Geodetski list, Vol. 69 (92) No. 4, 2015.
Izvorni znanstveni članak
Directional Remote Sensing
Tao Cheng
; Department of Mechanics, Automotive & Transportation Engineering Institute, Guangxi University of Science and Technology, Liuzhou, Guangxi, China
Sažetak
Concepts of directional remote sensing are put forward based on two-dimensional compressive sensing. Very little measured data are required to acquire and reconstruct change areas in directional remote sensing. The measured data in one-dimensional compressive sensing not only keep the energy of a sparse signal, but also inherit the sparse signal’s direction information. However, direction information can't be applied to reconstruction and test of a sparse signal in one-dimensional compression sensing. The two-dimensional compressive sensing model is proposed based on sparse features of change areas in remote sensing. Moreover, a sparse signal reconstruction algorithm (two-step reconstruction method, 2SRM) is proposed based on two-dimensional compressive sensing by use of the energy and direction information. The theoretical analysis and experimental results show the signal reconstruction ability of 2SRM is stronger. SNR (Signal to Noise Ratio) and PSNR (Peak Signal to Noise Ratio) of 2SRM increase by 16.57 dB as compared with a single traditional reconstruction algorithm at most.
Ključne riječi
directional remote sensing; change detection; two-dimensional compressive sensing (2DCS); structure prior information; two-step reconstruction method (2SRM)
Hrčak ID:
150715
URI
Datum izdavanja:
23.12.2015.
Posjeta: 1.539 *