| |
| Compressive sensing is introduced at first. Compressive sensing not only can compress signal, but also can encrypt signal. But the constraint is that the column of measurement matrix is equal to the row of the signal. In this paper, the model of semi-tensor compressive sensing is given. The model is composed of measurement matrix and auxiliary matrix. The measurement matrix is generated by Logistic chaotic system. The auxiliary matrix is generated by Tent chaotic system. The proposed model breaks the restriction of matrix multiplication, so the matrix can be multiplied when their dimensions do not match. Moreover, the size of measurement matrix of semi-tensor compressive sensing is smaller than compressive sensing, which reduces the transmission overhead and storage space greatly. The experimental simulations show the factor that affect the successful recovery of the signal, including sparsity and the number of measurements. When compression ratio is fixed, the larger the sparsity is, the smaller the percentage of successful recovery is. When sparsity is fixed, the more the number of measurements is, the larger the percentage of successful recovery is. In the experiment, the compression ratio is 0.5, when sparsity is less than 15, the percentage of successful recovery is equal to 1 regardless of the size of measurement matrix. |
|
| Keywords:signal and information processing;compressive sensing;semi-tensor compressive sensing;the percentage of recovery |
|
