KOMPASS AI checks visible edge chipping, missing corners and breakouts on manufactured parts. Bring the critical edges into view, distinguish damage from the intended profile and flag affected units before assembly or packing makes another inspection harder.
Edge chipping often sits outside the view used to inspect the main surface. A broken corner, damaged rim or small breakout along a side can pass a face check and reach the next operation unseen.
Our AI Product Inspection system, KOMPASS™, combines visual defect classification with modular vision hardware. The inspection is designed around the accessible edges, the intended profile and the smallest damage your process needs to reject.
An edge check needs a clear view of the region that matters. Plan the presentation, distinguish damage from intended geometry and keep the image with the inspection decision.
Plan camera views around the areas where damage matters. Check whether the fixture hides the lower rim, a sidewall or a corner, and whether the part needs to be turned.
The model uses production examples to distinguish the agreed defects from acceptable geometry and finish. A designed notch or chamfer belongs in the reference samples.
Connect the inspection result to the defined line response and retain the defect image. Quality teams can review where the damage occurred and whether the same location keeps appearing.
Show us a chipped part, an acceptable edge and the tolerance that separates them. We will assess the views needed to make that decision repeatable.
Get In TouchA method that works on a repeatable silhouette may be enough for one part. A complex edge with varying appearance needs a different combination of views and decision tools.
An edge defect is a departure from the accepted condition of the part's boundary. It can include a missing corner, a local breakout or another visible irregularity defined in your quality standard. Edge chipping detection focuses on unintended loss of material. The decision should account for size, location and the function of the affected region. Bring the drawing and representative parts to the review so the team can distinguish a cosmetic limit from a requirement that affects assembly or use.
Chipped edge detection starts with a clear definition of the intended profile. Good samples should include the chamfers, radii, notches and normal cut texture that production is allowed to have. Chipping inspection then assesses departures from that accepted appearance. Include borderline rejects and accepted parts that look similar. If the decision depends on a measured geometric tolerance, specify the required measurement and reference so the imaging and calibration can be assessed alongside the visual classification.
Only if its view exposes every region required by the acceptance standard. Edge quality inspection may need top, bottom and side views, or handling that rotates or flips the component. A fixture can hide an edge even when the camera has enough resolution to see the defect. Check the proposed presentation using damaged samples at each critical location. Coverage should describe the regions actually inspected, including any inaccessible area that needs a separate checkpoint.
Glass edge inspection needs an imaging arrangement that manages transparency, reflections and the position of the damage. Ceramic inspection also depends on the component's finish, contrast, dimensions and presentation. Jidoka lists glass, ceramic and machined parts within this use-case scope, but each application needs sample testing. Include chips on different faces and corners, together with acceptable variations. The trial should establish visibility and detection limits for the actual parts instead of assuming one setup works across materials.
Burr detection and broken edge detection concern different forms of damage. A burr adds unwanted material to the boundary, while a chip removes material. They can share an inspection station if its views and lighting expose both clearly. Keep the defect categories and acceptance criteria separate during validation. That makes it possible to review missed defects and false rejects for each task, and prevents a good overall result from hiding weak performance on one category.
Edge profile inspection requires a defined reference and suitable measurement capability. A visible outline can show that material is missing without establishing the full depth of the break. If acceptance depends on depth, radius, angle or a calibrated dimensional limit, specify those requirements during feasibility. The team can then assess whether appropriate 3D sensing or another measurement method is needed. Visual classification alone should not be treated as a dimensional certificate or proof of internal integrity.
Wafer edge inspection has specialised requirements for resolution, handling, cleanliness and defect classification. Those must be assessed for the particular process. The bearing video on this page demonstrates a multi-surface inspection approach and is not evidence of wafer capability. For electronics components, share the material, dimensions, critical edge regions and acceptance criteria with Jidoka. A feasibility study can establish whether the required views and detection performance are achievable for that specific component and production environment.
Validate chip inspection with a labelled set covering the smallest rejectable damage, accepted edge variation and each important location on the part. Use samples that were not part of training and test the intended speed, fixtures and orientations. Record missed defects and good parts rejected separately. Also verify that the configured response acts on the correct unit. Together, those checks establish whether the complete station can support your quality decision under production conditions.
Share your damaged samples, accepted edge profile and required production rate. We will assess how to expose the critical regions and validate the decision before the part moves on.
