Mid-infrared (mid-IR) imaging and spectroscopic techniques have been rapidly evolving in recent years, primarily due to a multitude of applications within diverse fields such as biomedical imaging, chemical sensing, and food quality inspection. Mid-IR upconversion detection is a promising tool for exploiting some of these applications. In this paper, various characteristics of mid-IR upconversion imaging in the femtosecond regime are investigated using a 4f imaging setup. A fraction of the 100 fs, 80 MHz output from a Ti:sapphire laser is used to synchronously pump an optical parametric oscillator, generating 200 fs mid-IR pulses tunable across the 2.7–4.0 μm wavelength range. The signal-carrying mid-IR pulses are detected by upconversion with the remaining fraction of the original pump beam inside a bulk LiNbO3 crystal, generating an upconverted field in the visible/near-IR range, enabling silicon-based CCD detection. Using the same pump source for generation and detection ensures temporal overlap of pulses inside the nonlinear crystal used for upconversion, thus resulting in high conversion efficiency even in a single-pass configuration. A theory is developed to calculate relevant acceptance parameters, considering the large spectral bandwidths and the reduced interaction length due to group velocity mismatch, both associated with ultrashort pulses. Furthermore, the resolution of this ultrashort-pulsed upconversion imaging system is described. It is demonstrated that the increase in acceptance bandwidth leads to increased blurring in the upconverted images. The presented theory is consistent with experimental observations.