Facsimile Bandwidth Compression Using Nonlinear Analog Processing

A system is presented for reducing the number of picture elements in a scanned image by an order of magnitude while preserving the detail contrast. Starting with high-resolution data a 3 × 3 matrix of input elements is compressed into a single output element using a nonlinear algorithm. Optimum results are obtained by including the data on the outer periphery of the input sample in the processing algorithm thus requiring a 5 × 5 input matrix. Experimental results are presented using weather satellite photographs. This system has application in those cases where the sensor has higher resolution capability than the communication link can support.