I spent more than a decade inside road maintenance contractors, trying to get regular
imagery of the road network into everyday inspection. Underneath every version of that
problem was the same thing: roads and public infrastructure are inspected periodically
and expensively, and between those inspections nobody really knows how conditions are
changing. A pothole opens, a sign fades, a drain blocks, and it can be months before
anyone with the right eyes drives past. If you've seen the same gap from your side of
the industry, as an authority, a contractor or a vendor, we want to hear how it looked
from where you sat.
Meanwhile, the answer has been sitting in plain sight the whole time. Millions of
ordinary vehicles already carry cameras, GPS and computers, including the trucks and
utes already out there maintaining the network every day. Contractor after contractor
tried to solve this, and each one rebuilt the same idea slightly differently, behind
their own walls, for their own contract, while vendors competed to sell a slightly
better version of the same thing. None of it added up to anything lasting. This market
isn't big enough for ordinary competition to fix that on its own, so we're doing it the
other way round.
We started OpenSight to fix that. It turns the vehicles already on
our roads into a distributed network of observation: an inexpensive, open appliance
captures everyday imagery as people drive, publishes it openly through platforms like
Mapillary, and feeds an open database where computer vision and human review turn that
imagery into evidence of assets, defects, and how they change over time.
The contract should end. The data shouldn't.
Outside my day job I build and run my own virtualised infrastructure, keep it properly
monitored and backed up, and have spent years building vehicle-mounted GPS and camera
capture tooling and geospatial software. I wrote OpenSight's first working prototype
myself, end to end, before asking anyone else to believe in it. We hold the whole
project to that same standard now: build it properly, then invite you in.
This is our commitment, and it's the whole reason we're doing this as a foundation
rather than a business. Every core OpenSight repository (the hardware designs, the
software, the schemas, the models) will be open source. Anyone, anywhere, can look at
it, use it, question it and improve it. When someone does improve it, we'll adopt that
improvement back into the project, properly credited, so the whole community benefits,
not just whoever found it first.
This is a call to action. To road authorities, contractors, technology companies,
universities and researchers: we want your expertise, your critique and your real-world
use cases. If you've already solved a piece of this problem, or can see a better way to
do it, tell us, and we'll look at folding it back into the project. And to the
individual engineers, GIS specialists, data scientists and everyone who simply cares
about the infrastructure their community depends on: you don't need an invitation or a
title to contribute. There's a seat at this table for you.
Part of that call is practical: to incorporate The OpenSight Foundation Ltd we need
three directors and a company secretary, and we already have a few people in mind who
are keen, so if that's you or someone you know, put yourself forward. Our focus for now
is Australia, where these roads are the proof case, but we'd welcome anyone in another
country who wants to start something similar there.
For now, the best way to reach us is LinkedIn, where messages get a personal reply, or
by email at hello@opensight.com.au. The
public GitHub repositories are coming very soon, and we'll share them the moment they're
live.
We're proud of what OpenSight already is, and even more excited about what it can
become with your hands on it too.
Clynton Husband
Founder, The OpenSight Foundation
