As a typical computational imaging modality, single-pixel imaging uses a single-pixel detector to measure the light intensities reflected or transmitted from the target after its interaction with a series of patterns. By calculating the correlation of the measured intensities and relevant patterns with different reconstruction algorithms, the target image can be recovered. Compared with multi-pixel detector (i.e. CCD or CMOS), single-pixel imaging overcomes hardware limitations and the detection efficiency is higher, and the response is faster in some special wavebands. Metasurfaces are a kind of artificial two-dimensional materials consisting of an array of subwavelength metallic or dielectric unit cells. In the optical wavelength regime, the metasurface can display various holograms by adjusting different degrees of freedom of incident light. In the microwave regime, the metasurface can couple with the waveguide and emit various radiating modes as patterns. The research background, imaging principle, reconstruction algorithms of single-pixel imaging, and the research background of metasurface imaging were reviewed. The discussion of relevant works was mainly focused on the combination of single-pixel imaging and metasurface imaging in optical and microwave regimes, and finally a perspective on the potential development in future was proposed.