Compressive Sensing (CS) is a modern technique that samples signals at a rate less than Nyquist rate.The sensing matrix will obey the Restricted Isometry Property (RIP) to determine the proper retrieval from the compressed measurements.In general, the Gauussian random matrix or Bernoulli random matrix constructs this type of matrix. Recently several researchers have been constructing measuring matrix based on chaotic sequences. In this work, we are extensively investigating the measurement matrices construction with various fractional-order hyperchaoticsequences.Fractional-order hyperchaotic sequences exhibit higher randomness and nonlinearity than integral-order chaotic sequences due to the complicated geometric representation.Weemploythe generated measurement matrices to compressive sensing of the image, and the reconstruction quality is evaluated for each generated measurement matrices. The experimentally observeddata reveals that the measurement matrices generated by fractional-order hyperchaotic sequences have rich dynamics and high randomness, making them ideal for constructing secure measurement matrices with good quality of reconstruction. Meanwhile, the performance of the different fractional-order hyperchaotic sensing matrices is nearly equal.
Volume 12 | 03-Special Issue