7
always expected to move the borescope downwards with
very limited horizontal movement. In such a way, the verti-
cal movement will show a constant positive value, and the
horizontal movement will have a value close to 0. However,
it is difficult to achieve with manual control. It can be
observed in Figure 8 that the borescope moved up and
down and rotated left and right with some spikes indicat-
ing sudden and large movements. In general, as compared
with the vertical movement, the horizontal movement is
generally much smaller with a few spikes. The top end of
a support pole has a slightly larger diameter than the bot-
tom end so that two support poles can be connected by
inserting the upper support pole into the lower one. Since
the connection is not rigid, unintentional rotation usually
occurs, especially when the insertion tube/cable is twisted.
Specifically, there are spikes in both vertical movement
and horizontal movement at a location slightly above 1.5
m from the roofline when scoping the dried borehole. The
vertical movement is negative, and the horizontal move-
ment is positive, indicating the borescope moved up and
rotated right. This potentially resulted from the operations
to disconnect the support poles. Normally, support poles
get disconnected by gravity and fall to the ground automat-
ically. However, it may get stuck with mud. The moving
pole would hit the floor. It needs to be lifted and discon-
nected with small twisting. Another problem encountered
during the recordings is that the connected support poles
can get stuck when moving in the boreholes, potentially
resulting from the shifting of roof strata or borehole devia-
tion. The borescope needs to move up and down to get
through, leading to the negative vertical movements.
In addition, Figure 8 shows that the averaged verti-
cal movement when scoping the freshly drilled borehole is
larger than that with the washed and dried borehole, which
is further larger than when scoping the freshly washed bore-
hole. This corresponds to the fact that it took the longest
time to scope the freshly washed borehole. It took about
103 seconds to scope the 2.56-m deep borehole under
freshly drilled condition, 243 seconds under freshly washed
condition, and 143 seconds under washed and dried con-
dition. Due to the water reflection in the freshly washed
borehole, the borescope moved slowly when scoping the
borehole, which took the longest time. With a longer time
to scope a borehole, there are more video frames to ana-
lyze and stitch, and relatively smaller offset values can be
expected between adjacent frames.
Furthermore, the borescope movement at the break
points was investigated. As marked in the plots in Figure 8,
there are 13 break points, and they were all from the video
recorded with the freshly washed borehole, while the videos
recorded with the freshly drilled borehole and the washed
and dried borehole were successfully stitched. It can be
found from the vertical movement in Figure 8 that, in
general, there were faster-than-normal movements at the
locations of break points when scoping the freshly washed
borehole. However, there were much faster vertical move-
ments or large rotations for the other two videos where the
images were still successfully stitched together. These obser-
vations indicate that, besides the borescope movement,
there are other factors affecting the image stitching process,
and the borehole condition is one of these potential fac-
tors. As shown in Figure 7 (b), water remains on the freshly
washed borehole wall and leads to reflection of camera light
on the borehole image, affecting the key point detection
and matching. This increases the difficulties in successfully
stitching the borehole image. The reflection can further
blur the borehole images with fast borescope movement.
Thus, the combination of abnormal borehole movement
and water reflection lead to the break points in the pan-
oramic image of the freshly washed borehole.
The Influence on the Identification of Lithology and
Geologic Features
The three videos recorded under different borehole condi-
tions were processed to generate the panoramic borehole
images, which are shown in Figure 9. It should be noted it
is hard to ensure that the side-view borescope followed the
same route and faced the same direction and minor differ-
ence may be observed in the panoramic borehole images
under different conditions. A few observations can be made
from Figure 9. Since there are multiple break points for
the freshly washed borehole, the corresponding panoramic
borehole image is longer than the other two borehole
images. Also, due to the large height (borehole depth) to
width (borehole diameter) ratio, it is difficult to demon-
strate the panoramic images in detail, and instead, two
marked locations were zoomed in to show the details.
The comparison of the three borehole images in
Figure 9 indicates that the presence of dusts has significant
and negative effects on the visualization and identifica-
tion of geologic features and lithology changes, while the
influence of water is significant and positive. The rock dust
generated during drilling could cover the borehole walls,
making it difficult to visually observe the geologic features
and lithology changes. This is clear when comparing the
enlarged view of three segments at the top marked location
in Figure 9. Also, Figure 9 shows that the rock dust almost
fills the fracture within the roof strata at the bottom marked
location. This adds difficulties in geologic mapping. In con-
trast, the washing of the borehole removed the dust and
always expected to move the borescope downwards with
very limited horizontal movement. In such a way, the verti-
cal movement will show a constant positive value, and the
horizontal movement will have a value close to 0. However,
it is difficult to achieve with manual control. It can be
observed in Figure 8 that the borescope moved up and
down and rotated left and right with some spikes indicat-
ing sudden and large movements. In general, as compared
with the vertical movement, the horizontal movement is
generally much smaller with a few spikes. The top end of
a support pole has a slightly larger diameter than the bot-
tom end so that two support poles can be connected by
inserting the upper support pole into the lower one. Since
the connection is not rigid, unintentional rotation usually
occurs, especially when the insertion tube/cable is twisted.
Specifically, there are spikes in both vertical movement
and horizontal movement at a location slightly above 1.5
m from the roofline when scoping the dried borehole. The
vertical movement is negative, and the horizontal move-
ment is positive, indicating the borescope moved up and
rotated right. This potentially resulted from the operations
to disconnect the support poles. Normally, support poles
get disconnected by gravity and fall to the ground automat-
ically. However, it may get stuck with mud. The moving
pole would hit the floor. It needs to be lifted and discon-
nected with small twisting. Another problem encountered
during the recordings is that the connected support poles
can get stuck when moving in the boreholes, potentially
resulting from the shifting of roof strata or borehole devia-
tion. The borescope needs to move up and down to get
through, leading to the negative vertical movements.
In addition, Figure 8 shows that the averaged verti-
cal movement when scoping the freshly drilled borehole is
larger than that with the washed and dried borehole, which
is further larger than when scoping the freshly washed bore-
hole. This corresponds to the fact that it took the longest
time to scope the freshly washed borehole. It took about
103 seconds to scope the 2.56-m deep borehole under
freshly drilled condition, 243 seconds under freshly washed
condition, and 143 seconds under washed and dried con-
dition. Due to the water reflection in the freshly washed
borehole, the borescope moved slowly when scoping the
borehole, which took the longest time. With a longer time
to scope a borehole, there are more video frames to ana-
lyze and stitch, and relatively smaller offset values can be
expected between adjacent frames.
Furthermore, the borescope movement at the break
points was investigated. As marked in the plots in Figure 8,
there are 13 break points, and they were all from the video
recorded with the freshly washed borehole, while the videos
recorded with the freshly drilled borehole and the washed
and dried borehole were successfully stitched. It can be
found from the vertical movement in Figure 8 that, in
general, there were faster-than-normal movements at the
locations of break points when scoping the freshly washed
borehole. However, there were much faster vertical move-
ments or large rotations for the other two videos where the
images were still successfully stitched together. These obser-
vations indicate that, besides the borescope movement,
there are other factors affecting the image stitching process,
and the borehole condition is one of these potential fac-
tors. As shown in Figure 7 (b), water remains on the freshly
washed borehole wall and leads to reflection of camera light
on the borehole image, affecting the key point detection
and matching. This increases the difficulties in successfully
stitching the borehole image. The reflection can further
blur the borehole images with fast borescope movement.
Thus, the combination of abnormal borehole movement
and water reflection lead to the break points in the pan-
oramic image of the freshly washed borehole.
The Influence on the Identification of Lithology and
Geologic Features
The three videos recorded under different borehole condi-
tions were processed to generate the panoramic borehole
images, which are shown in Figure 9. It should be noted it
is hard to ensure that the side-view borescope followed the
same route and faced the same direction and minor differ-
ence may be observed in the panoramic borehole images
under different conditions. A few observations can be made
from Figure 9. Since there are multiple break points for
the freshly washed borehole, the corresponding panoramic
borehole image is longer than the other two borehole
images. Also, due to the large height (borehole depth) to
width (borehole diameter) ratio, it is difficult to demon-
strate the panoramic images in detail, and instead, two
marked locations were zoomed in to show the details.
The comparison of the three borehole images in
Figure 9 indicates that the presence of dusts has significant
and negative effects on the visualization and identifica-
tion of geologic features and lithology changes, while the
influence of water is significant and positive. The rock dust
generated during drilling could cover the borehole walls,
making it difficult to visually observe the geologic features
and lithology changes. This is clear when comparing the
enlarged view of three segments at the top marked location
in Figure 9. Also, Figure 9 shows that the rock dust almost
fills the fracture within the roof strata at the bottom marked
location. This adds difficulties in geologic mapping. In con-
trast, the washing of the borehole removed the dust and