The conventional optical imaging system is based on Nyquist sampling theorem, which seriously wastes valuable sensing resources and increases overall system cost. Applying compressive sensing theory to practical imaging systems can overcome the conventional imager′s design idea. Hardware is used to implement the non-adaptive linear projection of target objects and compressive imaging by using a smaller number of photonic detectors is realized. This paper reviews several key problems in compressive sensing theory and discusses the existing difficult problems. The principle and difficulties of the compressive imaging systems are also discussed in detail.