跳转到主要内容

Knowledgebase Shape from Shading

The Shape from Shading method has established itself in Machine Vision as a powerful way to detect and evaluate features on objects with difficult surfaces.

Where there is light,
there is also shadow.

The optimal illumination of test objects is an essential prerequisite for the subsequent image recording and evaluation. Untypical reflections of the incident light are recorded by the cameras used and can be identified as defects by the image processing software. If test objects are illuminated successively from different directions and the camera images recorded are combined, even defects that are difficult to detect can be detected more reliably. This principle is called Shape from Shading.

Technical background

For the majority of all image processing applications, it is sufficient to illuminate the test objects with a light source from one direction and to position the camera in such a way that usable images are created which can be easily processed by subsequent software. However, there are also cases in which defect features cannot be detected so easily in this way. This applies, for example, to the inspection of components that have a shiny and diffusely scattering surface. If the application requires identifying the smallest scratches, cracks, grinding marks, deformations or other small defects on such surfaces, image processing systems with only one illumination often reach their limits.

The shape-from-shading method offers additional options for such cases by taking images of the test objects still with only one camera, but using multiple illuminations from different directions. Each illumination direction produces a different image because the light reflections reflected back from the object and its defects are different. How many illuminations are needed depends on the respective application - however, practice has shown that very good results can already be achieved with four light sources arranged at a 90° angle. If suitable software tools are used to create a combined overall image from all the individual images, this method can even be used to reliably detect fault features with a size of just a few micrometers.

Applications in the image processing industry

The Shape-from-Shading method is particularly suitable for testing textures and topographies of demanding surfaces, as they are frequently required in numerous industrial sectors such as, among others, the automotive industry, electronics manufacturing, the production of high-precision surfaces e.g. for turbine blades for energy generation or for aircraft engines, or also in the production of plastics. Shape from Shading is also a suitable method for detecting desired structures, such as Braille on colored backgrounds or the detection of raised lettering on metal or cast parts.

Shape from Shading with SVS-Vistek cameras

A prerequisite for the use of shape-from-shading technology is precise timing between the illuminations used and the camera. A separate controller for the lighting control is usually required for this, which, in addition to the actual image processing components, causes additional costs and increased effort for the coordination of the entire system.

By using suitable cameras from SVS-Vistek, this additional effort is eliminated: All our camera models contain an integrated strobe controller that provides up to four output channels with a maximum of 3 amps each. We call this concept 4IO Strobe Control. With cameras that have this feature, up to four lights can be controlled with high precision. The construction of economical shape-from-shading systems thus becomes child's play.

 

返回概览