Image Adaptive Reconstruction Based on Compressive Sensing and the Genetic Algorithm via ROMP

Compressive sensing (CS) effectively relieved the pressure of sensor network. CS proposed doing sampling and compression at the same time which avoided the waste of a large number of samples and large redundant data. In CS, reconstruction algorithm is a critical technology. In this paper, we use greedy algorithm Regularized Orthogonal Matching Pursuit (ROMP) to reconstruct signal. Because the recovery effect of the greedy algorithm depends on characteristics of the signal and the greedy algorithm needs a minimum measurement value, an adaptive image recovery method is proposed for the improvement of image recovery performance. In our method, we combine ROMP and the genetic algorithm (GA). The number of measurements N and the sparse of signal m in ROMP are respectively defined according to the GA and prior knowledge of signal. We compare the new method with other methods under the same test conditions, the new method can effectively reconstruct signal, reduce errors and reconstruction effects are superior to other methods.