ginngi engineering

All standard definition analog cameras produce interlaced video, which divides a single frame into two half-fields separated in time by half a frame. During display the two fields are 'combined' to reform the original frame. This combining is referred to as 'deinterlacing' or the effect of converting a frame from interlaced to progressive formats. Today virtually all digital video systems prefer to work with progressive format video, both for storage and playback. This is especially true for digital video surveillance systems. Combining two interlaced fields together is a non-trivial task since the second field is offset in time from the first. If the image has motion then this will be seen as blurred when combined and viewed using simple combining, which results in a significant loss of resolution.

Ginngi has developed a high-performance video deinterlacer IP for FPGAs that use complex motion detection algorithms to determine not only if there is motion but how much motion there is for each pixel in the frame. This 'motion-factor' is then used to optimally interpolate the pixel as a function of the spatial grid around the pixel and the temporal relationship to the corresponding pixel in the previous field. This results in a high-resolution, progressive image even for pixels that exhibit erratic motions.