Abstract:

The core problem of compressed sensing theory is how to find an efficient and fast reconstruction algorithm. The existing reconstruction algorithms (such as orthogonal matching pursuit) have some defects: slow reconstruction, the reconstruction algorithm is carried out under a given number of iteration conditions, and the adaptation is reduced by this compulsory stop. An improved genetic algorithm (IGA) combining with orthogonal matching pursuit (OMP) algorithm is carried out to construct the reconstruction matrix. First, an improved genetic algorithm is used to select the current maximum redundancy column vector from the measurement matrix columns with an optimal chromosome method. Then subtract the part of columns with optimal chromosome from the measurement matrix, and repeat iteration until it meets the reconstruction accuracy. Simulation results show that, compared with the existing reconstruction algorithms under the same conditions, timeconsuming of the algorithm reduces 5 s and the size of the measurement matrix reduces about 10%. This method can stop iteration adaptively under the condition of reconstruction signal with unknown sparseness.