THE POWER OF COAXIAL DOME LIGHTING OVDU-C120W

A recent user in Kyoto City contacted us to see if we could do something more about it.

At the experimental level, we knew that the coaxial dome illumination OVDU-C120W, which we frequently use in our experiments, could almost completely eliminate rolling streaks on metal surfaces. However, due to the fact that we were using a high pixel camera to capture images of about 50 mm square, we assumed that the angle of view would be insufficient with this lighting.

This time, when we brought it to the site for testing, we were able to install it without any particular problems, and it went straight into operation.

The effect is very effective. Random streaks (rolling streaks) on the metal surface, which used to occur randomly due to the directional nature of the illumination, are no longer visible at all, and the S/N ratio of the plating to be detected is now sufficient. The surface brightness variation σ is now less than 2, with a range of about 10 at ±3σ, giving the system an overwhelming detection capability. The person in charge of the project also commented that it was "perfect.

SINCE I STARTED USING THIS ILLUMINATOR REGULARLY, I HAVE COME TO FEEL THAT THE DIRECTIONALITY OF LED ILLUMINATORS DOES NOT ALWAYS HAVE A POSITIVE EFFECT ON IMAGE PROCESSING. ONE SHOULD NOT EXPECT THE LIGHT TO COME FROM A CERTAIN DIRECTION AIMED AT A SPECIFIC DEFECT. FORCING UP THE CONTRAST OF A DEFECT MEANS AMPLIFYING THE SLIGHTEST CHANGE, WHICH IN TURN MAKES EVEN THE IMAGE OF A GOOD PRODUCT UNSTABLE.

The illuminator that is really needed is one that can obtain a stable image from a good product and recognize defects on the image.
However, this is only possible with image processing technology that can detect the slightest change in the image.

*This illuminator reduces the three-dimensionality of the object and the contrast of scratches. It is not a universal solution. Methods to solve these problems are being studied separately.