A new registration error detection system for 3-tube HDTV camera has been developed to reduce chromatic aberration and to enhance resolution. The system detects registration error with live camera picture in real-time. Conventional pattern matching technique and newly developed false error rejection technique are adopted. Registration errors are corrected through Digital Registration Correction (DRC) system. Computer simulation confirms that the system performance is satisfactory during camera operations.
A design concept for compact HDTV camera is proposed. An exterimental camera has been developed in order to verify this concept. The HDTV camera employs three 2/3^- CCDs with spatially offset imaging between dual green channels. Limitting resolution of the camera is over 900TVL.
The development of an HDTV color camera employing three 2-millionpixel CCD image sensors is described. One unique characteristic of this camera is its high sensitivity, achieved by imaging devices overlaid with amorphous silicon. H-CCD registers are driven by a switching resonance method for low power consumption. Noise suppression for CCD image sensors is also described.
Learning capability of an optical associative memory using TFT-LCDs is experimentally shown. The system is based on a neural network model and it uses the orthogonal recollection method. This system can compensate for the optical ununiformity of each element, and can recoganize 10 chinese charactors. The optical crosstalk between two TFT-LCDs is the most important for learning capability.