Loading...
Loading...
Missoula's computer vision practice is the most academic of any Montana metro and that has practical consequences for how engagements are scoped here. The University of Montana's W.A. Franke College of Forestry and Conservation, the Numerical Terradynamic Simulation Group housed in the Stone Hall complex, and the Spatial Analysis Lab anchor a research bench that publishes regularly on remote-sensing CV problems — fire-burn severity classification from Landsat and Sentinel imagery, land-cover change in the Northern Rockies, and animal-movement detection from camera-trap networks. Around that academic core sits a private bench shaped by onX, the Missoula-headquartered outdoor mapping company in the Old Sawmill District, Submittable's document workflow product downtown, and a handful of forestry consulting firms tied historically to the Forest Service Region 1 office on Broadway. CV engagements here often start with a satellite-imagery question — what changed, where, how fast — rather than a real-time-camera question, and a Missoula partner who has shipped a remote-sensing pipeline through Google Earth Engine or Microsoft Planetary Computer is going to scope the project differently from a partner whose entire experience is in factory machine vision. LocalAISource maps Missoula buyers to the right archetype.
Updated May 2026
Forest Service Region 1, headquartered on Fort Missoula Road and West Broadway, has driven decades of forestry-imagery work in the Missoula valley, and the University of Montana's forestry school has built academic and applied capability on top of that base. Modern fire-burn severity mapping, fuels classification, and post-fire vegetation recovery monitoring all rely on CV pipelines applied to Landsat, Sentinel-2, and increasingly commercial Planet Labs imagery. The Numerical Terradynamic Simulation Group, founded by Steve Running, has produced the modeling and remote-sensing chain that downstream CV practitioners build on. Practical implication for buyers in the Missoula market: if your project touches fire, forest, or any large-area land-cover question, the local bench is unusually deep and unusually well-published. Pricing on a forest-imagery engagement that produces a defined deliverable — say, a beetle-kill detection model or a post-fire revegetation tracker — runs sixty to one-eighty thousand dollars depending on the area and the imagery cadence, with the higher numbers covering ongoing seasonal updates rather than one-shot baselines.
onX is the most visible private CV-adjacent employer in Missoula, building outdoor mapping products for hunters, off-roaders, and backcountry users out of the Old Sawmill District near Caras Park. The product surface — basemap accuracy, trail-detection from imagery, ownership-boundary clarity — is built on top of CV work, and the engineers who staff onX have produced spillover talent into the local consulting market over the last several years. onX alumni who left to consult independently or to build smaller geospatial products in Missoula make up a meaningful slice of the local senior CV bench. The implication for buyers outside outdoor recreation: if your problem is geospatial-CV-flavored — anything where the answer is 'detect this thing from aerial or satellite imagery and give me coordinates' — Missoula has more talent for that than its size would suggest. Engagement pricing for a defined geospatial-CV product or feature lands at fifty to one-fifty thousand dollars, and timelines run twelve to twenty-four weeks because basemap and georeferencing accuracy work always takes longer than naive scoping suggests.
Submittable, the application and grant-management product based at the corner of West Front Street and Higgins, anchors the small but real document-CV practice in Missoula. The product handles application forms, supporting documents, and reviewer workflows at meaningful scale, and the engineering team has built capability around document classification, layout extraction, and form-OCR that has trained a generation of Missoula engineers in applied document CV. Beyond Submittable, the Western Montana Innovation Initiative and the smaller downtown software shops working on document-heavy products have hired against the same talent pool. CV engagements in this niche differ from the forestry side: less Earth Engine, more LayoutLMv3 and Donut, more attention to OCR quality on degraded scans, more concern about end-to-end latency in user-facing review pipelines. Pricing for a document-CV engagement runs forty to one-twenty thousand dollars depending on form complexity and accuracy targets, with most projects landing in the eight-to-fourteen-week range. The Missoula AI Meetup, which hosts evenings at Imagine Nation Brewing Company on West Broadway, is a credible venue for vetting this side of the local bench.
Variable, and the difference matters. The W.A. Franke College and the affiliated research groups will engage with private partners through formal sponsored research agreements — typically ninety to one-hundred-eighty days from initial conversation to executed contract — and IP terms negotiate harder than buyers expect. Some faculty also consult independently on the side, which is faster but constrained by their academic load. The realistic structure for most private buyers is to retain a private consultant as the prime, with a UM faculty member as a named subject-matter advisor on a small fixed-fee engagement. That arrangement clears procurement, keeps timelines reasonable, and gets the academic depth where it matters.
Partly accessible. Active onX engineers are not, but the alumni network is. Several senior CV engineers who spent two to four years at onX have moved into independent consulting or smaller Missoula startups and take outside work. They are easy to find through the Missoula AI Meetup, the local LinkedIn outdoor-tech community, and the geospatial track at Imagine Nation. Buyers should ask candidates explicitly about their onX tenure and what they shipped there as a way to calibrate seniority against the local market.
For a defined area of interest at the scale of a few hundred thousand acres, a twelve-week engagement should yield a calibrated baseline classification model trained on at least two seasons of multispectral imagery, a ground-truth validation set with documented confidence intervals, an inference pipeline runnable on Google Earth Engine or a Planetary Computer notebook, and an integration into the buyer's existing GIS. What you should not expect in twelve weeks is a continuous near-real-time monitoring system; that requires a second phase and a different operating budget. A Missoula partner who promises both inside three months has probably under-scoped one of them.
It affects them more than buyers from out of the region expect. The Missoula CV bench overlaps with Forest Service Region 1 contractors, smokejumper logistics, and the broader fire-aviation community, and during a heavy fire summer — July through September — significant capacity gets pulled into emergency imagery and post-fire assessment work. That is good for a buyer whose project is fire-related and bad for a buyer whose project is not. Plan engagement timelines accordingly: schedule data-collection work in May and June, model development in October and November, and deployment in winter and early spring.
It can, but the bench is thinner for that work and the buyer should consider whether Bozeman is a better fit. Missoula's strengths are remote sensing, geospatial product CV, and document CV; pure industrial machine vision — line-rate defect detection at high frame rates with mining-spec or food-spec hardware — is closer to Bozeman's optics-cluster strength or, for serious volume, out-of-state. Honest Missoula partners will tell you this directly and refer rather than force-fit. Buyers should respect that referral and not interpret it as a lack of capability in this metro overall.
List your computer vision practice and get found by local businesses.
Get Listed