), multiple photo-finish cameras can be installed at different positions (finish line, 100 m mark, etc.). Each camera is synchronized to the same GPS 1 PPS reference via the [[photo-finish-camera-distribution-amplifier|timing distribution amplifier]]. This allows cross-event timing validation (e.g., confirming that a relay baton exchange occurred legally by comparing hand-off camera images across multiple angles).

Installation and Setup

A photo finish system is typically installed 5–10 meters before the actual finish line, mounted on a stable support (tripod or gantry). The camera is positioned at roughly 1.5–2 meters height, perpendicular to the expected path of runners. The lens is focused on the finish line with sharpness verified using a target card at race-start distance.

The capture station (computer with frame grabber and SSD) is placed in a weatherproof enclosure near the camera, or inside an adjacent timing booth. Power requirements are modest (500 W peak), running from standard 110/240 VAC mains.

GPS antenna is mounted on the roof of the timing booth or on the camera gantry, with a clear sky view (unobstructed 10+ degrees above horizon). Once GPS lock is achieved, the timing offset between the camera's local clock and GPS is recorded, allowing time-synchronization even if GPS signal is temporarily lost during the event.

Pre-Race Verification

Before competition, the timing system undergoes a verification procedure:

1. A test runner sprints past the camera at known speed, and the finish image is checked for focus sharpness and adequate exposure.
2. Timing accuracy is verified by recording a manual sync pulse (e.g., a strobe flash at known GPS time) and confirming the camera's timestamp matches within 1 μs.
3. Storage capacity is checked (1–2 TB is sufficient for 1–2 hours of continuous capture).
4. GPU software is tested with a replay of a previous race to confirm edge-detection accuracy.

Post-Race Analysis

After the race, the finish image is available within seconds. Race officials review the image, checking:

• Are leading edges clearly visible and distinct?
• Are timestamps accurate and properly recorded?
• Are the rankings consistent with photo edge positions?

If there are objections from runners (e.g., apparent false touch-screen or pacing error), officials can zoom, pan, and measure pixel positions with sub-millimeter precision, resolving disputed finishes. The complete image archive is retained for appeals processes, typically 30 days post-event.

Modern meets also project the finish image on large venue screens for spectators, showing the dramatic photo-finish ranking in near-real-time.