Algorithm 1: Maximum division algorithm.

Input: $Extent$ is the coordinate range of original RS images; $cellSize$ is the pixel value of RS images; $tileSize$ is the slice size of image tile.  Output: $level$ is final number of division layers. 1:Input coordinate range of original RS images $Extent$, and evaluate the width and height of RS images: $W_0\leftarrow Extent.width$, $H_0\leftarrow Extent.height$; 2:Evaluate resolution of i-th layer $(1\le i\le 20)$: $Re{s}_w=W_0/(2^i\times tileSize)$, then $i=i+1$; 3:If $cellSize.width+K\ge Re{s}_w$ (K is a correction constant), stop evaluation and achieve the value i as maximum value of division layers. Or if $cellSize.width+K<Re{s}_w$, repeat Step 2 to evaluate resolution of $(i+1)$-th layer; 4:Reassign $W_0\leftarrow H_0$, and evaluate $Re{s}_h$ according to processing method in Step 2 and Step 3; 5:Evaluate the maximum value $level\leftarrow max(Re{s}_w,Re{s}_h)$; 6:$\mathbf{Return}$ final number of division layers $level$.