Vision Based Wall Following Framework: A Case Study With HSR Robot for Cleaning Application

View full-text article in PMC

. 2020 Jun 10;20(11):3298. doi: 10.3390/s20113298

Algorithm 1 Wall Following and Obstacle Avoidance

1:
Input:
2:
Image capture from HSR vision system
3:
Output:
4:
Alert signal ( $A l e r t_H S R$ ) to HSR for obstacle avoidance and safe cleaning
5:
Initialize:
6:
C is an Camera
7:
F is an Image frame
8:
Segment_array: hold the segmented image
9:
Wall_Segment: hold the segmented wall image
10:
Floor_Segment: hold the segmented floor image
11:
Object_wall: hold the detected objects on wall
12:
Object_floor: hold the detected objects on floor
13:
Initialize End
14:
Begin:
15:
while (1) do
16:
F = Capture Frame (C)
17:
$S e g m e n t_a r r a y$ = FCN8_ segmentation(F)
18:
$W a l l_S e g m e n t$ = $S e g m e n t_a r r a y$ && $f l o o r_m a s k$ ; mask the floor
19:
$F l o o r_S e g m e n t$ = $S e g m e n t_a r r a y$ && $w a l l_m a s k$ ; mask the wall
20:
$O b j e c t_w a l l$ = SSD_MobileNet ( $w a l l_s e g m e n t$ ) ; Object detection on segmented wall region
21:
$O b j e c t_f l o o r$ = SSD_MobileNet ( $f l o o r_s e g m e n t$ ) ; Object detection on segmented floor region
22:
Distance_floor-obstacle_wall ( $w a l l_s e g m e n t$ , $f l o o r_s e g m e n t$ ) ; compute the distance of the detected floor object from wall
23:
Wall_Object-distance_ground( $w a l l_s e g m e n t$ , $f l o o r_s e g m e n t$ ) ; compute the distance of the detected wall object from ground
24:
if ( $D i s t a n c e_f l o o r - o b s t a c l e_w a l l$ ) then
25:
Alert_HSR( obstacle near by wall) ; Alert to HSR to avoid obstacle on path
26:
end if
27:
if ( $W a l l_O b j e c t - d i s t a n c e_g r o u n d$ ) then
28:
Alert_HSR( object on wall) ; Alert to HSR for cleaning modules safe operation
29:
goto $\overset{}{\to}$ Begin
30:
else
31:
goto $\overset{}{\to}$ Begin
32:
end if
33:
end while