In recent years, mid-infrared parametric upconversion imaging, a nonlinear optical method that involves converting mid-infrared light into visible images, has significantly advanced and has shown considerable potential for various applications, including biomedical imaging and remote sensing. While diffraction-based parametric upconversion imaging modeling in standard thin birefringence crystals have been addressed, the numerical framework developed so far fails to address long aperiodic poled crystals. Specifically, diffraction-based analysis of the recent broadband adiabatic frequency upconversion imaging, which allows simultaneous image upconversion of extremely broadband signals is still lacking. Here, we introduce a diffraction-based numerical simulation framework for predicting the evolution of the nonlinear image/signal generation in upconversion imaging systems. This generalized framework can handle both periodically and aperiodically poled crystal designs. Specifically, the model captures faithfully and addresses the varying image magnification arising from upconversion at a Fourier plane of a multiwavelength object. The numerical simulations are validated by experimental measurements of broadband upconversion 3-5 µm mid-IR images to the visible-NIR, showing a good agreement. Moreover, the model allows the exploration of the trade-offs in the spectral span when moving to the full visible range. Our numerical framework will be useful for the interpretation of experimental results obtained in an imaging setting with nonlinear optical elements.