To solve the problems of traditional relaying forward protocol: amplifying both the signals and the noise, possibly forwarding incorrect decoded signals and low reliability of decision on relay node in the two-way relay system, two Mutual Information Forwarding (MIF) schemes based on network coding were suggested. Firstly, the relay node forwards the hard decisions from two source nodes, and the reliability of these hard decisions. Then the receiving terminal detects and makes a decision on the combined result, which are based on the forwarding signal from relay node and its reliability, also the signal from another source node. The expressions for the received Signal-to-Noise Ratio (SNR) were derived, and numerical simulations were done in Additive White Gaussian Noise (AWGN) channel. At the same time, Monte Carlo simulation was conducted in AWGN channel and Rayleigh channel. Results show that both two proposed schemes under two kinds of channels have 1 dB and 1-2 dB gain than the Estimate-and-Forward (EF) scheme in Bit Error Ratio (BER) performance. The simulation results show that, the network coding scheme based on MIF significantly improves BER performance in relay forward of two-way relay network.

To prevent the leak of secrete information hidden in remote sensing image, a sample-based method for image inpainting was used in hiding secrete information in remote sensing image. The method selected the maximum region similar to secrete information image, and covered the secrete information block, thus hiding the secrete information. The method had not only strong invisibility, but also little influence on the use of remote sensing image. With reference to digital watermarking, whether users can see secrete information or not according to their authority status was realized, which facilitated the use of remote sensing image. Experimental results show that the method is feasible, and it improves the application security of remote sensing image.