One of the most important aspects of compressed sensing (CS) theory is an efficient design of sensing matrices. These sensing matrices are accountable for the required signal compression at the encoder end and its exact or approximate reconstruction at the decoder end. This paper presents an in-depth review of a variety of compressed sensing matrices such as random matrices, deterministic matrices, structural matrices, and optimized sensing matrices used in compressed sensing. Moreover, this paper presents insights into different research gaps which will provide the direction for further research in compressed sensing area.