Algorithm 1 Remote sensing image fusion via NSST

Input: the source remote sensing images A and B Output: fused image F Parameters: the number of NSST decomposition levels—N; the number of directions at each decomposition level—$D\left(l\right),l\in \left[1,N\right]$ Step 1: NSST decomposition The input images A and B are decomposed into low- and high-frequency sub-bands $\left\{L_A,H_A^{l,d}\right\}$ and $\left\{L_B,H_B^{l,d}\right\}$, respectively. Step 2: low-frequency band fusion rule (1) The saliency maps $\left(S_{L_A},S_{L_B}\right)$ and the corresponding weight matrices $\left(W_{L_A},W_{L_B}\right)$ of the low-frequency bands are calculated by Equations (1)–(5). (2) The fused low-frequency band LF is obtained by Equation (6). Step 3: high-frequency band fusion rule (1) The SML of the high-frequency bands is constructed via Equations (7)–(8). (2) The fused high-frequency band HF is computed by Equation (9). Step 4: inverse NSST and image reconstruction The fused image F is reconstructed by inverse NSST performed on the fused low- and high-frequency bands $\left\{L_F,H_F^{l,d}\right\}$.