Build inspection from CAD

Why CAD-Driven Inspection Actually Works

Design geometry → deployable QA logic for high-mix lines.

Rose City Robotics, led by Joseph Cole, PhD, helps factories catch defects earlier, cut scrap, and keep inspection stable—even as parts and finishes change. RoseVision CAD Studio trains from your design files and deploys on your edge hardware.

Injection molding (flash, sink, short shots) • Die casting (porosity, inclusions) • Stamped sheet (warp, surface)

What we do

We convert your CAD/STEP into deployable inspection logic. No brittle threshold rules. No waiting for a big dataset of bad parts. Pass/fail + defect localization outputs integrate to your PLC, robot, or HMI.

  • CAD → inspection model
  • Jetson Orin Nano/NX or x86—on‑prem
  • Works with FLIR/IDS USB3 industrial cameras

If your CAD defines what "good" looks like, your QA should start there—not after scrap shows up.

Our process

1

Upload CAD and context

Drag a STEP file into RoseVision CAD Studio and add the inspection context (part details, defects, features, environment).

2

Simulate, define, and validate

We build inspection zones and checks from geometry. You confirm tolerances, and we verify cycle‑time fit.

3

Deploy on your line

We package a production‑ready model for Jetson Orin Nano/NX or x86. Outputs are simple: pass/fail, defect class, location. When the drawing changes, update CAD → regenerate logic → redeploy.

Why factories choose Rose City Robotics

Catch defects before first shot

Build inspection before tooling is finalized to avoid post‑paint surprises.

High‑mix friendly

CAD‑driven logic adapts across variants without weeks of re‑tuning.

Keep your cameras

Upgrade inspection logic, not your entire station.

Real support from engineers

Built and supported by people who ship to factory floors—not demo teams.

Where it fits best

Applications

  • 1
    Injection molding & plastic enclosures: flash, sink marks, short shots, surface blemishes.
  • 2
    Die casting & machined housings: porosity indications, inclusions, missing features.
  • 3
    Stamped & formed sheet metal: warping, incomplete features, cosmetic damage.

Launch Scenarios

  • New mold/tool introductions
  • Replacing brittle rule‑based QA stations
  • Move first‑article inspection earlier

Hardware & integration

Edge Compute

NVIDIA Jetson Orin Nano / Orin NX, or x86 industrial PCs.

Cameras

USB3 (FLIR, IDS) and compatible industrial units.

Outputs

Pass/fail + defect class + location; integrable with PLCs, HMIs, and robot pick/segregate routines.

Environment

Built for production variance and real‑world lighting.

Built by engineers, led by Joseph Cole, PhD

Joseph has spent 20 years shipping machine vision and signal‑processing systems into real production environments—semiconductor inspection, ultrasound imaging, and factory automation.

He founded Rose City Robotics to make QA behave like good tooling: defined, repeatable, and adaptable.

Joseph Cole, PhD

How to get started

Feasibility Sprint

Ship a STEP file and a short video/photo of the inspection setup. We return a model preview, cycle‑time check, and deployment plan. If it meets your requirements, we finalize station integration with your team or preferred integrator.

Factory License / On‑Prem

For plants standardizing across lines, we provide on‑prem deployments with unlimited builds and SLA support.

See It In Action

Run your part through RoseVision.

See how CAD‑driven inspection works for your specific application.

Partner‑friendly

System integrators, camera vendors, and machine builders use RoseVision to de‑risk vision on quotes and deliver stations that hold up in the field. Bundle our feasibility sprint or include RoseVision in your cell to reduce support calls and speed installs.

Contact

Email
Engineering + Sales: hello@rosecityrobotics.com
Location
Location: Portland, Oregon
Lead time
Typical lead time: Feasibility in days; deployment aligns to your station build.

Software‑first QA. Built from your CAD. Deployed on your line.

Subscribe to our newsletter

The latest machine vision news and articles, sent to your inbox weekly.