Ricoh has developed a polarization camera that can acquire polarization information from a subject in real-time.
A polarization camera acquires polarization information from a subject. Light contains amplitude (brightness), wavelength (color), and polarization information. Although human beings are able to sense brightness and the color of visible light, we are not able to sense polarization information. Insects with compound eyes have the ability to sense polarization, and it is said that they perceive a world quite different from our own.
Acquisition of polarization information with a normal camera allows only changes in intensity as a polarization filter rotates in front of the camera to be viewed. Ricoh has developed a polarization camera that captures polarization images in real-time without mechanically adjusting the polarization filter or other parameters. A polarization camera makes it easier to acquire information on the surface orientation of hard-to-distinguish single-colored objects and to recognize hard-to-distinguish differences between transparent surfaces and non-transparent surfaces. A broad range of applications is possible, including factory automation (FA), of course, but also the security, pharmaceutical, and food industries.
By precisely mounting an area-dividing, sub-wavelength structure (SWS) polarization filter on a CMOS sensor, a Ricoh polarization camera can acquire polarization information in real-time, something that has been difficult to capture with conventional cameras.
Until now, it was difficult to determine the shape of a single-colored subject from a captured image in some cases. A polarization camera, however, makes it easier to detect differences in the surface orientation of single-colored subjects. This is because the polarized state of light from a subject differs depending on the surface orientation of that subject. These differences can be captured as polarization images with a polarization camera. Potential applications include object recognition for single-colored objects in a production line.
Figure 1: Usage scenario for capturing single-colored subjects
Figure 2: Imaging comparison between a normal camera (left) and a polarization camera (right)
With a polarization image, distinguishing the shape of single-colored subjects is possible.
Until now, it has been difficult in some cases to determine the presence of transparent subjects from a captured image. A polarization camera, however, makes it easier to detect the presence of transparent subjects. This is because when light passes through a transparent subject, its polarization changes depending on the transmittance of that subject. These differences can be captured as polarization images with a polarization camera. Potential applications include object recognition for transparent objects on a production line.
Figure 3: Usage scenario for capturing transparent subjects
For example, as shown in Figure 4, the polarization camera makes it easier to detect presence of transparent glass plate.
Figure 4: Imaging comparison between a normal camera (left) and a polarization camera (right)
With a polarization image, distinguishing the shape of transparent glass plate is possible.
Also, as shown in Figure 5, the polarization camera makes it easier to detect the presence of foreign material such as a plastic wrapper inserted in a glass bottle. This is because the polarization camera can capture a difference of polarization between light being passed only through the glass bottle and light being passed also through the plastic wrapper.
Figure 5: Imaging comparison between a normal camera (left) and a polarization camera (right)*
With a polarization image, distinguishing the shape of transparent plastic wrapper inside glass bottle is possible.
(* The polarization image is colored for easy recognition.)
Furthermore, as shown in Figure 6, in the case of a package of pills where a plastic layer is accumulated on a thin metal such as aluminum, the polarization camera makes it easier to detect stamped marks on the plastic layer.
Figure 6: Imaging comparison between a normal camera (left) and a polarization camera (right)
With a polarization image, distinguishing the stamped mark on a package of pills is possible.
Sorted by : field “Machine Vision”