Abstract
Color images can be obtained from a single solid-state sensor by covering the photosites with a repetitive pattern of color filters. This paper reviews the major issues in selecting appropriate filter patterns and compares different one-chip camera approaches. The processing used to decode the color signals and form acomposite television signal is discussed. The tradeoffs between using primary or complementary filters are considered. Checkerboard geometries are shown to be superior to stripe geometries, and the advantages of horizontally staggered sampling are explained. The constraints imposed by the sensor architecture and by the need for interlaced readout are examined.

This publication has 0 references indexed in Scilit: