How to read a shift change on a mine haul road from overhead imagery

If you're trying to benchmark a competitor's haul fleet from outside the fence, the first thing that trips people up is the shift change dip. Truck count on the haul road drops hard around crew changeover, and if you're reading a single overhead frame without context, it's easy to mistake that dip for something it isn't: a stoppage, a weather hold, a maintenance event on the primary route.

Here's what's happening during that window. During crew changeover, trucks queue at the pit or park near the change-out point while outgoing drivers hand off to incoming ones. For a window that typically runs somewhere in the twenty-to-forty-minute range depending on the operation's changeover protocol, fewer trucks are moving on the haul network even though the fleet hasn't gone idle. Count them from a single satellite pass taken during that window and you'll read a materially lower number than the same road ten minutes earlier or later.

What a shift change looks like in a single daily frame

A VHR pass gives you one moment in the day, not a curve. At sub-0.5 m resolution you can usually tell a moving loaded truck from an empty one by the shadow and the load profile, and you can count stationary trucks parked near the pit entrance or the crib area distinct from trucks queued at a crusher or stockpile. What you can't do from one frame is tell whether that parked cluster is shift change, a safety stand-down, or trucks waiting on a grader to clear a wet spot.

The tell is position, not just count. Shift change parking clusters at a known change-out point, often near the pit access road or a designated muster area. A line of trucks strung out along a haul segment, backed up behind a single choke point, points to a different story: a maintenance hold or road closure stacking the fleet in a queue along the route itself.

Separating a shift change dip from a stoppage

Road condition matters here too. A wet or recently graded section slows cycle times and can produce a count dip that looks similar to shift change if you're only looking at raw truck numbers. Reading the haul road surface in the same frame, rutting, standing water, a grader working a section, helps rule that out. If the road looks clean and trafficable and the dip still shows up at a consistent time of day, that's a much stronger shift change signal than a dip with visible surface disruption nearby.

The single biggest limitation worth naming plainly: one daily pass can't confirm a shift change pattern on its own. It can only tell you the count was low at that particular overpass time. The pattern comes from stacking several days of passes at a roughly consistent capture time and watching whether the dip repeats. A one-off low count could be anything. A low count that shows up at the same time of day across a run of daily images, with trucks clustered at a known change-out point rather than strung along the route, and a clean road surface, is a shift change.

Building the daily pattern across passes

The benchmarking exercise works like this: a run of daily counts, each tagged with timestamp and road condition, stacked one on top of the next. Overlay enough days and the shift change dip stops looking like noise and starts looking like a schedule, two dips a day for a two-shift site, one for a single-shift operation, or a different rhythm entirely depending on how the site runs its roster. That daily rhythm is what you're benchmarking against.

Mine Fleet Activity is built around that stacking exercise, a daily truck count on a defined haul network paired with a road condition read, so the pattern builds itself out day over day instead of you reverse-engineering it from one-off frames. If you're scoping a haul fleet benchmark from outside the fence, see how the daily count and road read works on a defined network.