Two-dimensional system identification of a digital laser electrophotographic system

Digital laser printing transforms a desired two-dimensional image into a printed image using two main processes, i.e., image processing and marking. The required image processing plays a role similar to a controller in a control system, which converts a two-dimensional continuous-tone image (reference command) to a two-dimensional halftone image (control input). The marking process or system plays the role of an open-loop plant, which accepts a halftoned or binary image as its input and generates printed image as its output. For identification of the two-dimensional marking system, the average tone of an input image may be regarded as an operating point. A linearized model or MTF corresponding to each tone level can then be identified and used for subsequent design of the halftone algorithm. In this paper, we revisit methods previously used for performing MTF identification. We improve upon those methods by taking into account the need for varying the amplitude of the sinusoid at different frequency region for different tone levels so as not to excite the nonlinearity of the system and preserve acceptable S/N ratio.