High Dynamic Range is typically reconstructed merging the dynamic content of different images taken at different exposure time. The different exposure time capture different parts of the dynamic range of the scene under acquisition. An example is given in Figure 1, where four images with different exposure times are taken and merged to obtain the final HDR image.
Figure 1: First row, four photographs taken at different exposure time. Second row, the High Dynamic Range image with the reconstructed dynamic range.
Once this vast dynamic range has been acquired its need to be stored. This step involves an encoding operation that consists to use floating-point representation. This consists typically to use 32/48-bit representation per colour channel. This type of encoding, allows storing more fine details, contrast, colours increasing the quality representation of the acquired scene when compared with traditional 8-bit digital imaging. The use of floating-point representation obviously is increasing the amount of memory used to store this type of data. Depending on the type of application, an image/video compression step is required that is backward compatible with standard imaging/video compression techniques i.e. jpeg, mpeg etc.
Manipulation - Visualisation
The HDR content, once stored, may be manipulated for providing different visualisation solutions. Typically we have available two types of display technologies, traditional 8-10 bits and HDR displays. When using traditional display technology, the vast dynamic range available in the input HDR content needs to be manipulated to be adjusted in the narrow dynamic range available in this type of displays. This allows keeping as much as possible the contrast and details of the original HDR content avoiding underexposed and overexposed areas in the output image/video. This process is called Tone Mapping and typically is manipulating only the luminance information. When an HDR display is available, direct visualisation of the input HDR content may be possible. This depends on the dynamic range reproducible by the HDR display, and soft manipulation of the input HDR content may be required. In the case the HDR content, has been previously encode (compressed), an encoding (decompression) process is required before its manipulation and visualisation.
XDepth is working on the development of the entire HDR pipeline, Figure 2, from acquisition, storage-manipulation, to delivering-visualisation.
Figure 2: XDepth HDR Pipeline, offering wide range of products from the acquisition to the delivering-visualisation steps.
The XDepth HDR pipeline is offering a wide range of solution for still images and video, providing high quality and computational performances that make its products mature for being delivered on the market.
Click on the links below to more information about our technology: