1
00:00:01,230 --> 00:00:02,340
Hello, everybody.

2
00:00:02,340 --> 00:00:05,490
Welcome back to another code tutorial.

3
00:00:05,490 --> 00:00:08,880
This week, we're gonna learn
how to filter for clouds,

4
00:00:08,880 --> 00:00:10,770
so how to make cloud-free imagery,

5
00:00:10,770 --> 00:00:14,300
and also how to make
false color composites.

6
00:00:16,200 --> 00:00:19,350
So last week, we learned how
to make true color composites

7
00:00:19,350 --> 00:00:21,810
and how to make an ROI.

8
00:00:21,810 --> 00:00:23,493
So again, here's my ROI.

9
00:00:25,410 --> 00:00:28,290
Let's use the same ROI
that we used last week,

10
00:00:28,290 --> 00:00:31,620
and you can use the same script
that you started last week

11
00:00:31,620 --> 00:00:33,210
or you can make a new script.

12
00:00:33,210 --> 00:00:36,600
I'm just gonna continue on
in the script from last week.

13
00:00:36,600 --> 00:00:40,590
So in order to make this
a new script for Module 3,

14
00:00:40,590 --> 00:00:45,030
I'm gonna say save as
and save it as Module 3.

15
00:00:45,030 --> 00:00:47,490
So that M2 script still exists.

16
00:00:47,490 --> 00:00:50,433
I'm just creating a new one
called M3 that I'll work in.

17
00:00:52,830 --> 00:00:55,260
I also wanna do a little
bit of housekeeping

18
00:00:55,260 --> 00:00:57,900
in terms of organizing my script.

19
00:00:57,900 --> 00:01:02,790
So first off, it's
important to label your,

20
00:01:02,790 --> 00:01:05,400
the top of your script
with some information.

21
00:01:05,400 --> 00:01:10,400
So for example, I might
say created by my name,

22
00:01:10,590 --> 00:01:14,670
created and the date,

23
00:01:14,670 --> 00:01:17,160
oh, 2024,

24
00:01:17,160 --> 00:01:22,160
and then I'm also gonna do
some block organization.

25
00:01:23,280 --> 00:01:28,170
So I'm creating these blocks of comments,

26
00:01:28,170 --> 00:01:32,253
and I'm gonna label last
week's code as TRUE COLOR.

27
00:01:33,540 --> 00:01:35,220
So this is just like a way

28
00:01:35,220 --> 00:01:40,220
to visually break apart
and organize your code,

29
00:01:40,410 --> 00:01:43,590
and then for the new code this week,

30
00:01:43,590 --> 00:01:48,590
I'm gonna label that as
CLOUDS and FALSE COLOR.

31
00:01:48,690 --> 00:01:51,210
So now when I'm just, like,
quickly looking at my code,

32
00:01:51,210 --> 00:01:53,583
I can see the different
sections really easily.

33
00:01:57,900 --> 00:02:01,860
I also am going to, well,
I guess we can run the code

34
00:02:01,860 --> 00:02:04,210
from last week to remind
ourselves what we did.

35
00:02:07,050 --> 00:02:07,883
Run.

36
00:02:10,590 --> 00:02:12,993
It's taking a minute to load,

37
00:02:14,760 --> 00:02:18,610
but last week, remember we
used the Landsat 8 satellite

38
00:02:19,980 --> 00:02:22,710
and used the red, green and blue bands

39
00:02:22,710 --> 00:02:24,630
in the red, green and blue channels

40
00:02:24,630 --> 00:02:28,110
and created this true color image.

41
00:02:28,110 --> 00:02:30,600
I'm gonna start with some new code today.

42
00:02:30,600 --> 00:02:33,270
So I'm actually gonna
comment out this code

43
00:02:33,270 --> 00:02:34,650
that we did last time,

44
00:02:34,650 --> 00:02:38,100
and I'm gonna import a new
satellite to work with.

45
00:02:38,100 --> 00:02:42,030
So last week, we used Landsat
8 and the top of atmosphere.

46
00:02:42,030 --> 00:02:44,430
We learned in the lecture this
week the difference between

47
00:02:44,430 --> 00:02:47,160
top of atmosphere and surface reflectance.

48
00:02:47,160 --> 00:02:50,850
So let's grab a surface
reflectance dataset this week.

49
00:02:50,850 --> 00:02:55,230
I also wanna use a
different Landsat sensor

50
00:02:55,230 --> 00:02:57,450
to just show the differences between

51
00:02:57,450 --> 00:02:59,160
Landsat 8 and another one.

52
00:02:59,160 --> 00:03:00,350
So let's use Landsat 5.

53
00:03:01,890 --> 00:03:03,240
Again, we don't want raw scenes

54
00:03:03,240 --> 00:03:06,000
'cause those are typically lowest quality,

55
00:03:06,000 --> 00:03:08,190
and then the top of atmosphere,

56
00:03:08,190 --> 00:03:10,950
as we talked about in lecture,
can have some haze effects.

57
00:03:10,950 --> 00:03:14,250
So what we're gonna use is this Tier 1,

58
00:03:14,250 --> 00:03:17,010
and we can click on it
and see this represents

59
00:03:17,010 --> 00:03:20,670
the atmospherically corrected
surface reflectance values.

60
00:03:20,670 --> 00:03:22,380
So great. That's what we want.

61
00:03:22,380 --> 00:03:24,930
Some other important things
to look at when you're looking

62
00:03:24,930 --> 00:03:28,080
at datasets is the date
range of your dataset.

63
00:03:28,080 --> 00:03:32,580
So Landsat 5 goes from 1984 to 2012.

64
00:03:32,580 --> 00:03:33,840
That's good to know.

65
00:03:33,840 --> 00:03:37,950
We can also look at some other
qualities of this dataset,

66
00:03:37,950 --> 00:03:39,693
including the band names.

67
00:03:41,100 --> 00:03:43,410
So the band names are named
a little bit different

68
00:03:43,410 --> 00:03:46,110
in this dataset than
they were in the last.

69
00:03:46,110 --> 00:03:49,740
So because these represent
surface reflectance values,

70
00:03:49,740 --> 00:03:54,363
the band names are SR_
and then the band number.

71
00:03:55,500 --> 00:03:57,930
I also wanna point out that
if you remember last week,

72
00:03:57,930 --> 00:04:02,550
Landsat 8's red, green and
blue were B4, B3 and B2,

73
00:04:02,550 --> 00:04:05,910
but for Landsat 5, because they
don't have the coastal blue

74
00:04:05,910 --> 00:04:09,750
that in Landsat 8 is labeled B1,

75
00:04:09,750 --> 00:04:13,800
the B3 band is red, B2 is green,

76
00:04:13,800 --> 00:04:15,180
and B1 is blue.

77
00:04:15,180 --> 00:04:16,980
So that's important to note.

78
00:04:16,980 --> 00:04:21,060
Also, you might notice
that the min and max values

79
00:04:21,060 --> 00:04:25,800
are different in this satellite
than they were in Landsat 8.

80
00:04:25,800 --> 00:04:27,200
So that's important to note.

81
00:04:28,260 --> 00:04:30,330
I also wanna go to IMAGE
PROPERTIES really quick

82
00:04:30,330 --> 00:04:32,880
and show you this property
name called CLOUD_COVER

83
00:04:32,880 --> 00:04:34,980
'cause we'll be using it today.

84
00:04:34,980 --> 00:04:37,320
So the name is CLOUD_COVER,

85
00:04:37,320 --> 00:04:40,590
and then the description
is that it represents

86
00:04:40,590 --> 00:04:44,190
the percent cloud cover
in each Landsat tile

87
00:04:44,190 --> 00:04:46,323
between 0 and 100.

88
00:04:47,730 --> 00:04:51,120
Great. So I'm gonna import
this image collection.

89
00:04:51,120 --> 00:04:52,170
I'm gonna name it L5,

90
00:04:55,590 --> 00:04:58,200
and now we can work with it down here.

91
00:04:58,200 --> 00:04:59,070
So first things first,

92
00:04:59,070 --> 00:05:04,070
I'm gonna create a variable
called L5_notcloudy,

93
00:05:04,200 --> 00:05:06,990
and so within this variable,

94
00:05:06,990 --> 00:05:11,990
we're gonna summarize that
entire Landsat 5 image collection

95
00:05:12,780 --> 00:05:16,980
into just one single image
that we will visualize.

96
00:05:16,980 --> 00:05:18,780
So we're gonna start off

97
00:05:18,780 --> 00:05:21,090
with that Landsat 5 image collection.

98
00:05:21,090 --> 00:05:24,690
First, I'm gonna filter it
again to the bounds of my ROI.

99
00:05:24,690 --> 00:05:27,303
So my ROI is called FortCollins.

100
00:05:28,890 --> 00:05:31,473
I'm gonna filter it by date as well.

101
00:05:32,670 --> 00:05:35,130
This is not necessary this week,

102
00:05:35,130 --> 00:05:38,490
but I'm just doing it
because it gives the computer

103
00:05:38,490 --> 00:05:42,210
less data to sift through, and
so it'll just process faster.

104
00:05:42,210 --> 00:05:44,790
So I'm using this filterDate function,

105
00:05:44,790 --> 00:05:46,860
and within this filterDate function,

106
00:05:46,860 --> 00:05:48,510
you can put just years.

107
00:05:48,510 --> 00:05:51,480
You could put in entire date ranges,

108
00:05:51,480 --> 00:05:53,850
so including days, months and years.

109
00:05:53,850 --> 00:05:55,770
For now, I'm just gonna put the year.

110
00:05:55,770 --> 00:06:00,770
I'm gonna get data from 2000 to 2002.

111
00:06:01,020 --> 00:06:02,460
Again, this is just kind of random

112
00:06:02,460 --> 00:06:05,700
because I don't want to
work with so much data,

113
00:06:05,700 --> 00:06:09,690
and then finally, now we're
going to filter by clouds,

114
00:06:09,690 --> 00:06:14,690
and the syntax for filtering
clouds is a little clunky.

115
00:06:14,910 --> 00:06:17,850
First we're gonna tell it that
we're filtering something,

116
00:06:17,850 --> 00:06:22,850
and then we're gonna use
this Filter.lte function.

117
00:06:23,400 --> 00:06:27,720
This represents less than or equal to.

118
00:06:27,720 --> 00:06:29,490
If we wanna learn more about this, again,

119
00:06:29,490 --> 00:06:33,030
we can go to the docs and search

120
00:06:33,030 --> 00:06:34,953
for ee.Filter.lte,

121
00:06:36,390 --> 00:06:40,027
and click on it, and it says,

122
00:06:40,027 --> 00:06:43,830
"Filter on metadata less than
or equal to a given value."

123
00:06:43,830 --> 00:06:46,140
So there's two arguments
that are required,

124
00:06:46,140 --> 00:06:48,990
the name of the property
that you're filtering

125
00:06:48,990 --> 00:06:50,340
and then the value.

126
00:06:50,340 --> 00:06:53,730
So in our case,

127
00:06:53,730 --> 00:06:57,540
the name of the property is CLOUD_COVER,

128
00:06:57,540 --> 00:07:02,540
and then we can add a number
to represent the cloud cover

129
00:07:02,640 --> 00:07:04,950
that we want less than or equal to

130
00:07:04,950 --> 00:07:07,470
to be included in our
final image collection.

131
00:07:07,470 --> 00:07:08,763
So I'm gonna put 20.

132
00:07:10,290 --> 00:07:14,073
I also wanna point out
that there's ee.Filter.gte.

133
00:07:15,990 --> 00:07:19,290
So same concept, it'll
just do all the numbers

134
00:07:19,290 --> 00:07:21,540
above the number that you list.

135
00:07:21,540 --> 00:07:23,220
There's also greater than.

136
00:07:23,220 --> 00:07:24,690
So there's a lot of different options here

137
00:07:24,690 --> 00:07:27,423
that you could explore, and then again,

138
00:07:28,770 --> 00:07:32,250
if I left this variable as is right now,

139
00:07:32,250 --> 00:07:34,410
it would still be an image collection.

140
00:07:34,410 --> 00:07:37,800
So it might contain a couple of images

141
00:07:37,800 --> 00:07:41,910
that are within my filter
bounds, within my filter dates,

142
00:07:41,910 --> 00:07:44,190
and that have cloud cover
that's less than 20.

143
00:07:44,190 --> 00:07:47,310
So I need to summarize
that image collection

144
00:07:47,310 --> 00:07:49,710
into just one image somehow.

145
00:07:49,710 --> 00:07:52,320
So I'm gonna put median.

146
00:07:52,320 --> 00:07:55,053
So I'll take a median
of all of those images.

147
00:07:56,910 --> 00:07:59,220
Great, so now we can add

148
00:07:59,220 --> 00:08:04,113
that notcloudy Landsat 5 image.

149
00:08:05,190 --> 00:08:07,020
We can add it to our map.

150
00:08:07,020 --> 00:08:10,170
So I'm gonna first put the name

151
00:08:10,170 --> 00:08:13,410
of my image variable.

152
00:08:13,410 --> 00:08:16,263
Then I'm gonna add some
visualization parameters.

153
00:08:17,100 --> 00:08:21,840
So the important one to
put here is the band names.

154
00:08:21,840 --> 00:08:26,070
So SR_B3 is red.

155
00:08:26,070 --> 00:08:29,853
SR_B2 is green.

156
00:08:31,900 --> 00:08:36,450
SR_B1 is blue, and then this is a list,

157
00:08:36,450 --> 00:08:38,163
so I have to put it in brackets.

158
00:08:39,510 --> 00:08:42,060
Let's also put the min and max values.

159
00:08:42,060 --> 00:08:44,820
So if you remember, min they said was 1.

160
00:08:44,820 --> 00:08:48,060
Max was around 60000.

161
00:08:48,060 --> 00:08:50,100
We can edit that in a second,

162
00:08:50,100 --> 00:08:53,733
and then let's add a name to
our map, so I'm gonna say L5,

163
00:08:54,960 --> 00:08:58,563
and then this is a true color image.

164
00:08:59,940 --> 00:09:01,533
So let's run that.

165
00:09:03,750 --> 00:09:05,370
Oh, okay.

166
00:09:05,370 --> 00:09:10,170
So what happened here is
I forgot to clip my image

167
00:09:10,170 --> 00:09:12,630
to exactly to my ROI.

168
00:09:12,630 --> 00:09:17,630
So what it's giving me here
is the entire Landsat tile,

169
00:09:18,210 --> 00:09:20,310
and actually it's given
me two Landsat tiles

170
00:09:20,310 --> 00:09:22,293
'cause it seems like my ROI,

171
00:09:23,430 --> 00:09:27,330
there's two tiles that
both sit on top of my ROI.

172
00:09:27,330 --> 00:09:31,407
So after this variable, I need
to put .clip(FortCollins),

173
00:09:32,790 --> 00:09:37,173
and then it'll clip it
just to my little ROI.

174
00:09:38,310 --> 00:09:40,650
Okay, so one of the first
things I'm seeing here

175
00:09:40,650 --> 00:09:42,693
is that this image is pretty dark.

176
00:09:43,560 --> 00:09:46,320
So I'm gonna go to the Inspector tool

177
00:09:46,320 --> 00:09:47,943
and click around a little bit.

178
00:09:48,870 --> 00:09:51,033
I can see the values here.

179
00:09:52,290 --> 00:09:57,290
So you can click on some
low reflectance value pixels

180
00:09:58,050 --> 00:10:00,960
and see what those are.

181
00:10:00,960 --> 00:10:05,370
You can click on some high
reflectance value pixels,

182
00:10:05,370 --> 00:10:08,580
and this is just giving me
an idea of what the, like,

183
00:10:08,580 --> 00:10:11,193
max, min and max might be.

184
00:10:12,120 --> 00:10:15,570
I could also manually look in here

185
00:10:15,570 --> 00:10:20,460
and say stretch the image by 100%,

186
00:10:20,460 --> 00:10:25,460
and it'll give me the range of my image.

187
00:10:25,680 --> 00:10:28,563
Okay, so that's good to know.

188
00:10:29,850 --> 00:10:32,190
I'm gonna change my min to, let's say,

189
00:10:32,190 --> 00:10:37,190
5000 and my max to 20000,
and I'm gonna run it again,

190
00:10:38,130 --> 00:10:40,050
and yep, this is a bit clearer.

191
00:10:40,050 --> 00:10:42,600
I can see more in the
landscape now, which is good.

192
00:10:44,640 --> 00:10:48,480
Okay, so again, we
filtered for cloud cover

193
00:10:48,480 --> 00:10:51,420
to get less than 20% clouds.

194
00:10:51,420 --> 00:10:54,130
What if we did less than

195
00:10:55,050 --> 00:10:58,503
1% clouds?

196
00:10:59,340 --> 00:11:00,690
So it kinda looks the same.

197
00:11:01,590 --> 00:11:04,380
Colorado's not the cloudiest place ever,

198
00:11:04,380 --> 00:11:06,967
but just for fun, we could say,

199
00:11:06,967 --> 00:11:11,967
"What if I want an image that
has greater than 30% clouds?"

200
00:11:13,050 --> 00:11:15,390
I'm gonna run that, and let's see.

201
00:11:15,390 --> 00:11:18,420
Yeah, so just to show
you that the code works,

202
00:11:18,420 --> 00:11:22,050
if I put greater than 30% cloud cover,

203
00:11:22,050 --> 00:11:25,140
it'll give me a pretty cloudy photo.

204
00:11:25,140 --> 00:11:29,583
So let's go back less than
or equal to 20% cloud cover,

205
00:11:32,580 --> 00:11:37,470
and now I want to visualize

206
00:11:37,470 --> 00:11:40,980
this image but with a
false color composite,

207
00:11:40,980 --> 00:11:42,960
not a true color composite.

208
00:11:42,960 --> 00:11:47,340
So I'm just gonna copy-paste this code,

209
00:11:47,340 --> 00:11:49,080
and then I'm gonna edit it down here.

210
00:11:49,080 --> 00:11:51,780
So this is my false color.

211
00:11:51,780 --> 00:11:54,273
This was my true color.

212
00:11:55,380 --> 00:11:57,960
So in order to make this false color,

213
00:11:57,960 --> 00:12:00,150
I need to change the band names.

214
00:12:00,150 --> 00:12:04,509
So I'm gonna make this
an infrared false color,

215
00:12:04,509 --> 00:12:07,260
infrared,

216
00:12:07,260 --> 00:12:10,350
and to do that,

217
00:12:10,350 --> 00:12:14,103
I need to assign the red
channel to the infrared,

218
00:12:16,050 --> 00:12:19,020
the green channel to the red band,

219
00:12:19,020 --> 00:12:21,303
and the blue channel to the green band.

220
00:12:22,320 --> 00:12:25,710
I'm gonna make sure I label my map

221
00:12:25,710 --> 00:12:30,093
so I know what it's
visualizing, and now I can run,

222
00:12:35,430 --> 00:12:36,960
and the image looks quite different.

223
00:12:36,960 --> 00:12:41,830
So the areas that were
green before are red

224
00:12:45,400 --> 00:12:49,830
in this image, and I'll let
you think about why that is.

225
00:12:49,830 --> 00:12:52,230
So in the assignment for
this week for Yellowdig,

226
00:12:52,230 --> 00:12:55,680
I asked you to create
a infrared false color

227
00:12:55,680 --> 00:13:00,360
and a short wave false
color image within your ROI.

228
00:13:00,360 --> 00:13:04,470
So what you'll do next is take this code

229
00:13:04,470 --> 00:13:06,120
and then change the bands again

230
00:13:06,120 --> 00:13:10,173
to represent a short
wave false color image.

231
00:13:12,090 --> 00:13:14,163
So don't forget to save your code,

232
00:13:18,570 --> 00:13:21,180
and then again, when
you're sharing your code,

233
00:13:21,180 --> 00:13:25,563
click Get Link and then
copy-paste what's here.

234
00:13:28,680 --> 00:13:30,903
All right, thanks, and see you next week.