Ajile DepthScan Imager Pairs with PolyWorks for Robot-Mounted 3D Inspection

A robot-mounted Ajile DepthScan 3D Imager now feeds point-cloud data directly into PolyWorks inspection software, measuring complex part geometry against CAD models and engineering drawings. The workflow registers individual captures in the robot's coordinate frame and combines them into a full-colour point cloud, which PolyWorks then evaluates feature by feature.

The combination targets a specific failure mode in automated optical inspection: critical features captured at poor viewing angles yield unreliable dimensional data. Bores, bolt holes and side faces typically demand near-normal viewing angles before measurements become usable.

How does the cell get calibrated?

Deployment starts with hand-eye calibration. Ajile Calibrate walks the user through the procedure that establishes the imager's position and orientation relative to the robot's tool flange. A reference artefact then verifies the measurement cell before production inspection begins.

The robot's movement programme, positioning and safety arrangements sit outside the metrology chain and are configured separately. At each programmed inspection position, the robot signals arrival, the imager captures the view, and the robot moves on.

Hand-eye calibration plus artefact verification is the gate: no production data flows until the cell passes both.

What does PolyWorks do with the scan?

Inspection planning happens in PolyWorks against the component's CAD model and drawing. Existing CAD-based measurement plans can serve as a starting point, but scan-based inspection requires feature extraction settings reconfigured for point-cloud data. Once built, the plan saves as a template and is reused for subsequent components.

On the shop floor, operators work in Ajile Inspect: select the part, enter its serial number, initiate the cycle. The robot executes the programmed sequence of views; PolyWorks evaluates the captured data and flags any control that falls outside its specified limits.

The software reports deviations from nominal values and a pass-or-fail status for each specified control. Results save against individual parts and export in batch reports, giving quality teams documented traceability per serial number.

The takeaway: one operator action — part selection and serial entry — triggers capture, evaluation and archived results.

Why develop on a robot before fixing the imagers?

The robot cell doubles as the development platform for a faster production configuration. Once viewing positions are established and validated on the robot, the same inspection geometry transfers to a fixed Multi-DepthScan frame surrounding the component.

In that arrangement, multiple imagers capture the part simultaneously rather than sequentially, cutting data-acquisition time compared with robot positioning through each view in turn. The trade is clear: the robot cell offers flexibility during process development, while the fixed multi-imager frame targets shorter cycle times where inspection speed governs throughput.

Ajile positions this as a staged adoption path — flexible robot-guided inspection first, fixed-cell measurement once the inspection geometry proves out. More information is at www.ajile.ca.

The open question for adopters is cycle time: whether simultaneous multi-imager capture, once validated, meets takt constraints that a sequentially positioning robot cannot.