Algorithm 1:$N_{ij}$ calculation algorithm in 3D space.
1:	Initialize pedestrian detection results, target detection boxes and confidence level.
2:	Create a List $L_{\mathit{targets}}$ of every unit, a centroid variable C and a pedestrian count variable $N_{count}$.
3:	While ($i_{\min }$ < $i$ < $i_{\max }$ && $j_{\min }$ < $j$ < $j_{\max }$) do
4:	If (target detection boxes) then
5:	$L_{\mathit{targets}}$ = $L_{\mathit{targets}}$ + 1
6:	for each $L_{\mathit{targets}}$ do
7:	If (C is under the range of the corner coordinates of $L_{\mathit{targets}}$) then
8:	$N_{count}$ $=N_{count}$ + 1
9:	end if
10:	end for
11:	$N_{ij}$ =$N_{count}$
12:	End while
13:	Return $N_{ij}$