ABOUT THE SPEAKER
Lian Pin Koh - Drones ecologist
Lian Pin Koh expands conservation efforts by championing the use of low-cost autonomous aerial vehicles.

Why you should listen

Lian Pin Koh is a relentless tinkerer and science fiction movie geek, though most know him as an environmental scientist. His dreams of combining these interests led him to cofound, with colleague Serge Wich, the site ConservationDrones.org, a project dedicated to gathering intelligence on forests and wildlife through the use of low-cost unmanned flying machines.

Ground surveys are expensive, and are not conducted at a sufficient frequency. Furthermore, some remote tropical forests have never been really surveyed for biodiversity. Koh's machines have already collected valuable information in Sumatra, Congo, Gabon, and Madagascar.

He is an assistant professor of applied ecology and conservation at the ETH Zurich.

More profile about the speaker
Lian Pin Koh | Speaker | TED.com
TEDGlobal 2013

Lian Pin Koh: A drone's-eye view of conservation

Filmed:
643,849 views

Ecologist Lian Pin Koh makes a persuasive case for using drones to protect the world's forests and wildlife. These lightweight autonomous flying vehicles can track animals in their natural habitat, monitor the health of rainforests, even combat crime by detecting poachers via thermal imaging. Added bonus? They're also entirely affordable.
- Drones ecologist
Lian Pin Koh expands conservation efforts by championing the use of low-cost autonomous aerial vehicles. Full bio

Double-click the English transcript below to play the video.

00:13
When we think of Nepal,
0
1360
1932
00:15
we tend to think of the snow-capped mountains
1
3292
2455
00:17
of the Himalayas,
2
5747
1272
00:19
the crystal-clear still waters of its alpine lakes,
3
7019
3449
00:22
or the huge expanse of its grasslands.
4
10468
3046
00:25
What some of us may not realize
5
13514
2207
00:27
is that in the Himalayan foothills,
6
15721
2324
00:30
where the climate is much warmer
7
18045
2174
00:32
and the landscape much greener,
8
20219
2088
00:34
there lives a great diversity of wildlife,
9
22307
2426
00:36
including the one-horned rhinoceros,
10
24733
2736
00:39
the Asian elephant
11
27469
1674
00:41
and the Bengal tiger.
12
29143
2075
00:43
But unfortunately, these animals
13
31218
1852
00:45
are under constant threat from poachers
14
33070
2663
00:47
who hunt and kill them for their body parts.
15
35733
3312
00:51
To stop the killing of these animals,
16
39045
2209
00:53
battalions of soldiers and rangers
17
41254
2598
00:55
are sent to protect Nepal's national parks,
18
43852
3249
00:59
but that is not an easy task,
19
47101
1625
01:00
because these soldiers have to patrol
20
48726
2671
01:03
thousands of hectares of forests on foot
21
51397
2991
01:06
or elephant backs.
22
54388
1641
01:08
It is also risky for these soldiers
23
56029
2485
01:10
when they get into gunfights with poachers,
24
58514
2381
01:12
and therefore Nepal is always looking
25
60895
2016
01:14
for new ways to help with protecting the forests
26
62911
3296
01:18
and wildlife.
27
66207
1977
01:20
Well recently, Nepal acquired a new tool
28
68184
2509
01:22
in the fight against wildlife crime,
29
70693
2713
01:25
and these are drones,
30
73406
1706
01:27
or more specifically, conservation drones.
31
75112
3274
01:30
For about a year now, my colleagues and I
32
78386
2514
01:32
have been building drones for Nepal
33
80900
2260
01:35
and training the park protection personnel
34
83160
2498
01:37
on the use of these drones.
35
85658
1978
01:39
Not only does a drone give you
36
87636
1632
01:41
a bird's-eye view of the landscape,
37
89268
2474
01:43
but it also allows you to capture detailed,
38
91742
3035
01:46
high-resolution images of objects on the ground.
39
94777
3982
01:50
This, for example, is a pair of rhinoceros
40
98759
3509
01:54
taking a cooling bath on a hot summer day
41
102268
2937
01:57
in the lowlands of Nepal.
42
105205
1963
01:59
Now we believe that drones have
43
107168
2231
02:01
tremendous potential,
44
109399
1808
02:03
not only for combating wildlife crime,
45
111207
2778
02:05
but also for monitoring the health
46
113985
1963
02:07
of these wildlife populations.
47
115948
2539
02:10
So what is a drone?
48
118487
1850
02:12
Well, the kind of drone I'm talking about
49
120337
2509
02:14
is simply a model aircraft
50
122846
2101
02:16
fitted with an autopilot system,
51
124947
2445
02:19
and this autopilot unit contains a tiny computer,
52
127392
4269
02:23
a GPS, a compass, a barometric altimeter
53
131661
4054
02:27
and a few other sensors.
54
135715
2259
02:29
Now a drone like this
55
137974
1419
02:31
is meant to carry a useful payload,
56
139393
2206
02:33
such as a video camera
57
141599
1783
02:35
or a photographic camera.
58
143382
1988
02:37
It also requires a software that allows the user
59
145370
3349
02:40
to program a mission,
60
148719
1701
02:42
to tell the drone where to go.
61
150420
2406
02:44
Now people I talk to are often surprised
62
152826
2077
02:46
when they hear that these are the only
63
154903
2331
02:49
four components that make a conservation drone,
64
157234
3107
02:52
but they are even more surprised
65
160341
1504
02:53
when I tell them how affordable
these components are.
66
161845
3500
02:57
The facts is, a conservation drone
67
165345
2286
02:59
doesn't cost very much more than
68
167631
2259
03:01
a good laptop computer
69
169890
1977
03:03
or a decent pair of binoculars.
70
171867
3259
03:07
So now that you've built
your own conservation drone,
71
175126
2972
03:10
you probably want to go fly it,
72
178098
2162
03:12
but how does one fly a drone?
73
180260
1933
03:14
Well, actually, you don't,
74
182193
1592
03:15
because the drone flies itself.
75
183785
2490
03:18
All you have to do is to program a mission
76
186275
3130
03:21
to tell the drone where to fly.
77
189405
2260
03:23
But you simply do that by clicking on
78
191665
2319
03:25
a few way points on the Google Maps interface
79
193984
3249
03:29
using the open-source software.
80
197233
2250
03:31
Those missions could be as simple
81
199483
1947
03:33
as just a few way points,
82
201430
2026
03:35
or they could be slightly longer
and more complicated,
83
203456
3377
03:38
to fly along a river system.
84
206833
2421
03:41
Sometimes, we fly the drone
in a lawnmower-type pattern
85
209254
3849
03:45
and take pictures of that area,
86
213103
2120
03:47
and those pictures can be processed
87
215223
1971
03:49
to produce a map of that forest.
88
217194
2739
03:51
Other researchers might want to fly the drone
89
219933
2603
03:54
along the boundaries of a forest
90
222536
2142
03:56
to watch out for poachers or people
91
224678
2171
03:58
who might be trying to enter the forest illegally.
92
226849
2918
04:01
Now whatever your mission is,
93
229767
1880
04:03
once you've programmed it,
94
231647
1473
04:05
you simply upload it to the autopilot system,
95
233120
2983
04:08
bring your drone to the field,
96
236103
1640
04:09
and launch it simply by tossing it in the air.
97
237743
3215
04:12
And often we'll go about this mission
98
240958
2342
04:15
taking pictures or videos along the way,
99
243300
3060
04:18
and usually at that point,
100
246360
1934
04:20
we will go grab ourselves a cup of coffee,
101
248294
2217
04:22
sit back, and relax for the next few minutes,
102
250511
2930
04:25
although some of us sit back
and panic for the next few minutes
103
253441
3511
04:28
worrying that the drone will not return.
104
256952
2218
04:31
Usually it does, and when it does,
105
259170
2273
04:33
it even lands automatically.
106
261443
3406
04:36
So what can we do with a conservation drone?
107
264849
2907
04:39
Well, when we built our first prototype drone,
108
267756
3393
04:43
our main objective was to fly it over
109
271149
2518
04:45
a remote rainforest in North Sumatra, Indonesia,
110
273667
3889
04:49
to look for the nest of a species of great ape
111
277556
3049
04:52
known as the orangutan.
112
280605
2156
04:54
The reason we wanted to do that was because
113
282761
2656
04:57
we needed to know how many individuals
114
285417
2029
04:59
of this species are still left in that forest.
115
287446
3387
05:02
Now the traditional method of surveying
116
290833
2186
05:05
for orangutans is to walk the forest on foot
117
293019
3602
05:08
carrying heavy equipment
118
296621
1887
05:10
and to use a pair of binoculars
to look up in the treetops
119
298508
3650
05:14
where you might find an orangutan or its nest.
120
302158
3418
05:17
Now as you can imagine,
121
305576
1726
05:19
that is a very time-consuming, labor-intensive,
122
307302
2970
05:22
and costly process,
123
310272
1718
05:23
so we were hoping that drones
124
311990
1796
05:25
could significantly reduce the cost of surveying
125
313786
3325
05:29
for orangutan populations in Indonesia
126
317111
2567
05:31
and elsewhere in Southeast Asia.
127
319678
2467
05:34
So we were very excited when we captured
128
322145
2355
05:36
our first pair of orangutan nests on camera.
129
324500
3170
05:39
And this is it; this is the first ever picture
130
327670
3164
05:42
of orangutan nests taken with a drone.
131
330834
3040
05:45
Since then we have taken pictures
132
333874
2024
05:47
of dozens of these nests
133
335898
1618
05:49
from around various parts of Southeast Asia,
134
337516
2769
05:52
and we're now working with computer scientists
135
340285
2243
05:54
to develop algorithms that can automatically count
136
342528
3919
05:58
the number of nests from the thousands
137
346447
2017
06:00
of photos we've collected so far.
138
348464
3119
06:03
But nests are not the only objects
139
351583
1958
06:05
these drones can detect.
140
353541
1790
06:07
This is a wild orangutan
141
355331
2309
06:09
happily feeding on top of a palm tree,
142
357640
2521
06:12
seemingly oblivious to our drone
that was flying overhead,
143
360161
3552
06:15
not once but several times.
144
363713
2384
06:18
We've also taken pictures of other animals
145
366097
3018
06:21
including forest buffalos in Gabon,
146
369115
3121
06:24
elephants, and even turtle nests.
147
372236
3215
06:27
But besides taking pictures
of just the animals themselves,
148
375451
3540
06:30
we also take pictures of the habitats
these animals live in,
149
378991
4001
06:34
because we want to keep track
150
382992
1614
06:36
of the health of these habitats.
151
384606
2954
06:39
Sometimes, we zoom out a little
152
387560
2219
06:41
and look at other things that might be happening
153
389779
2253
06:44
in the landscape.
154
392032
1464
06:45
This is an oil palm plantation in Sumatra.
155
393496
3300
06:48
Now oil palm is a major driver of deforestation
156
396796
3242
06:52
in that part of the world,
157
400038
1522
06:53
so we wanted to use this new drone technology
158
401560
3057
06:56
to keep track of the spread of these plantations
159
404617
3115
06:59
in Southeast Asia.
160
407732
2010
07:01
But drones could also be used to keep track of
161
409742
2363
07:04
illegal logging activities.
162
412105
2484
07:06
This is a recently logged forest,
163
414589
2441
07:09
again in Sumatra.
164
417030
1355
07:10
You could even still see the processed
165
418385
2159
07:12
wooden planks left on the ground.
166
420544
3628
07:16
But perhaps the most exciting part
167
424172
2289
07:18
about taking pictures from the air is
168
426461
2566
07:21
we could later stitch these pictures together
169
429027
2426
07:23
using special software to create a map
170
431453
2863
07:26
of the entire landscape, and this map
171
434316
2767
07:29
gives us crucial information
172
437083
2015
07:31
for monitoring land use change,
173
439098
2268
07:33
to let us know where and when
plantations might be expanding,
174
441366
3747
07:37
where forests might be contracting,
175
445113
2277
07:39
or where fires might be breaking out.
176
447390
3382
07:42
Aerial images could also be processed
177
450772
2232
07:45
to produce three-dimensional
178
453004
2309
07:47
computer models of forests.
179
455313
1997
07:49
Now these models are not just visually appealing,
180
457310
2953
07:52
but they are also geometrically accurate,
181
460263
2501
07:54
which means researchers can now measure
182
462764
2777
07:57
the distance between trees,
183
465541
2125
07:59
calculate surface area, the volume of vegetation,
184
467666
3295
08:02
and so on, all of which are important information
185
470961
3199
08:06
for monitoring the health of these forests.
186
474160
3313
08:09
Recently, we've also begun experimenting
187
477473
2453
08:11
with thermal imaging cameras.
188
479926
2064
08:13
Now these cameras can detect
189
481990
1889
08:15
heat-emitting objects from the ground,
190
483879
1974
08:17
and therefore they are very useful
for detecting poachers or their campfires at night.
191
485864
7372
08:25
So I've told you quite a lot about
192
493236
1884
08:27
what conservation drones are,
193
495120
2216
08:29
how you might operate one of these drones,
194
497336
2317
08:31
and what a drone could do for you.
195
499653
2009
08:33
I will now tell you where conservation drones
196
501662
2236
08:35
are being used around the world.
197
503898
2455
08:38
We built our first prototype drones in Switzerland.
198
506353
3053
08:41
We brought a few of these to Indonesia
199
509406
2141
08:43
for the first few test flights.
200
511547
2143
08:45
Since then, we've been building drones
201
513690
1815
08:47
for our collaborators from around the world,
202
515505
2682
08:50
and these include fellow biologists
203
518187
2180
08:52
and partners from major conservation organizations.
204
520367
3928
08:56
Perhaps the best and most rewarding part
205
524295
2290
08:58
about working with these collaborators
206
526585
2126
09:00
is the feedback they give us
207
528711
1480
09:02
on how to improve our drones.
208
530191
2643
09:04
Building drones for us is
209
532834
2131
09:06
a constant work in progress.
210
534965
2164
09:09
We are constantly trying to improve them in terms of
211
537129
3077
09:12
their range, their ruggedness,
212
540206
2614
09:14
and the amount of payload they can carry.
213
542820
2678
09:17
We also work with collaborators
214
545498
2042
09:19
to discover new ways of using these drones.
215
547540
3287
09:22
For example, camera traps are a common tool
216
550827
3280
09:26
used by biologists to take pictures of shy animals
217
554107
3895
09:30
hiding in the forests,
218
558002
1561
09:31
but these are motion-activated cameras,
219
559563
2334
09:33
so they snap a picture every time an animal
220
561897
2641
09:36
crosses their path.
221
564538
1672
09:38
But the problem with camera traps
222
566210
1552
09:39
is that the researcher has to go back to the forest
223
567762
3743
09:43
every so often to retrieve those images,
224
571505
3453
09:46
and that takes a lot of time,
225
574958
1652
09:48
especially if there are dozens
226
576610
1834
09:50
or hundreds of these cameras placed in the forest.
227
578444
3128
09:53
Now a drone could be designed to perform the task
228
581572
2811
09:56
much more efficiently.
229
584383
1724
09:58
This drone, carrying a special sensor,
230
586107
2670
10:00
could be flown over the forest
231
588777
2341
10:03
and remotely download these images
232
591118
2747
10:05
from wi-fi–enabled cameras.
233
593865
4137
10:10
Radio collars are another tool
234
598002
2137
10:12
that's commonly used by biologists.
235
600139
2248
10:14
Now these collars are put onto animals.
236
602387
2667
10:17
They transmit a radio signal which allows
237
605054
2801
10:19
the researcher to track the movements
of these animals across the landscape.
238
607855
5147
10:25
But the traditional way of tracking animals
239
613002
2657
10:27
is pretty ridiculous,
240
615659
2222
10:29
because it requires the researcher to be walking
241
617881
2944
10:32
on the ground carrying a huge
and cumbersome radio antenna,
242
620825
4114
10:36
not unlike those old TV antennae we used to have
243
624939
2873
10:39
on our rooftops. Some of us still do.
244
627812
2968
10:42
A drone could be used to do the same job
245
630780
2265
10:45
much more efficiently.
246
633045
1659
10:46
Why not equip a drone
247
634704
1706
10:48
with a scanning radio receiver,
248
636410
2003
10:50
fly that over the forest canopy
249
638413
1994
10:52
in a certain pattern
250
640407
1538
10:53
which would allow the user or the operator
251
641945
2792
10:56
to triangulate the location
252
644737
2582
10:59
of these radio-collared animals remotely
253
647319
2987
11:02
without having to step foot in the forest.
254
650306
2973
11:05
A third and perhaps most exciting way
255
653279
2503
11:07
of using these drones
256
655782
1530
11:09
is to fly them to a really remote,
257
657312
2950
11:12
never-explored-before rainforest
258
660262
2276
11:14
somewhere hidden in the tropics,
259
662538
1824
11:16
and parachute down a tiny spy microphone
260
664362
3933
11:20
that would allow us to eavesdrop on the calls
261
668295
2446
11:22
of mammals, birds, amphibians,
262
670741
3322
11:26
the Yeti, the Sasquatch, Bigfoot, whatever.
263
674063
4277
11:30
That would give us biologists
264
678340
1739
11:32
a pretty good idea of what animals
265
680079
1886
11:33
might be living in those forests.
266
681965
2550
11:36
And finally, I would like to show you
267
684515
2049
11:38
the latest version of our conservation drone.
268
686564
3101
11:41
The MAJA drone has a wingspan
269
689665
2139
11:43
of about two meters.
270
691804
1674
11:45
It weighs only about two kilograms,
271
693478
1950
11:47
but it can carry half its weight.
272
695428
2461
11:49
It is a fully autonomous system.
273
697889
2595
11:52
During its mission, it can even transmit
274
700484
2387
11:54
a live video feed back to a ground station laptop,
275
702871
3963
11:58
which allows the user
276
706834
1543
12:00
to see what the drone is seeing in real time.
277
708377
2930
12:03
It carries a variety of sensors,
278
711307
2422
12:05
and the photo quality of some of these sensors
279
713729
2338
12:08
can be as high as one to two centimeters per pixel.
280
716067
4210
12:12
This drone can stay in the air for 40 to 60 minutes,
281
720277
3681
12:15
which gives it a range of up to 50 kilometers.
282
723958
3124
12:19
That is quite sufficient for most
283
727082
1915
12:20
of our conservation applications.
284
728997
2999
12:23
Now, conservation drones began as
285
731996
2070
12:26
a crazy idea from two biologists
286
734066
2191
12:28
who are just deeply passionate
about this technology.
287
736257
3285
12:31
And we believe, strongly believe,
288
739542
2150
12:33
that drones can and will be a game changer
289
741692
3347
12:37
for conservation research and applications.
290
745039
3034
12:40
We've had our fair share of skeptics and critics
291
748073
3175
12:43
who thought that we were just
fooling around with toy planes.
292
751248
3521
12:46
And in a way, they are right.
293
754769
1426
12:48
I mean, let's be honest,
294
756195
1629
12:49
drones are the ultimate toys for boys.
295
757824
3688
12:53
But at the same time, we've also gotten to know
296
761512
2406
12:55
many wonderful colleagues and collaborators
297
763918
2249
12:58
who share our vision
298
766167
1548
12:59
and see the potential of conservation drones.
299
767715
2856
13:02
To us, it is obvious that conservation biologists
300
770571
3599
13:06
and practitioners should make full use
301
774170
2258
13:08
of every available tool, including drones,
302
776428
3534
13:11
in our fight to save the last remaining forests
303
779962
2641
13:14
and wildlife of this planet.
304
782603
1787
13:16
Thank you.
305
784390
1998
13:18
(Applause)
306
786388
5722

▲Back to top

ABOUT THE SPEAKER
Lian Pin Koh - Drones ecologist
Lian Pin Koh expands conservation efforts by championing the use of low-cost autonomous aerial vehicles.

Why you should listen

Lian Pin Koh is a relentless tinkerer and science fiction movie geek, though most know him as an environmental scientist. His dreams of combining these interests led him to cofound, with colleague Serge Wich, the site ConservationDrones.org, a project dedicated to gathering intelligence on forests and wildlife through the use of low-cost unmanned flying machines.

Ground surveys are expensive, and are not conducted at a sufficient frequency. Furthermore, some remote tropical forests have never been really surveyed for biodiversity. Koh's machines have already collected valuable information in Sumatra, Congo, Gabon, and Madagascar.

He is an assistant professor of applied ecology and conservation at the ETH Zurich.

More profile about the speaker
Lian Pin Koh | Speaker | TED.com