Digital beamforming synthetic aperture radar (SAR) in elevation is a promising technique to realize the increasing requirements of high-resolution and wide-swath imaging in remote sensing, although it suffers from a dense sampling rate and heavy data volume. This poses serious challenges to onboard transmission and storage of satellites. Additionally, the traditional beamforming technique suffers degraded imaging capacities caused by the incoherence between the received signals of different channels. To overcome these deficiencies, this paper proposes an improved method for the focusing of multichannel SAR raw data based on the framework of compressed sensing (CS). Through adaptive measurement matrix construction for each subchannel, the proposed approach can take the misregistration information of all received signals into account, providing an accurate imagery recovery of sparse reflectivity. Compared with conventional processing, the suggested technique not only provides a more efficient data acquisition scheme for orbital configurations, but also carries the potential to eliminate the migration effects of imaging amplitude and resolution. Furthermore, it is demonstrated that the proposed implementation is equipped with additional imaging benefits, such as superresolution and lower sidelobes. Numerical and experimental results have validated these favorable performances of the suggested strategy.