Landsat NDVI Trend Analysis Tutorial: Linear Regression + Mann–Kendall Test in Google Earth Engine

So hello everyone welcome

welcome to the

14 days online training program on

okay can you see my screen is it visible

can you see my

Yes.

>> Okay. Thank you. So let's click

continue. Yeah.

So basically this is a course title

research based training on mastering GIS

and remote sensing analysis with Google

Earth Engine. So basically in this

online training program I will try to

cover lot of research topic. Okay. as

well as uh if you want you can also

publish some scientific paper regarding

this type of topic then it will be more

better okay so we can easily complete

all of those okay so basically we take

the class only for the one day in a week

only for the Monday we take the class

okay for 3 hours okay is uh day we

provide the 3 hours session and we try

to cover all of research based topic

Okay. And every topic we can publish the

scientific paper.

Okay. And every Monday we try to join at

the same time. Always every initial we

try to join at the same time. And

further we simply discuss our topic.

So first of all I try to discuss about

that uh what types of content we try to

cover in this syllabus.

So first of all we discuss about that

rooley model

how we can use and then how to compute

the soil erosion or soil loss mapping

for a hilly agriculture area using rley

model. Then we'll discuss about the

change detection mainly NDVI change

detection for the forest degradation

monitoring.

uh in this case we can detect the forest

health changes over 20 years or 30 years

in a national park using landset data

set. Then we talk about the rainfall

pattern. Okay, rainfall pattern or

monthly rainfall pattern, seasonal

rainfall pattern or annual rainfall

pattern and analyze the rainfall

variability in broad prawn region using

the charts satellite data or any other

satellite data which is provided with

the precipitation. you can also use no

problem as well as discuss about the

land use land cover classification using

machine learning. So basically this is

our topic. Um we can also use the Google

engine but now we also discuss about the

some advanc for Python library how we

can use for the Python to make the

supervised or unsupervised

classification further how to make the

urban extension or urban spiral. So this

type of thing we discuss about that and

discuss about the also LST NDVI time

series analysis for urban heat island

okay and compare the vegetation health

and temperature trend for the urban

versus ruler area we also discussed and

satellite based aridity index for

drought assessment we also discuss about

that map severity for agricultural

region using modis data set okay and

also discuss about the LST monitoring

using landset data for generate the

seasonal LST map summer season rainy

season different types of seasonal we

also discuss about that and linear

regression and trend analysis for

climate variable so different types of

climate variable bioclimatic different

types of variable we also discuss then

main candle uh end trend analysis we

also discuss detect the long-term

greening and the growing the trend for

the semi region

And also discuss about that different

types of drought index suppose

vegetation condition index, temperature

condition index or vegetation health

index to monitor the drought scenario

for our study area and deforestation

dete uh detection using machine

learning. We also discuss about that

mainly identify illegal logging areas.

Um suppose some forest area or some

place illog areas. We also discuss about

that. Then slope classification area

calculation NDVA classification raster

restriction. Then monitoring the

bioclimatic variable

then transpiration. Then soil serenity

how to monitor we also discuss about

that. Then landslide substability

mapping. Okay. Flood mapping sent on one

side also discuss about that okay NDI

trend analysis for the c of health

analysis also discuss land suitability

mapping livestock suitability mapping

then future land use land cover

prediction using CMR model and land and

buildup okay and flood bricks mapping so

it's mainly cover lot of topic okay and

every topic we can easily publish the

scientific paper if you want or you can

also combine suppose more two or three

topic and make a paper easily. So in

this session uh basically this training

program I will try to guide you

methodology section method section how

we can apply this type of methodology

and show you the practical analysis

practical analysis. So further uh you

can read different types of paper

regarding your topic and then literature

review or also others you can do. So for

that I also guide you. Okay. So how we

can easily work for that? We also

discuss about that. No problem.

So this is the overall summary of our

total foods course syllabus and every

concept or every topic we our main focus

for research research based and in the

research based session I try to show you

that methodology section. Okay. Method

and also result this section I try to

discuss every topic I try to discuss. So

further topic mainly introduction then

literature review or also others data

method. So different part of the

research paper you also try to do this

type of things. Okay. But I just cover

methodological section how to perform

this type of analysis how to do this

type of analysis. So further you can uh

easily learn the method and this method

you can easily use it for um your

research paper. Okay. So this is the

full course content and um you can get

the course e certificate after

submitting your assignment and what is

the assignment. So our assignment is

that we try to publish minimum four

paper or three paper minimum three or

four paper after that you are able to

get a certificate from this and you try

to publish the best journal Q1 journal

or Q2 journal also no problem. So you

try to publish it. Okay. After that you

are able to get a certificate. Okay. So

it's very important all of those because

this top it's a research based training

program. So already one base already

done. So this is our second base. Okay.

So one B already tried to publish some

paper. We already submit some journal.

So they also send our feedback. We try

to I hope it will be published some

paper from the first B also. So this is

our second base program. Okay. Second B

is running. So in the second P we try to

cover all of topic I already discuss

okay and I hope you can also get the

idea or anyone any question for the

course content you can ask me or any

confusion you can also ask me then I'll

try to uh recapture again any question

or any confusion about the course

content you can ask me.

Yes, please please please

>> uh sorry I was late I was kind of busy

but uh regarding the course uh so and

the subjects that you have already

explained uh the topics uh so are you

going to go with the in the B with us

with the basic of uh coding or are the

codes are going to be ready? You're just

going to explain it for us.

>> Okay. So basically this session um we

are start from that advanced level some

advanced level. Okay. For the basic

coding we have the another session. It's

called the 7 days training session. For

the basic if if you have no knowledge

about that uh coding then we have the

another session. It's called the 7 days

training session. But this one is a

research based training session.

>> I see. So uh do we need to sign up to

that session also or is going to be uh

the same session but in a different

time?

>> So we need to also registration for the

another

okay for the another program. So this is

another program

>> for that basically if you show that I

will show you this. This is another

program. So this is this one. So

basically this is a Google Earth engine.

So this new base will be start from that

17th January. Okay. So it's a totally

beginners who don't have ABC knowledge

of uh coding. Okay. So in this program

we try to discuss about that all of uh

things from beginners to advanced level.

Suppose how to run the code, how to

write the code, everything will be

completed and as a beginners who don't

have any knowledge, they can easily

learn these type of things. But now uh

this are this is our researchbased

training session. So in the research

training session we simply focus on the

application

application of different types of

project on Google.

>> Okay. Thank you very much.

>> Okay. So let's start. Yeah.

Okay.

Okay. So let's start our first class and

every class you can get the regret

session as well as you can also get that

uh practice code all of materials. So,

so today discuss about this topic mainly

20 yards data how I can get suppose it's

very important. So this data set you

need when you perform any types of

analysis suppose you want to make the

land use land cover or NDVI or anything

suppose I need the data uh for 20 years

data or 30 years data of landset

imagery. So how we can get this type of

data then we can easily apply different

types of analysis no problem. So first

of all we discuss about that how we can

easily get the data suppose for 30 years

or 20 years. Okay we discuss about that.

So I think here all of you already

complete the registration with Google

Earth Engine because it's our u advanced

training session. So you already know

about the Google Earth Engine how it

will work. So first of all we discuss

about that how to get the data. Suppose

how to get the 20 years data or 35 years

data for landset. And if you check

suppose click on here and click the

browse data catalog we can find out the

lenset section

and if you check the landset datas are

available from uh you can see different

time period. So you can see landset 1 to

5 it's available between this time

period landset 4 this time period. lands

at five this time period and also lands

at 7 this time period then lands at 8

also this time period so lands at 9 also

this time period so now how I can

merge this type of lens set okay and get

the data suppose I need the data for 20

years or 30 years so how I can get this

type of data and when I have this type

of data then I can easily perform

different types of analysis I can easily

perform the historical land use land

cover. I can easily perform the

historical uh land surface temperature.

I can easily perform the analysis

historical um NDVI different types of

things I can do. So it's very important.

So that's why I start from the data

collection how I can get the data and if

you have the data so this data further I

can use the different purpose I can

easily use for the land use land cover.

I can easily use for the land surface

temperature. I can easily use for NDVI

monitoring or visitation monitoring. So

different purpose I can easily use this

type of data. So let's go ahead to

start.

So first of all we discuss about that uh

data collection

data collection. Okay. How we can easily

get the data mainly landset

landset.

Suppose I need the data from

uh 1980 to

20

25. Okay. Suppose every year's data I

need. So this is my main target data

collection. So I take the long-term data

or historical data. So this type of data

I need from 1980 to 2025.

So now how I can set this type of data.

So it's very important because uh when

we have this type of data we can easily

perform the different types of analysis.

I already know um land land use land

cover, NDVI, LST, NWI. So different

types of analysis you can easily do for

historical data analysis. And if you

check the different types of paper so

always they perform the historical data

analysis. So here I already open lot of

paper. So you can see change detection

visitation using remote sensing and GIS.

So if you check this paper also use the

data from 1990 to 2010. Okay. So it's

historical data they take from the NDVI.

Okay. And if you check every paper. So

basically today I discuss about this

topic. So that's why I already open

different types of paper related this

type of topic mainly vegetation change

analysis or vegetation trend analysis.

Uh these things I try to explain today's

session. So here I already opened this

type of relevant paper and I will send

you all of paper because our main target

is to publish the scientific paper. So

this type of relevant paper I also send

you uh then you try to read later. Okay

every paper. So today discuss about this

type of paper method how we can perform

the analysis.

So let's go try to data collection.

Suppose I need the data from landset

1980 to 2025.

So first of all I need to create a

observation year. Observation

year. Suppose in the observation year I

need to set a start year and in year. So

simply take a variable. So I think you

already know about the variable. Suppose

I put the variable name is start year

and I put suppose year 1990

or 1980 any year you can put then

suppose I put the end year end year and

put the year

suppose 2025.

So this is my observation year. Okay you

can also increasing or decreasing this

time no problem. So this is my stud

year. This is my end year. So I need the

data for every year. So just I make a

list. I just make a sequence sequence

suppose 1990

19001 19002 19003 19004 and up to 2025.

So all of data. So take a variable

suppose years equal and make a list

sequence. For that we use e dot list dot

sequence. So it this this function may

create a list or sequence. So it start

from 1990

and up to 2025.

Just I copy this variable and paste

here. So now what happened if you check

just I print the years and run the code

it just make a list it list computing

total 36 years. Total 36 years it

printed. If you check all of years 1992

all of years you find out in here. So

now what is my requirement? My

requirement is that I need this every

year satellite image. I need every year

satellite image from 9925.

Every year satellite image I need. So

for that

for that if you check the satellite

image this satellite image all of

satellite image is not the it start from

you can see 2021 to present so it's

available 2013 to present it's available

from 1980 uh 2024 it's available from

1984 to 2012 so now I need to merge okay

I need to merge multiple collection

suppose I need to merge land set four 5

7 8 and 9. So this four or five image

collection lands at four lands at five

lens at 7 L at 4 8 and L at 9. So this

five image collection for different

different year I just need to merge I

need to I need to merge when I take the

data from 990 to 2025. So now we discuss

about the how I can merge. I also write

the comment also because I also send you

the code link then you also try to read

this type of comment. It also help you

for your study.

So how to merge

L set 9 L set 8 lens 7 L set five and L

set four. Okay. So how to merge this

type of collection. Okay. If you check

uh it's a different different collection

lens set at four lens at set five lens

at 7 lens at 8 lens at 9. So it's a

different different collection. So now I

want to merge all of collection

together. So how I can merge? So it's

very important because if you check the

band combination mainly bands or

spectral bands suppose let me open for

lancet 9.

Lancet 9 if you check we find out the

surface reflectance. So I want to work

with surface reflectance this data and

if you check this data the serial of the

band. So there are total seven spectral

bands B1 to B7 and B1 is the B1 is the

blue or coastal aerosol band or alter

blue. B2 is the blue. B3 is the green.

B4 is the red. B5 is the near infrared.

B6 is the short wave infrared one. And

B7 is the short wave infred 2. Okay.

This is the sequence of band for lens

set 9 and we also find out the same

sequence also for lens set 8 satellite

data. If you check the lens set 8 again

that was lens 8. If you check

this B1 ultra blue then blue green red

infrared short infrared and short 2. So

B1 to B7 also same sequence we also find

out for the lens set 8 and Lancet 9. So

the the lens set 8 and lancet 9 image

collection are the similar

band sequence we find out. Okay. But if

you check for the lancet 7 it's some

different. If you check the lens 7

surface reflectance

this bands if you check it show the

result B1 is the blue it's a blue

but lancet 8 and 9 B1 is postal aerosol

or ultra blue but for the lens 7 if you

check the lens 7 image collection this

band it's a blue then green then red

then near infrared then short wave

infrared and short wave infrared 2.

Okay. So in this time this sequence is

not similar with lancet

8 and lancet 9. This is a different

sequence. Lancet 8 and 9 similar but

lancet 7 is the different you find out

the difference because you can see

lancet 7 one band is short basic.

Basically lit 7 it is not provide you

with the coastal aerosol. Okay. So

that's why SRB1 is the blue band. So B1

is the blue for Lancet 7. For Lancet 8

and 9, B2 is the blue. Okay, we can get

the some difference between the lids 8

and 9 and Lancet 7. So you can check the

uh it's a sequence and we find out the

similar sequence for L set five. If you

check the lens set five

set five, if you check lens set five and

seven are the similar sequence. If you

check

so for the lens uh five, it's a lens at

five image collection. If you check it

from the B1 many blue then you can see

B2, B3, B4, B5, B7. So we find out the

similar sequence for the lancet 5 and

lancet 7 they are spectral bands are the

seven. So from here you can get a

decision. So lancet 9 and lancet 8. So

these two image collection are the

similar sequence of bands but lancet 7

and lancet 5 also similar. Lens set five

and l seven is similar. Lenset 8 and

lens 9 is also similar.

Okay. So if I just make a note about

that. Suppose we find out that lens set

8,

Lancet 9 and Lancet 8 equal similar. We

find out the similar bands or similar

spectral band sequence and also find out

that Lancet 7

and Lancet 5

equal to similar

similar bands sequence. Okay, it's very

important when you merge. So lens at 9

and lens at 8, it is a similar bands.

Lens at seven and lens at five, it's a

similar bands. But lens at 7 and lens at

9, it's a different, it's a different

lens at 5 and lens at 8, it also

different. So we find out that lens at 9

and lens at 8 is a similar bands. But

lenses 7 and five also similar. Okay, we

find out this type of things from here.

So first of all I need to create the

similar bands. So mainly when you merge

when you merge landset 5 7 8 9 make the

all of all of uh all of collection all

of collection

need similar

sequence. Okay. So when you merge all

collection suppose land set 7 8 9 or

five you need the all collection need

the similar sequence it's very important

otherwise it not create the composite we

need the similar sequence so now how I

can make the similar sequence so let's

discuss about that so I hope you also

get the idea from uh for this lecture uh

how I say about these things you can

also get the idea okay or any question

or any confusion you can always ask me

the question I try to explain again. If

you have any confusion or any question

you can ask me anytime. Okay. So I'll

try to give the answer inshallah you can

get the solution. So first of all I take

a variable

I put the variable name is.

So here I find out I simply put the

variable name. Suppose I put the

variable name is

L L5

bands or I put suppose L5

bands. You can put any name as a

variable and make a list and make a list

all of bands. Suppose I put the list

SR B1,

SR

B2,

SR

P3,

SR,

SR

P4,

then SR

P5,

And there is no SRB 6, SRB 7. We write

SRB 7. Okay. So I simply write SRB 1,

SRB 2, SRB 3, SRB 4, SRB 5, SRB 7. Just

make a list for Lancet five band names.

So I take another variable. How about I

put the variable name is that

L5

names. I just write the name all of band

name. So I put the name suppose SRB1 is

the blue band right SRB2 is a green band

for lenset 5. I put the name then I put

the name SRB3

is the red band. SRB 4 is the near

infrared band. Then SRB 5 is uh S WR1

and then SRB

7 is S WR2 short wave infra two. So just

I make the band and band name band and

band name for lens set five also keep do

the same things for lens set 7. So

Lancet 7

and Lancet 5 is similar. We already

check it. So just I copy the same line

of code

and then

paste here. I simply change the name.

It's a L 7. It's a L set 7. Put the L7

L7 because band sequence are the similar

for the L set 5 and L 7 that is no

problem. So now we write the code for L

set 8 and 9. So take a variable same

way. So I put the variable name suppose

L8

bands

and I put the all of Lancet 8 bands

name.

So Lancet 8 band names. So I start from

SRB2. SR BB2. Why I start SRB2? Because

SRB2

is blue band for landset A. SRB2 is the

blue band. So for that I simply put the

SRB2. I start from SRB2 because SRB2 is

the blue band. Okay. Then put here

SR B3.

Then put here SR B4.

Then SRB 5.

SRB5,

SR

B6

and SR

B 7. Then I also put the band name. So

take a variable

L8

names. So now I simply write the name

SRB2 is blue band. Okay. SRB 3 is the

green band. SRB 4 is red band. Then SRB

5 is near infrared band. SRB 6 is short

wave infrared one band. SRB 7 is short

wave infrared two band. Okay. And Lancet

8 and Lancet 9 is the similar. So that's

why I simply copy the same line of code

and paste. And I simply change the name

it's L set 9 band L set 9 band name. Now

if you check the last line every band

last line. So this one also this one

also this one

and finally this one is the similar.

Okay you can see the similar. So it's

very important because when you make the

merge all of collection you need the

similar sequence of band similar

sequence of band we need when you merge

land set 5 7 8 9 when you merge you need

the similar sequence all time okay all

time we need it and this it's important

because uh when you just develop this

code for first time then other time not

need just we simply change our study we

can easily get the result okay so first

time that I simply explain all of those

line by line But next time when I work I

simply use the same code. I simply use

the same code all time. Not need to

change just I simply change my study

area and get the result from here. Okay.

So now if you check we find out that

this is the similar sequence. We make

the similar sequence. Let five lands at

7 L set uh 8 then land set 9 have the

similar sequence. There is no problem.

It's a similar sequence.

So now we simply call the collection our

landset collection.

Take a variable suppose L9

and if you check the image collection

suppose L9 just call

just surface reflectance just click.

So this is the ciphet ID of this image

collection just I copy

and then paste here. Okay just I call

the lenset 9 image collection. So this

lenset 9 image collection I need to

filter. So how I can filter? So for that

I need to import a stud area. Okay. So I

where I need the image suppose where

boundary or where I need the image. So

for that here I already import a shape

file in my asset tab. So for if you

simply okay here I already upload a

shape file. So I want to use it to

display this image. Just I click on

here. So it's imported in in here as a

table. So I simply change and I put the

name suppose ROI. So I need the image

suppose 1982 2025 for this study area

for this boundary I need this image. So

for that just I apply this uh function

in here. So it call the dot filter

filter bounce function. it intersecting

with my ROI then dot

filter date function put for the

specific time period. So now if you

check this landset 9 image is available

between this time period. So 2021 so

this time period. So I use this time

period full time period. Okay it give me

the all of image between this time

period. So just I use this same time

period.

Okay.

just we try to paste here. So start date

it will be start from 2021

uh 10 31 and put this time 26 uh 1st

January uh 28. So I put this type of

time period

and put the another filter metadata. I

set the cloud cover

cloud cover

less than I set less than

less than one. Choose a low cloud cover

less than one. And after that simply

just I use the similar code all time

just copy.

So now I simply call the lancet 8

collection just paste here. I put the L8

and [snorts] change the collection. So

how I can change it? Same way just uh

open the lens set 8 section eight

collection. If you check uh this is the

lens set 8 collection

just I click on here. So this landset 8

surface reflectance I just take and copy

this

just copy

and change it.

So this landset 8 we take and filter

bound. So simply change the time period

date is available from 2013

of it's uh March

18 months from we can take this result

and if you check this time period

it's available from uh 2026 uh January

3rd January I put the 3rd January no

problem okay so this landset 8 done so

then take the another variable for land

set 7. So now I use the same line of

code just I copy

and paste. So now simply call the lancet

7. So just open the lancet 7 collection

again.

If you check the lens set 7 this one.

So just I call the surface reflectance

and copy this ciphet ID of lancet 7

and change it

and it's available if you check it's

available from

uh this time

just I copy this time period

and change it

and also

with the last time period

I change it. Okay. So nice. So now L is

five similar just I copy

and paste and I put it's a lens set

five.

Okay. So now simply call the lenset 5

collection

the landset. This is the lens set

just back

lens set five. This one

and then copy the surface reflectance.

This one just open.

Then copy the cipet ID of lens set five

and paste here. Okay. Okay. So, let's

set five and it's available from if you

check uh 98416

just copy the start date

change here

and also call the

2005 Z just take OP

here.

Okay. So this is the common things we

simply apply for the lens set 7, Lancet

5, Lancet 8 and L set 9. So now this

type of bands I need to connect with my

collection. So how I can connect? So for

connecting just use the select function

select and call the lens set

9 bands. Just I copy.

Okay. Then put the comma sign and call

the lancet 9 name because we put the

similar name. So this thing we simply

copy and paste one by one all of code.

So now lancet 8 just paste here. So

simply change it's a lens set 8 bands

and lancet 8 name. So then simply lset

7. So L set 7

bands and L set name. Same L set five.

Just I connect the band this I already

created L set five bands and L set uh

five name okay so it's done

so you can get this type of things okay

I hope you also get the idea how it will

work yeah so now uh I want to merge all

of collection all of lands set

collection I simply merge

so for that take a variable suppose

full collection. I put this type of name

full collection and I simply merge all

of collection. I simply merge L set five

L set 9 merge with L set 8 merge.

Okay, merge function simply merge L set

at 7

then merge L set five. Okay, so I simply

merge L at 9 8 7 and five. I want to

check how many collection I find out.

How many collection I find out total

total number of collection after March

total number of collection I want to

check it. So for that just use the print

function

and call the full collection

and use the size function. So let's

computing how many total collection.

I think some wrong spelling I put the

lens set. Okay, let me check. Okay, I

put the name. So, it's a name. I put the

name. Similar name I put name. So now

run.

My god, it show the problem. Lame is so

I put I think I use the different

spelling. So L5 I put the names. It's a

name. I put the name. No problem. So I

put the similar. So this variable and

also use the same variable also in here

otherwise it create some problem. So now

it's very good. We find out that total

373

images. Okay. Total 373

images in my collection after Mars lands

at 9 8 7 and five. Total 373 image

collection. And this 373

image collection is less than 1% cloud

cover. Less than 1% cloud cover. You

find out the total how many image? So

you can I put the all time less than 1%

less than 1% less than 1%

less than 1%. So I take only for less

than 1% cloud cover and I find out how

many image we find out that total

uh 373

image collection I find out.

Okay, this is my image collection I find

out.

So now this image collection I want to

convert into annual. Suppose I have

[clears throat] total 300 373

images. So this image I need to convert

into 36

me.

Okay the 36 images. So and it will be it

will be uh annual it's many 36

annual images.

Okay. So I have total 373 images. I

already make the list. Total 36 years.

Total 36 I already make the list. And

now I want to create a function. Suppose

9990. How many image I have? I don't

know.

1991 how many image I have I don't know

how many image I have 992 I don't know

so now I create a function this function

filter this function filter how many

images are available for every year for

every year how many images are available

so I create a function this function

help me compute how many images are

available for every year I find out

Total 373 images but 1990

how many images are available in this

one year 1991 how many images are

available in one year I don't know so

for that I need to filter I need to

filter this image collection and find

out that how many images are available

in this year how many images are

available in this year how many images

are available in this year and up to

2025 how many images are available in

this

Okay. So for that um we use a function.

Suppose I have image collection.

Now I want to convert the another image

collection. So I have total 373

images. So now this 373 images I just

convert 36 images by annual by annual

every year

all of images. Suppose 1990 all of

images 1992 all of images. So year by

year I just make the composite or median

composite all of images.

So for that we simply take a variable

suppose I put the variable name is uh

annual

collection I put this type of name

annual collection. So now I have a

collection this collection I to convert

another image collection. So then we

simply use a function. It's called the

e dot

image collection

from

images.

So this function is simply apply a dot

image collection from images

function. So I have image collection. So

I want to create the 36 images from this

image collection. So I want to create

this 36 image collection from this image

collection. So this function e.m image

collection from images. So I want to

create 36 images by annually

from 30 373 images. So just we use this

function e. Collection from images.

Okay.

So in this function uh we simply uh

write our argument. So first of all I

create a function. I create a function.

So let's go create a function. So I

create a function.

I put the function

and I put the element name suppose y and

then put the ki bracket in this kal I

put my argument. So what is my argument?

I want to filter. I want to filter this

collection. I want to filter this images

mainly this collection total 373 images

by year by year. So for that take a

variable suppose annual

and call the full collection because it

is stored total 373 images full

collection and this collection I want to

filter. How I can filter? Art engine dot

filter dot calendar

range.

Now I put the time here

year to year mainly one year to another

year year to year and unit is year. So

now what happened? So now what happen if

you check

I have a full collection total 373

images. So in this full collection it

just filter how it filter year by year

suppose 9990

1992 all of year by year it's just

filter and every filter image suppose uh

990 I have suppose three image it create

the composite so I put the median median

composite it create the median composite

as well as also clip with my shape file

which is R O I okay

so now what I return from this function

so I want to return

this function I want to return

annual [snorts]

annual it's set with

here I put the element name is y so now

it's very complex as a beginners if you

have no knowledge about the coding so

That's why we have the another session

how to learn all of part of the code.

Okay. So I hear all of uh some student

already uh doing my previous course for

7 days. So I hope they also get this

type of code easily. So in this code

what I write in here if you check I

create a function. Okay I create a

function. So in this function

I want to filter the image year by year.

Okay. And I already make a list all of

years. I already make a list all of

years. So if you check previous first

time I already write the list of the

years. So just I copy this collection of

year because you can see this variable

is stored total 25 or 36 images 36

years. So just I call the years and this

years I connect with my function just I

simply connect and how I can connect

using the map function. I connect start

the bracket and also close the bracket.

So now what happened? So now this years

it stored all of year 1990

1991 all of year is stored in this.

Okay. So now from this collection one by

one year just come one by one year just

input in my function and filter. Okay.

and filter. Suppose the first time it

just come 1991 or 9990. So this one this

year suppose 1990 this year just come in

my function and filter 9990. Suppose it

find out total

three image. Okay. In this year suppose

three image I have. So this three image

just create the median composite. Same

1991 this year from this from this list

just come in function and create the

median and up to all of year 2020 all of

year one by one just [clears throat]

come in this tool or just come in this

function and create the median

composite.

So now if I check suppose just a print

this annual collection. So now let's

check how many image I find out. Just

check it. Just use the size function and

run the code. So now it's computing how

many image I find out 36. It created the

total 36 images because this years total

36 years. I already created the list of

total 36 years. So now this full

collection basically this number of

image collection it filter into total 36

years annually annually filter and all

of this year create the median

composite.

If you check all of image information

just

run the code it create the image

collection

36 element 36 images.

Then if you check

if you check we find out that okay total

six bands every image we find out there

are six bands. So just click on here. If

you check it show the image 9990.

Okay first image 9990. Then if you check

suppose second image 1992

and all of image if you check last image

225 and all of image have the total six

band. Six band blue green red short we

can easily get it there is no problem

okay we can get that all of image total

36 years image now I have

so I want to display this type of image

in my map let's try to show you that how

we can display

for that uh just if you want just map

dot add layer and call the annual

collection ction

and just run the code then see what

happened. There is no problem. It's just

loading but not focus on my given study

area. So further simply apply map dot

center

object object is

uh ROI zoom level I put the 10. So I

directly focus on my given study area.

Okay, which I put

and let's see what happened.

Or if you want just click on here.

And if you check

you can see it show the all of bands but

I want to display the image every year

year by year. Suppose I need to display

the image uh 1991 1992 this type of

things. So for that simply apply the two

bands function. So now run the code.

So now total 36 bands it's added. If you

check we find out. So just I open

and click on here. If you check it start

from the zero and if you check

35 or 36 three bands I want to create

suppose you can see zero mainly 90.

Okay. First serial if you check the

serial number of the list zero zero only

9990. And if you check 35 is the 2025.

Suppose I want to set the band

combination for 2025.

So how I can set just I select the 35.

If you check the 35 where is it? 35 35

30 35 35 red

then I put the 35

uh red then green

and 35

35 35 blue RGB natural color 35 blue I

keep the stressing keep the 98% level

but you can see it show the digital

number okay after some time discuss

about that how we can easily convert

into reflectance value. It show the

digital number. You can see it show the

8,000 13,220.

It's a digital number. But this digital

number we can convert in reflectance

value. Okay. Uh every bands have the

reflectance value. So we can also

convert. So I'll show you that after

some time just a minute. So just apply.

So then if you check uh the result you

find out this type of imagery and this

image for 2025 because you can see I put

the 35 35 35 red 35 green 35 blue. So

you can easily get the 2025 images.

It's just for visualization. It's not

the proper way to add the image. So I'll

show you the proper way also. No

problem. Suppose I want to display the

image for uh image for 9990. Okay. So

for the 9990 if you check it's a zero

serial zero zero zero. I simply set the

zero.

Suppose zero

uh red

then zero

uh [snorts] green

and zero

blue. Okay, just use it and it's going

to give 98%. So it still show the

digital number. So after some time I

will show you that how we can confirm

reflectance value. Just I apply.

So now you can get this type of imagery

for 990.

Okay.

Okay. So this is the one way to

visualize the image. But we face a

problem. Uh it show the digital number.

Okay. But if you check if you check the

satellite images basically it show the

reflectance value this type of a unit

0.45 45 0.52 0.63 63 0.77

but for the digital number it show this

type of min and max value. So now how I

can convert digital number it's called

the DN value. So this DN value I can how

I can convert into reflectance value.

So for that they already give you a

scale. So this is the scale. This is the

scale and this is the offset value. This

is the scale value and this is the

offset value. So this scale value and

offset I want to use when I convert TN

to RV mean reflectance value. This scale

value and this offset I need to apply

otherwise it gets some problem. So how I

can do it? So let's go ahead and

discuss.

So this thing we need to add in the

first part. So I put the it's a how to

convert.

So how to convert DN value?

How to

convert DN value to

reflectance

value?

So for that we need to use a uh we need

to create a function. Okay we need to

create a function. So this function

mainly convert dn to reflectance value.

Okay so let's go try to discuss about

that how we can easily convert.

So for that create a function suppose I

put the function variable name is that

apply

scale factor you can put any name I put

the name apply scale factor 12

I create a function

yeah I put the element name suppose

emails and what is my argument my

argument is two argument I

If you check it's very important because

um uh we can also if you check I'll show

you that just a minute. So if you check

SRB 1 to SRB 7 all of the similar okay

escal the similar similar similar okay

there is no problem similar similar also

offset also similar okay and it's

similar let's set 5 7 8 9 all of

collection are the same lens at 5 lens 7

lens at 8 lens at 9 all image collection

are the similar scale and similar offset

for optical ical bands. So it's called

the optical bands. Okay, it's called the

optical bands. But another bands it's

called the

thermal band. Okay, thermal band. What

is the thermal band? So thermal band

basically this one it's called the

thermal band. And you can see the

surface temperature it's called the

thermal band. So this band we need to

you can see it's another you can see

it's a this is the scale. This is the

offset. Okay. So thermal band is a

different scale different offset but

optical band if you check similar scale

and similar offset okay for the lenset 7

and lancet 8 lancet 5 all are the

similar for optical band so it's called

the optical bands and it's called the

thermal band so if you check the thermal

band are different not only for that

thermal band name is also different so

HTP6 this is the name for lancet 5 and

if you check for the lens set 8 it's the

another name Lancet 8 if you check

suppose Lancet 8

just check the thermal band so it's

another name

this band name is uh if you check STB10

so it's the thermal band okay but SKL

and offset is the similar it's not

problem it it's similar for lens set 5

lens set 7 lens set 8 lens set 9 it's

similar

But Lancet 8 and 9 this band name is STB

10. STB 10 uh thermal band for Lancet 8

and Lancet 9. Okay. And STB 6, Lancet 7

and Lancet 5. These two collection

thermal band is

STB6.

If you check STB6 this one it's create

some problem but we need to apply the

same similar band. So let's go ahead try

to show you that how we can use it. So I

hope you also get the idea what I say.

So in this case I simply show you that

uh lens set 8 and lens 7 lens at 9.

Okay. So this

optical

bands

optical bands

scale

and

offset

similar.

Okay. Not only for that we can also

write here some lens set 8 l set 7 also

landset five also similar. So we face a

one problem in here. This problem is

lens set

lens set

lens set 8

and lens at 9.

Here

the

bands

it name is

ST B10.

Okay.

Then L seven and L set five. This

thermal band name is

this thermal

bands

name is

ST

B6. Okay. So this difference we find out

but this scale about this scale

scale

and offset

offset similar

similar with

landset

landset

five

lens 7 thermal okay thermal thermal band

thermal band thermal bands okay so you

also get the idea what I say so laid 8

975 optical bands mainly SK and offset

similar lens at 8 9 thermal bands called

the STB 10 and L 75 thermal band is the

STB6 but STB 10 and STB 6 are the

similar are the similar

scale and offset value

okay similar value okay So now we create

a function. Okay. So we create a

function.

I put the things take a variable. So I

put the variable name is

optical optical bands

12. I call my element and select

select the all optical band. I just make

a list,

blue,

green,

red,

infrared.

S WIR 1

and S WIR 2. Okay. So these are optical

bands and what is my requirement? I want

to multiply I want to multiply with my

scale value and offset. So just I

multiply

this scale value. I already have this is

the scale value.

Just I copy

just I copy

paste here and add with offset value.

What is the offset value? This is the

offset value.

Just take copy

and paste here. So now optical is done.

So now another one thermal.

So take another variable.

Suppose I put the variable name is

thermal

bands.

and I call the element name image

and select the

band name. So in this case if you check

we already have the list all of bands.

Okay all of bands we already have the

list. So now I just use a common name.

Suppose this name is similar all time.

Okay. So for that I simply set a common

name. Common name I put suppose uh all

of band similar common name I put the

name any name you can put I put the name

suppose BT for brightness temperature

okay for the similar name I just put

so I put the name this band name is BT

brightness temperature but lit five and

seven it's come from

ST T B6

also for lens 7

ST

B6

then lens at 8 and 9 it's come from STB

10 because it's a different name okay ST

B 10 same

ST

P 10 okay so I can see now so STB6 okay

this is the name of lenset file Right?

But I put the name BT. So now you can

see similar. We find out the similar

sequence all of bands. Similar sequence.

Okay. BT. So now what can I do? I simply

put here uh BT BT because I put the name

is the BT in here. All of bands BT and

BT and these things I want to multiply

with the scale. So just put here the

multiply

the scale value scale value. If you

check the beta scale value is that this

decal value [snorts] just copy

and paste and add with offset value

offset value. This one is offset value

just copy

paste.

Okay. So it created function.

So now from this function what I return.

Okay. So what I return from this

function? I want to return

this function image

add bands

of optical bands just I call this

variable.

Okay. And put that null

proof. Same like as for add bands for

thermal band. put the add bands

of thermal bands

and put the null

and true. So now this is the function.

Okay, this function I need to connect

with my collection. I already have the

collection. If you check I already have

the collection all of image collection.

I simply connect using the map function.

Just connect also connect

lancet 8 then also connect lens 7 just

connect this function me just call the

function variable name and connect help

of map function

okay

okay there is no problem so now let's

try again and run the code to check the

value again what's happened

so now we try again to visualize the one

image and check the value

just click on here

suppose I put that uh red

green and blue still I give 98% so now

you can get that proper reflectance

value okay 0.05 05 something you can see

not very high value the reproed or scale

value just I apply

we can get that this is the image for

990 okay you can see you can get the

image 9090

so now we can also get the value just

click on here and then check then you

can get the all of reflectance value for

this specific coordinate we find out

that total six bands or seven bands we

find

display. So you can see find out that

okay so now okay it's not find out it

show that all of bands 252 bands okay

total 36

uh image okay 36 image total this type

of bands okay so now I want to display

okay suppose I want to display the image

uh 990 okay and also 2025

or any any years from the list any years

I can easily display the image. So how I

can display? So let's go to discuss

about that. So

for that take a variable. Suppose I put

the variable name is

what the name I put. You can put any

name. I put the name suppose uh image

90.

So all of image if you check in this

variable is stored. All of image total

36 image in this variable is stored.

Just a call and paste here. And I want

to filter the first year image. So I put

art engine dot filter dot equal. What's

equal? What? I put the year and set the

year 9990. So now what happened? So now

I make the mistake. So it's a annual

collection. Just copy and paste here

annual collection. So now from this

annual collection I just filter

year equal to 199. It means that image

of 9990

basically this image I simply choose and

store it in this store it in this

variable.

Okay.

And I want to use all of image and

create the composite. I use the mosaic

composite because I already create the

median composite. I put the mosaic. So

now I to display just a map dot add

layer and call the image

990 and paste here. So then it's not

need just I comment or remove you can

also remove that is no problem. So now I

just add the layer only for the image

9990. Okay just run the code. So now

it's the added one layer adding the

image of 9990

look like that. Okay so now I want to

add the band combination. Okay, band

combination very easy. Just click on

here. So now you can see it just show

the only one image. Okay, only one image

total seven bands or eight eight bands I

find out it's called the channel. So I

put the first channel it was uh which

one I put the red and green and put the

blue band RGB color I set and stress 98%

I keep to get the visualization just

apply.

So now you can see it loaded the image

1990 and when I click the import if you

see when I click the import a new line

of code added in my code editor just I

click it added just I copy

and further

just I paste after the object and if you

check it show the layer one but I want

to display the name so I put the name

suppose image uh 9990 just from the

port. So now we can get this is this is

the image

we can get this type of imagery.

So now I display suppose image 2025.

What can I do? Similar just copy the

same line of code and then paste. I

change the variable name uh image 2025

and I put the year equal to 2025 and I

want to display similar way map dot add

layer and image

2025

and run the code. So now it's added

another layer image 2025.

First time it show this type of color.

Okay. Then we simply set click on here

and I add the color suppose uh natural

color um I put suppose red

green and blue the stressing I keep has

a 98%.

and then further apply

and then import and further copy the

variable name you can this variable I

just copy and put the variable name

after the layer name you can see just I

put this name in here and put the layer

name I put the it's image 2025 so now

run the code so now I can add the two

image

who in one email you find out

from here.

So this is the way you find out this is

the image 2025

and this is the image for 1990.

So now you can easily do different types

of work for annually. Suppose you can

compute the vegetation, you can compute

the land use land cover, you can compute

the LST, you can compute the different

types of things. Not only for that you

can add the more image similar way.

Suppose I want to add the image for uh

1990 then 2000 just I copy

or 2010 just I put suppose image

image uh image

suppose 2010

and I set year equal to 2010 and for

that similar way just add the layer so

map dot add layer of image 2010 just I

copy Okay.

And now so now it's added the another

layer for image 2010.

So similar way just add the color band

combination. Click on here

and add suppose red green and uh blue

stressing keep as a 98%.

and further apply.

So then import this collection. Okay. So

now using this way we can easily add any

year of images and it's mainly anyone

media composite image.

It's image

20.

Okay. So now anyone need question you

can ask me the question about this code

or any confusion you can also ask me

about that please

and it's very useful because lot of time

we are use the same code for our

different types of research work or

project work for the LST historical LST

or historical NDVI Historical land use

can cover a lot of time we use the same

concept

for uh creating the different types of

work. Okay. Same concept we use.

>> Okay. Anyone question you can ask me.

>> Yes please.

[clears throat]

>> Hello dear dctor.

>> Yeah please.

Victor uh my project is soil erosion.

Can I use this code for NDVI for C

factor?

>> Yes. Yes. Yes. Yes. So we also discuss

about the soil erosion also. Okay. You

can also use it for that. No problem.

You can also use it. No problem.

>> Okay. For MDVI result for the soil

erosion. You can also use this type of

code. Okay. So we use this same code for

the different uh concept. Okay. Use the

same code for the different concept. So

inshallah I will try to explore all of

those in our total 14 days session. I

hope you can easily get the idea and you

also try to practice. So now in this

case I also set an assignment or

homework you also try to do. So for your

city

any any study area or for your city for

your city

make a

make a

make a suppose uh 30 years

in collection

and display

display the layer

display the layer 2090

then 2000

2010

2015 and 2025 try to display okay so for

your any city also try Because uh this

type of assignment you also do then you

can get the certificate and how how can

you submit the FM? I also try to show

you that also uh because it also help

you to uh more better study for when you

just practice and and not need to

memorize all of those. So you try to uh

understand all of code. Okay. And then

try to uh write your code your own hand.

Okay. Not copy and paste for the first

time. Okay. So when you become a

developer then copy and paste no

problem. But first time when you work

then try to write the code your hand not

copy and paste. Okay.

So you also try to for your city make a

30 years image collection. Okay. Uh and

also display this type of uh year layer.

Okay. Similar way you also try. Okay.

This similar way you also try. So then

you can easily display this type of

layer also in here.

So now we find out this type of imagery.

So basically now how I can compute the

NDVI suppose for NDVI NDVI basically

normalized difference vegetation index.

So NDVI so how I can make the NDVI

suppose annual NDVI or annual vegetation

map how I can create.

So for the NDVI when you create

we need to create a function we need to

create a function.

So how I can create a function okay for

NDVI? So let's get to create a function

for NDVI. So always you can see when I

write the function I add the top part.

So it's it's it's basic things because

this function further also connect with

our another collection. So you can see

this my function this function is simply

connect with my collection this

collection. Okay. So that's why this

function I simply write above the

collection. Okay. If I write the

function uh below the collection then it

will be not work. Suppose I write the

function in here below and then I

connect with my collection then it will

be not work. So always we try to write

our function above the collection. So

this is my collection. Let's say 9 87.

So I I always try to add the function

above the collection. I write the

function in here. So just I write the

function. Suppose I want to compute N

DVI. Okay.

How to compute NDVI

NDVI? How we can say compute

NVI? So for that how I can compute the

NDVI.

Okay. So for the NDVI uh take a variable

add band for add bands and create a

function. I put the function element

name suppose image and put my curly

bracket. In this curly bracket I put my

argument. So what is my argument for

this? I simply right in here.

So now what is my argument? So my

argument is I want to compute N DVI

using this element.

So take a variable suppose NDVI.

I call the element name element name and

apply normalize difference function

normalize

difference function between two bands

basically near infrared band and you can

checkh I put the name is N okay just use

the same name otherwise create some

problem so put the same name N

and red okay so I put the NDVI and put

the image it's my element name normalize

difference and I rename it I put the

rename function rename function I put

the name is NDVI

so now what I return so I want to return

this image add a bands add a bands

with this variable just like copy and

paste here so now it's one kind of

function okay this function I need to

connect with my collection so if you

check this is my collection. I already

have the collection. For connecting the

collection, use the map function always

the function and also connecting just

use add the function

and also use the connecting put here

this function.

Okay. As well as also connecting use

here this function and paste here.

Okay. So now what's the benefit I find

out from here? So now every year it

create the new band of NDVI.

If I check it just run the code. So then

if you check it from here suppose just

check it

uh total 36 image collection and I find

out that every years suppose you can say

years 990 it create the new band of

NDVI. Okay. every image, every image

create the new band of NDVI.

Okay. So now I want to display the NDVI

map. Suppose I want to display the NDVI

map for uh 2025.

Okay? Because if you check 2025, so this

is the image of 2025.

2025 and this image already have the

band of NDVI.

So now I can easily display the NDVI

map. Okay. So let's go to figure that

how I can display. So for that take a

variable ND vi 2025. I put this type of

name.

Okay. So, let's call it and I call the

image 2025. You can see I already have

the image 2025. Okay. This one I already

have because you can see I already

filtered the image 2025. Okay. Just I

call the image 2025. Okay. Just I call.

So from this image I just select only

one band of NDVI. Okay. So now I want to

add the layer. So my added layer of NDVI

2025

and run the board. You can check there

is no problem. It's added the new layer

of NDVI map 2025.

So let's create the NBMF look like that

without color.

So now we try to add the color.

So let's go. So for the color click on

here

and I customize just ST 98%.

And I the color suppose I can put any

color. I put the color suppose uh for

that low vegetation show this type of

color.

So greater than NDVI pixel it show this

type of color.

So greater than it shows this type of

color. So I put the three color

combination can also put the more color

no problem just I apply. So now this

type of green color indicating about the

vegetation area. So you can see this

type of green color indicating about the

vegetation area. Just import okay just

import. So then new line of code will be

added in my map. If you check it just

added the new map just I copy and paste

after the object name

and I put the name. It's a NDVI 2025. So

I want to add the NDVI map for uh 9990.

So how I can add? Same way take a

variable NDVI

1990. I already have the image 1990.

Just I copy and paste here. So from here

I select select only for the one vent.

This is NDVI and further just I copy.

and paste and now call ND 91 2000 I just

put and paste here and further I put the

name suppose NDVI

say 1990 and I [snorts] add the color

later so for that I simply put the empty

and run the code so now I add the color

later to get the better visualization

Okay. So then it added this type of NDV

map. So I want to add the color. Click

on here again.

Okay. Just Okay. It also better we try

to use the same color. Suppose this

color I want to also apply in this map.

So how I can apply? Suppose this color

this color I to apply this map. So for

that click on the setting button and

click on the pen sign and copy the color

port just like copy and further

just click on here and open the setting

button

and then further uh click on the plated

click on the pen sign and paste. Okay.

and also stressing I keep the similar 98

and further apply and also import

and further just copy the variable name

this one

and it's connect with your

one and now run the code so now it show

the result to NDVI map if you check the

two time period

So you can see this is the NDVI this

type of place is the vegetation green

colors the vegetation and this is the

result for 2025 you can see

a lot of resident decrease between this

time period.

So let's discuss this type of things one

by one. How to calculate the change

detection? How to calculate the NDVI

trend. So all of things I try to

explain.

So you can see this is scenario. So this

is the map 1990 of NDVI and this is the

map 2025.

Okay. So now you also try to uh I put

the another assignment you also try. So

compute NDVI

compute the NDVI

for,000

and 2025

for your city

then add the layer this type of layer

for 20 and 2025 for your own city. Okay.

So this is another homework. [laughter]

So I always put the aw. So aw means the

homework. You can also uh a means home

you also try. Okay. When you practice

you also try to practice this type of

homework. Okay. So this is the way. So

now you can easily add all of years any

years any years you can easily add from

the NDVI. Not only for the NDVI suppose

I need the NDWI. What is the NWI? NWI is

that normalized difference water index.

Okay. mainly uh surface water we can

easily get the idea from the NDWI. So

how I can get the NWI similar way just

make a band in here. So I already

created a function if you check. So I

just create the band take a variable

NDWI

call the image

normalized

difference between two band

green band

and

band. So this is a formula about the NWI

and I rename it

as a

NDWI and also add bands put here the add

bands of NDWI.

So now what happened? So now all of

image connecting a new band of NDWI. So

if you check the image collection

all of years all of years image

collection added the new band suppose

this is the image for 1995 and if you

check it added the new band of NDWI so

now I can easily make the NWI map okay

and we can get the water body map so let

me show you the one example for water

body map suppose I want to create the

NWI map for same same here okay just I

copy the same line of code because it

okay just I put the same line of code

and paste here. So I put suppose it's a

you can put any name I put the name

suppose NDWI

and also put the NDWI.

I simply choose the band name NDWI.

Choose the band name NDWI

and I put that say ND

NDWI

2025

and ND WI 19990 and call this NI 2025

and NWI

9990.

So I add the color letter I put the

empty also put the empty. So now run the

code.

So now we simply get the another layer

of NWA

1990 and 2025.

So now we can easily get the water body

map from here from this tool layer. So

not only for that any index any index

you can compute this way for any years.

So let's try to add the more other index

because remote sensing have the lot of

index. So let's go try to add the color

map.

So I want to add the color for 990. Just

a click

and sting I keep as a 98% I keep

and further add the color. Suppose where

there is no water I would display as a

this type of color.

Okay. And g then I to display this type

of color.

So get than I to display this type of

green color.

And where is the positive value? It's a

water. I want to suppose this type of

blue color for water body and display

the water body. And I copy.

We copy it later. Just I apply

or if you want you can also change the

stressing levels. I put the stress

levels of three sigma. So stressing

level it show the different types of

value. Okay. For displaying your map.

I think this stressing level is more

better because we can easily extract the

water body as a blue color. Okay. Just I

keep it import.

further uh just I

call it

just copy the variable name

and paste after the object name for

9990.

So now add the similar color for that

2025 of NWI. Same way just click on the

pen sign

and copy

and then uh 2025 this layer I want to

add just click

I put the station with three sigma again

I put the same

and then just

sign

and apply okay and also apply

and also import.

So this thing I want to add

after the object name just I copy the

variable name and paste

after in here just I remove the empty

and then paste. Okay. So this is the way

to adding the color. So now run the

code. So now it's added the another two

map of 10 DWI

we can get it this type of

This is the map 9990 you can see lot of

water body we find out and 2025 very

less water body you can see

so this are 1990 water body map this are

2025 Okay.

Okay. Any question please

or any confusion you can ask me please.

Okay, I hope there is no problem. So

just we make a break for 10 minutes. So

after 10 minutes we'll meet again. Okay,

just make a break. Short break for 10

minutes.