This thesis summarizes the knowledge of the discrete tomography and introduces thecompressive sensing theory which is developing fast during these years. The paper applies thesparse recovery theory to the discrete tomography and the simulation usingl1algorithm showsits valid in low noise conditions. Meanwhile, the paper researches the image recovery from lossdata by using compressed sensing. The experimental results show that we can get the idealreconstructed pictures while both the speed and the complexity have the advantages and theapplication prospects. |