Loading...
Loading...
Butte's computer vision economy does not run on retail loss prevention or autonomous shuttles — it runs on the open pit. Drive past the Berkeley Pit overlook on Continental Drive and you can see the visual problem the local CV community keeps getting paid to solve: hundreds of acres of acidic water, slumping highwall, and ongoing remediation work that nobody wants to walk and nobody wants to instrument by hand. Montana Resources, the active copper-and-molybdenum mine northeast of town, runs haul-truck and shovel telemetry alongside drone imagery for pit-wall stability and inventory measurement. Pit Watch and Atlantic Richfield's Superfund operations need ongoing visual monitoring of the Berkeley Pit and the Anaconda smelter site under EPA oversight. That niche is what gives Butte CV any density at all — a small but unusually deep bench of engineers who know how to deal with dust, blast-vibration, sub-zero cold, and federal documentation requirements. Most of them came through Montana Tech's College of Mines and Engineering on West Park Street, which runs undergraduate and graduate work in metallurgical engineering, mining engineering, and more recently a software-engineering track that has been quietly producing CV-capable graduates for the past decade. LocalAISource pairs Butte buyers with the practitioners who can read a haul-road camera feed and a 43-101 report in the same week.
Updated May 2026
If you want to understand why CV in Butte looks different from CV in Missoula, drive up to the Continental Pit fence line and watch what happens during a shift change. The visual workload here is photogrammetric drone surveys for end-of-month pit inventory reconciliation, fixed-camera monitoring of tailings impoundments for slump and seepage, haul-truck driver-attentiveness systems running on mining-spec ruggedized hardware, and time-lapse work on the Anaconda smelter cleanup that has to be defensible to EPA Region 8 reviewers. Montana Resources, the Atlantic Richfield-owned remediation operation, and several smaller exploration outfits in Walkerville and Whitehall make up the buyer pool. The technical implications are sharp: imagery is captured on Phantom and Matrice drones with RTK GPS, processed in Pix4D or Bentley ContextCapture, and the CV layer typically does change-detection across surveys rather than novel object classification. A Butte CV partner who has not actually walked a bench, who does not understand why a mine surveyor cares about a five-centimeter discrepancy in toe-of-slope, and who has never had to defend a result to a state regulator at the Montana DEQ in Helena will produce work that looks fine in a notebook and gets rejected in the field.
Montana Tech is the reason Butte has any CV practice at all, and the reason that practice does not scale past a certain point. The school's undergraduate Computer Science and Software Engineering programs, plus the Geophysical Engineering and Mining Engineering departments, produce graduates who can do real CV work, particularly anything photogrammetry-adjacent. The Mineral Research Center on campus and the High Bay Lab inside the Natural Resources Building have hosted sponsored research with Montana Resources and Rio Tinto on automated rock-fragmentation analysis and shovel-load monitoring — both classical CV problems that have aged into deep-learning territory in the last few years. The constraint is volume: graduating classes are small, and most strong students leave for Bozeman, Boise, Salt Lake, or Denver within two years. The realistic Butte staffing plan for a serious project is one or two senior independent consultants who chose to stay, supplemented by Montana Tech graduate students on a project-funded basis, with the understanding that any team needs to be locked in early because the bench is shallow.
Senior Butte CV engineers bill in the one-fifty to two-thirty per hour range, lower than Bozeman by a measurable margin and reflective of a smaller, more industrial buyer pool. A typical pit-imagery or tailings-monitoring engagement runs twelve to twenty weeks and lands between thirty-five and ninety thousand dollars, with the cost driver being deployment hardware, not engineering hours. Mining-spec enclosures, IP67 industrial cameras from Basler or FLIR, sealed Jetson AGX Orin compute units in NEMA 4X housings, and the labor to install them on a pit perimeter or a tailings dam crest add fifteen to forty thousand dollars to most projects. Edge inference is non-negotiable because a working pit's connectivity is dust-attenuated 900 MHz radio at best, and the Berkeley Pit area in particular has documented coverage gaps. A Butte CV partner who knows the local installation contractors — the same crews that put up the original pit cameras for security and slope-stability work — saves weeks of scheduling friction. The Butte-Silver Bow Chamber's Manufacturing and Mining Roundtable, which meets quarterly at the Butte Public Archives, is a reasonable place to find references for those crews.
Some can, and the ones that can are the only ones worth hiring for that scope. Anything tied to the Anaconda Smelter NPL site or the Berkeley Pit complex falls under EPA Region 8 oversight, which means imagery, models, and decision logs need to be reproducible, version-controlled, and defensible to a federal auditor years after capture. A Butte partner who has worked under an EPA Administrative Order on Consent or a state DEQ Record of Decision will speak this language fluently. Ask specifically about prior Superfund or hard-rock-mining regulatory work in the kickoff meeting; if you do not get a concrete answer, the partner is probably not the right fit.
Less than buyers from coastal metros assume. Butte sits at fifty-five hundred feet on the Continental Divide, and winter winds at the pit rim regularly exceed the operating envelope of a Phantom 4 or Matrice 350. Realistic flight calendars at active mines here lose two to four weeks of December through February to weather, and summer afternoons get smoke-impacted in fire years. Local CV partners design pipelines that gracefully handle these gaps — interpolating from prior surveys, falling back to fixed cameras, and scheduling change-detection runs to align with predictable atmospheric windows. A vendor who promises continuous monitoring without acknowledging this is not telling the truth about Butte conditions.
Roughly fifteen to twenty-five percent below Bozeman for engineering hours, but often equal or slightly higher for full deployed-system pricing because the hardware burden is heavier. Butte projects spend more on ruggedization, more on installation labor in difficult terrain, and more on travel for any team members not based in Silver Bow County. The honest comparison is to look at total project cost and time-to-first-useful-output, not hourly rates. Buyers who optimize purely on hourly rate without modeling the deployment overhead end up surprised by the all-in number.
Yes, and the Mineral Research Center is the main door. Sponsored research projects with the College of Mines and Engineering can take on commercial CV problems, particularly in fragmentation analysis, ore-grade estimation from imagery, and tailings-impoundment monitoring. The contracting cycle runs ninety to one-hundred-twenty days, and IP terms favor the university more than private buyers expect, but the cost is well below private consulting rates and the resulting work is publishable, which matters for some industrial buyers. Several Butte private consultants moonlight as Montana Tech adjuncts and can structure these arrangements cleanly.
Modest, and that is by design. Butte is not the right metro for a buyer who needs cutting-edge vision-transformer research or large-model pretraining work — the local bench is built around applied, deployment-focused engineering on industrial problems. For those harder research questions, even the best Butte partners will subcontract or refer out to teams in Bozeman, Boise, or Boulder. What Butte does at a high level is hardened, regulator-defensible CV systems on real pits and real tailings dams, and that is where buyers should aim their budget.
List your Computer Vision practice and connect with local businesses.
Get Listed