MEMPHIS is an experimental millimeter-wave synthetic aperture radar (SAR) system that acquires cross-track multibaseline interferometric data at high resolution in a single pass, using four receive horns. In this paper, we present the SAR system and navigation data, and propose a processing chain from the raw data input to a digital surface model (DSM) output. This processing chain includes full bandwidth reconstruction of the stepped- frequency SAR data, azimuth focusing with an Extended Omega-K algorithm, generation of interferograms for each available baseline, phase unwrapping using the multibaseline data, and phase- to-height conversion. The hardware and processing chain were validated through the analysis of experimental Ka-band data. The SAR image resolution was measured with point targets and found to be ￼ ￼ ￼ and ￼ ￼ ￼ ￼ coarser than the theoretical value in range and azimuth, respectively. The geolocation accuracy was typically better than 0.1 m in range and 0.2 m in azimuth. Observed depression angle- dependent interferometric phase errors were successfully removed using a correction function derived from the InSAR data. Investiga- tion of the interferometric phase noise showed the utility of a multi- baseline antenna setup; the number of looks and filter size used for the DSM generation were also derived from this analysis. The results showed that in grassland areas, the height difference between the ~2 m-resolution InSAR DSMs and the reference ALS models was 0 ±0.25 m.