Solutions / Bore Inspection Robocell

Inside the part,
not just on the surface.

A standard cell for automatic inspection of internal channels. A robot inserts the borescope into every bore and two AI models analyse every frame in real time.

Send us your parts Request a demo

2cells in production on hydraulic manifolds

8 mmminimum bore diameter we can inspect

3dedicated operators freed per cell

~2 yearsestimated payback

The starting point

A defect inside a bore is found by your customer

Manual borescope inspection is the hardest check there is: it demands constant concentration, is physically punishing and is still only a sample. Training an operator takes months, and the instruments are expensive.

In hydraulic manifolds and valve bodies, a ridge or a lump inside a channel can't be seen from outside, yet it compromises everything downstream.

The cycle

Four steps, no sampling

Loading

The part is placed on the rotary table, by hand or automatically.

Scanning

The robot inserts the borescope into every channel, with 0° and 90° optics.

Analysis

The AI models process every frame in real time.

Result

Pass or fail, with the image of the defect saved for traceability.

The cell's robot with the borescope mounted — The robot with the borescope, inside the cabin.

Defects detected

Frames from the bores

Real borescope frames, with the defects marked by the model: ridges in red, grit in blue, lumps in green. Hover over an image, or tap it, to see the original frame.

Large ridge on the wall, segmented by the model — Ridge
large, on the wall

Ridge where two channels meet, segmented by the model — Ridge
where two channels meet

Small ridge on the edge of the hole, segmented by the model — Ridge
small, on the edge of the hole

Thin ridge, side view, segmented by the model — Ridge
thin, side view

Ridge, front view of the bore, segmented by the model — Ridge
front view of the bore

Grit residue in the channel, segmented by the model — Grit
residue in the channel

Large protruding lump, segmented by the model — Lump
large, protruding

Small lump on the wall, segmented by the model — Lump
small, on the wall

The models

Two models, two ways of looking

Both models are validated on an independent dataset before production, and keep improving as parts are inspected.

Segmentation

The typical defects of your process

Pixel-by-pixel analysis trained on the defects of your process, such as ridges, grit and lumps. It returns the exact extent of the defect, not just its presence.