CID Processor Chip Development.

摘要

TV bandwidth reduction is required to achieve jam resistance in guided weapon and remotely piloted vehicle applications. One way to achieve bandwidth reduction is to eliminate redundant information prior to transmission. The high degree of redundancy in most images can be reduced by employing transform techniques to uncorrelate the scene information. A number of coding techniques can then be used to allocate the available bandwidth to the significant spatial information. A number of transforms such as the Hadamard, Fourier, Haar, Cosine and Slant could be used. Until recently, however, the real time realizations of these transformations were questionable. The use of large numbers of logic, memory, and linear devices would be required to process the video signal in a conventional manner. Charge-injection device technology can be used to combine such widely diverse functions as imaging, memory, and signal processing into a single device to achieve low system cost, size, weight, and power consumption. The process of computing the Hadamard transform requires linear combinations, sums and differences, of the picture information. This process can be easily mechanized on the focal plane of a CID image sensor by sensing combinations of the signal charge from a number of image sensing sites (pixels) rather than the individual pixel magnitudes.