ORGSONAR → 1ST-PLACE EVERBRIDGE FREAKY FRIDAY → 1909 VERMONTVILLE INSPIRATION → AI + MAVIC 2 PRO

From Fictitious Enterprise to Live In-Flight Software: Progressive Focus Dissolve

To demonstrate OrgSonar (pairing formal compliance milestones with continuous, day-to-day operational health signals across teams), I wanted to invent a mission-driven company inspired by working at organizations like Everbridge (whose Critical Event Management platform helps public safety agencies and enterprises keep people safe during emergencies). Back at Everbridge, teams ran Freaky Friday projects to pitch cool new features for the CEM platform; I pitched real-time aerial overlay maps for search-and-rescue and first-responder events, and my team won first place. Drawing historical inspiration from a 1909 freight train collision in Vermontville, Michigan, I created the fictitious 1,800-employee company Aeris Overwatch Systems, and years later partnered with AI and grabbed my consumer-grade DJI Mavic 2 Pro outside by my house to bring to life what I could only imagine back then.

1st-Place Everbridge Freaky Friday Concept

0 Post-Processing Live In-Flight RTMP Map

Consumer Mavic 2 Proof of Concept

2.68 cm/px Nadir Belly Clarity · 0 Snap

Act I: Origin Story (OrgSonar → 1909 Inspiration → Aeris Overwatch → Live Proof of Concept)

How a 1st-place hackathon idea, a 1909 rail collision clipping, and AI helped turn a fictitious 1,800-person company into live aerial software.

ORIGIN 01 / ORGSONAR LIVE SONAR ORBIT

Continuous Operational Health Signals Across Teams

In an earlier LinkedIn post I shared OrgSonar. Making that demonstration credible meant modeling a complete enterprise (10 departments, 39 teams, 65 services, and 1,800 employees) around a noble public-safety mission I’d genuinely want to join.

ORIGIN 02 / CONNECTED SERVICES & EVERBRIDGE CEM

1st-Place Freaky Friday Innovation Concept

At Everbridge, teams ran Freaky Friday projects to pitch cool new features for the Critical Event Management platform. I pitched real-time Search & Rescue aerial overlay maps, and my team won first place.

The Vermontville Echo - Oct. 13, 1909 Freight Train Collision

ORIGIN 03 / THE VERMONTVILLE ECHO (OCT. 13, 1909)

Historical Rail Wreck Reimagined Today

This historic 1909 Michigan Central freight train collision in Vermontville inspired the scenario: reimagining how modern responders arriving at a wreck like that today would need instant live aerial overlays of blocked roads and perimeters.

01 / MAP THE UNKNOWN

Yesterday’s Map → Today’s Reality

Satellite tiles are months or years old. High-altitude survey passes stream live RTMP frames and dissolve just-in-time ground overlays directly into place over the stale basemap.

02 / HOLD THE HORIZON

Cue Dedicated Live-Video Drones

Once the just-in-time orientation overlay reveals blocked lanes, damaged structures, or shoreline hazards, secondary aircraft lock onto priority sectors with continuous live video feeds.

03 / MAKE THE INVISIBLE VISIBLE

Layer Sensor & Response Overlays

With accurate just-in-time geometry underneath, commanders fuse downwind atmospheric plume envelopes, environmental sensor contours, traffic detours, and responder staging on one picture.

Bringing Imagination to Life With AI: Phase 01 on that fictitious site promised: “Watch the stitched derailment imagery dissolve into place.” Even on expensive enterprise RTK drones, aligned maps usually require landing and sending photos back for offline or cloud post-processing rather than appearing live out of the box from an RTMP video stream. Years later, AI helped me bring that concept to life as a real proof of concept on my consumer-grade Mavic 2 Pro by resolving six cascading physics & UX hurdles in software.

Act II: 6 Real Flight Clips (Cause & Effect Engineering Walkthrough)

Click each stage (or press 1–6) to watch the actual Mavic 2 flight & map-build clips pointing directly at each technical hurdle.

1. Technical Challenge

2. Root Physical Cause

3. Effect on Live Map

4. Engineering Fix

Act III: Rendered 1080p LinkedIn Demonstration Video & Post Kit

Ready to upload natively to LinkedIn (1920×1080 H.264 MP4 with burned-in story captions & real Mavic 2 aerial footage).

⬇ Download 1080p First-Person LinkedIn Video (.MP4) ✨ Inspect Stage 6 Close-Up Dissolve & Wipe

Ready-to-Post First-Person LinkedIn Narrative

Connects OrgSonar → 1st-place Everbridge Freaky Friday & 1909 Vermontville inspiration → AI + Mavic 2 proof of concept → Progressive Focus Dissolve.

You may have seen an earlier post where I shared screenshots of OrgSonar, which I built to help organizations pair formal compliance milestones with continuous, day-to-day operational health signals across teams.

To demonstrate OrgSonar realistically, I needed a fictitious company. Having had the privilege of working at organizations with truly noble missions, including Everbridge, whose Critical Event Management (CEM) platform helps public safety agencies and enterprises keep people safe and coordinate response during emergencies, I wanted to invent the kind of mission-driven company I’d genuinely want to join.

Back at Everbridge, teams ran Freaky Friday projects to come up with cool new features and ideas for the Critical Event Management platform. I pitched building real-time aerial overlay maps for search-and-rescue and first-responder events, and my team actually won first place for it. For the fictitious scenario here, I drew inspiration from a historic 1909 freight train wreck in Vermontville, Michigan, reimagining how modern responders arriving at a rail collision like that today would urgently need an immediate overhead picture of blocked roads, nearby homes, and downwind evacuation perimeters rather than months-old satellite imagery.

So I created a fictitious company called Aeris Overwatch Systems, with 1,800 employees across 10 departments and 39 teams. Phase 1 on that fictitious website included a deceptively simple line:
"Watch the stitched derailment imagery dissolve into place."

And so here I am, years later, and AI has helped me bring to life what I could only imagine back then. To test it as a real working proof of concept without buying a super-expensive enterprise RTK mapping drone, I grabbed my consumer-grade DJI Mavic 2 Pro outside by my house to build that capability in software (which still doesn’t have a product name yet!).

While an enterprise-class RTK drone can eliminate raw consumer GPS drift, aligned orthomosaic maps still typically require landing, pulling an SD card, or sending imagery back to the cloud or a workstation for post-processing rather than appearing live out of the box from an RTMP video stream. Proving this live, in-flight concept on a consumer Mavic 2 meant six real physics and UX hurdles had to be resolved in software:

1️⃣ Zero-GPS RTMP + Takeoff & Return-to-Home Contamination: Live RTMP streams raw video frames with zero GPS or gimbal tags, so an automatic cruise detector had to wait out the 50-second driveway climb/spin before Waypoint 1 and cut off the instant post-Waypoint-8 Return-to-Home spun 45° back across the neighborhood.
2️⃣ 2-Story Roof Speed Illusion & Open Lake Water: Without GPS, speed is measured from pixel motion, and 8.5m two-story roofs sit closer to the lens and rush ~27% faster across the frame than street pavement while Pass 4 hits 65% rippling water. Speed tracking had to lock strictly to flat street pavement and coast smoothly across open water.
3️⃣ Ghosted Double-Rooftops & Sliced Street Trucks: Opposite survey passes viewed pitched roofs from opposite sides, smearing dormers and slicing street trucks in half until I restricted stitching to the dead-center nadir strip with Dynamic Programming seam carving.
4️⃣ Live In-Flight "Visual-RTK" Without Post-Processing: Consumer GPS dropped stitched shoreline riprap ~4.5m into the lake, and standard workflows normally send imagery back for offline/cloud post-processing. To align live in the air out of the box with zero RTK hardware, I built a color-agnostic Sobel "Visual-RTK" pass that locks structural curbs and shoreline edges directly onto the satellite map.
5️⃣ The Live "Rigid Snap" Dealbreaker: Re-fitting one rigid mosaic sheet yanked already-drawn houses and shoreline ~5m under the operator’s cursor right as Pass 4 finished.

6️⃣ Progressive Focus Dissolve ✨
• Stationary Earth from frame one through 100% lock (zero pixel snap).
• A soft peripheral recon wash paints ahead of the aircraft for immediate situational orientation.
• Directly beneath the drone’s belly, crisp 2.68 cm/px detail smoothly dissolves into permanent focus, turning bare dirt lots on old satellite tiles into today’s completed houses, driveways, and vehicles.

#SearchAndRescue #Drones #ComputerVision #AI #PublicSafety #CriticalEventManagement #Geospatial #OrgSonar #AerisOverwatch