Daily haul-truck counts and road condition, read from imagery

A count of trucks on each segment of your haul network, plus a read of the road surface, from VHR imagery and with no site access.

Files in the pilot delivery

File
Daily truck count by segment
Format
CSV
What's in it
One row per segment per day: date, segment ID, trucks counted, and a flag for days with no usable image. A count is the trucks visible on that segment at the time of the pass. Lands in your benchmark spreadsheet or your own model.
One row per
segment × day
File
Road condition read
Format
CSVGeoPackage
What's in it
A condition note per segment per day for what shows on the surface at VHR: standing water, spillage on the running surface, broken or missing edge berms, wet or dry sections. It is a visual read, not a rut-depth or roughness measurement.
One row per
segment × day
File
Activity trend
Format
CSV
What's in it
The daily counts rolled into a trend by segment, with gaps left as gaps. Read it next to your own dispatch cycle times as a cross-check, or by itself where you have no such records.
One row per
segment
File
Cycle-time proxy
Format
CSV
What's in it
A rough index built from the count and the route length of each segment, for comparing one period or network with another. It is a proxy, not a cycle time. No truck is timed.
One row per
segment
File
Benchmark table
Format
CSVspreadsheet
What's in it
The segment counts, condition notes and trend laid out for an operations review or a research note, so two periods or two networks can sit side by side.
One row per
period or network

How it works

  1. You supply

    You define the haul network

    Send the pit access route and haul roads as a GeoPackage or shapefile, or mark them on a map with us. For an outside read, mark what you can see from public imagery.

    GeoPackageshapefile

  2. We do

    We acquire a daily pass

    One VHR image over the network each day the pilot runs, from satellite or aerial.

    VHR RGB

  3. We do

    Trucks are counted and the surface read

    Trucks are counted segment by segment and the road surface is read. A person checks the output before it goes out.

  4. You receive

    You receive the files

    Daily counts, condition notes and the trend arrive as CSV, with the segment layer as a GeoPackage.

    CSVGeoPackage

Limits

Cloud and darkness
An overcast or dark day gives no usable image. That day's row is flagged and left empty, never filled with an estimate, and the trend shows the gap.
Resolution
At 0.5 m or finer a full-size haul truck is a clear object on the road, enough to count it and to tell it from a light vehicle. It is not enough to tell models apart, or to see what a truck is carrying.
A snapshot, not a shift
One pass a day is one moment. It says how many trucks were on the network then. It does not see the rest of the shift, and it is not real-time.
What it does not measure
No production tonnage, no payload, no queue times measured from a truck. Road condition is what is visible on the surface; potholes and rutting finer than the pixel are not traced.
Pilot scope
Coverage is one defined haul network for a pilot period. It is not yet a continuous fleet-wide service.

The reason this exists

Your dispatch system records every cycle on your own fleet, and it has nothing to compare against. Another operator's numbers are private. An outside analyst has quarterly statements and a site-tour photograph. Nobody can see the haul network from outside, so a long queue at the load point or a cut-up ramp shows up as a missed month, well after the fact.

A loaded haul truck driving away up a gravel haul road between waste rock piles, a loader working behind it
Scenery: a loaded truck on a haul road, seen from the ground.

Who it's for

For

  • Mine planning engineers and superintendents benchmarking their own fleet against an outside read
  • Operators without dispatch data on a site, such as a contractor or a recently acquired pit
  • Equity and commodity analysts estimating a competitor's haul activity from outside the fence

Not for

  • Teams that need production tonnage or payload
  • Teams that need a continuous live feed of truck positions
  • Teams after truck model or fleet composition identification

Questions from planning and procurement

What happens on a cloudy day?

No usable image, so no count. The row is flagged and the trend shows the gap.

Can you tell truck models apart?

No. At this resolution a truck is countable and distinguishable from a light vehicle. Model identification is out of scope.

Does this give me production tonnage?

No. It counts trucks and reads the road surface. Tonnage is not derived from it.

Is it real-time?

No. It is one image pass a day.

How do I check it against my own records?

Line the daily counts up against your dispatch records for the same segments and days. Discrepancies are worth raising in the pilot.

Can you watch a site I do not operate?

Yes, where the haul network can be defined from imagery. You are working from outside the fence, so the count and trend are the read, with no access to the site's systems.

Who owns the imagery and the output?

Usage terms for the imagery and the delivered files are agreed at the start of the pilot.

How is it priced?

A pilot is scoped per network and period. Pricing comes from that scope, and a person will go through it with you.

Start a pilot on your network

Tell us the haul network and whether you operate it or are watching it from outside.

Start a pilot