verified Patent Pending · PAT 63/923,797
Coolsquid Defense Product

Vector

Range

Real time projectile impact detection and scoring using computer vision. Know where every round landsinstantly.

System Active · 0 FPS
The Problem

Fire. Wait. Walk. Repeat.

Standard range protocol: fire at a target, wait for the range officer to call the line cold, walk downrange to inspect your target, then try to remember what you felt when the shot broke. By the time you see the result, the feedback loop is broken.

Is the shooter anticipating recoil? Pulling the trigger too fast? Does the optic need adjustment? These questions can't be answered from memory. They need to be answered in real time, at the moment of the shot.

close
Existing Approach

Manual inspection. Broken feedback loop. Slow iteration. No performance data. High range time waste.

check_circle
Vector Range

Instant impact location. AI driven improvement feedback. No range cold required. Works with any target. No wires to shoot off.

Deployment Scenarios

military_tech Military & Training

Range Training & Qualification

Service members receive immediate AI driven feedback during drillswithout waiting for range cold, without manual intervention from a range officer. Performance data is captured in real time, enabling faster skill development and objective qualification scoring.

  • chevron_right Range officers can monitor multiple shooters simultaneously from a single dashboard
  • chevron_right Qual tests scored digitallyrange officer verifies, not manually tallies
  • chevron_right AI identifies technique issues: recoil anticipation, trigger pull timing, grouping drift
  • chevron_right Multi target detection from a single wide-FOV camera
Hardware Setup
videocam

~4 cameras mounted above the range, looking down at targets. Each shooting lane gets an iPad for the shooter's live view. Simple retrofit. No target modification required.

phone_iphone Civilian & Recreational

Smartphone App

Available to any shooter, at any range, with any target. Mount your phone to a spotting scope, open the app, and start shooting. Vector Range tracks every impact in real time, shows grouping data, and gives actionable coachingall without replacing anything you already own.

  • chevron_right Works with any standard targetno proprietary hardware required
  • chevron_right Tested at 10 to 100 yards across a wide range of calibers
  • chevron_right Real time grouping analysis and shot by shot history
  • chevron_right Native iOS and Androidpotential OEM bundling with optic manufacturers
Supported Calibers
.22LR .22MAG .223 9MM .357 12GA Slug 7.62x54r

How It Works

// YOLOv8 Dual Model Architecture
01
Stage One

Target Detection

The first YOLOv8 model locates and frames every target in the camera's field of view. Trained on ~1,000 images, it achieves 80%+ real-world accuracy across varying backgrounds, lighting conditions, and target types. When multiple targets are presentas in military range configurationsall are detected and tracked independently.

Each detected target establishes a coordinate space for impact calculations: distance from center, scoring ring intersection, and multi target grouping comparison.

Model Stats
80%+
Real-World Accuracy
~1,000
Training Images
Any
Standard Target Type
Target detection model — bounding box annotation
FIG. 2 — Target detection training sample
02
Stage Two

Bullet Impact Detection

The second model identifies and localizes individual bullet impacts. Trained on ~5,000 targetsincluding dirty splatter targets captured at outdoor ranges from 10 to 100 yards across multiple spotting scopes and weather conditionsit achieves ~70% accuracy across the supported caliber set.

Data augmentation during training ensures reliable performance on inputs that are scratched, off center, tilted, or captured under inconsistent zoom levels. The model handles real-world range conditions, not just clean lab data.

Model Stats
~70%
Impact Detection Accuracy
~5,000
Training Images
10–100
Yards Tested Range
Bullet impact detection — annotated training sample
FIG. 3 — Impact detection training sample
YOLOv8 training performance metrics
FIG. 5 — Training performance metrics
03
Stage Three

Combined Real Time Output

Both models run in parallel to deliver a unified real time overlay: current impact location, last known impact marker, running grouping size, and session historyall rendered at 30 FPS. Tested on iPhone 13 mini and iPhone Air, the system runs on consumer hardware with no external compute required.

The civilian app deploys via iOS and Android. The military system pairs with dedicated edge hardware for air gapped, low-latency range environments.

Performance
30 FPS
Real Time Processing
Tested Devices
iPhone 13 mini
iPhone Air
Deployment
iOS · Android · Edge Hardware

Why Vector Range Wins

Feature Vector Range ShotMarker USMC Smart Target
Any target type check close close
No wires check close close
No shootable components check close close
AI improvement coaching check close close
Grouping analytics check close close
Multi target from single camera check close close
Low installation cost check close close
Consumer smartphone app check close close
verified Patent Pending · PAT 63/923,797

Ready to deploy
Vector Range?

Whether you're outfitting a training facility or building optic integrationsreach out for a full technical briefing.

mark_email_read Request a Briefing