Supervised Classification with Sentinel 2 Satellite Images | Earth Engine | Machine Learning

Подписаться 13 тыс.

Просмотров 35 тыс.

50% 1

Learn more: spatialelearning.com
Enroll in the full course: bit.ly/earth-engine-mega-course
Subscribe to our channel for more tutorials: / spatialelearning
In this tutorial, we will discuss how to apply supervised machine learning classification with Sentinel 2 images in Google Earth Engine.
In this tutorial, you will learn how to:
1. Import Sentinel 2 data from the Google Earth Engine data library
2. Create the best cloud-free image
3. Create land cover training data
4. Apply CART (Classification and Regression Tree) supervised classification algorithm
5. Calculate the confusion matrix and overall accuracy
6. Display and visualize classification result
7. Download the classification result
Blog: spatialelearning.com
Courses: courses.spatialelearning.com/
RU-vid: / spatialelearning
Facebook: / spatialelearning
Twitter: / spatialelearn
#earthengine #remotesensing #sentinel2

Опубликовано:

31 июл 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 57

@spatialelearning 3 года назад

Visit our website to learn more: spatialelearning.com Hope this video was useful to you. What video topic would you like to see next?

@ishanweerakoon9838 3 года назад

Please do a video about Parallelepiped classifier

@harshelsuares3194 2 года назад

It's a wonderful recording. But I have a query related to ROI. So if I want the shapefile of British Columbia or Vancouver what will be the FeatureCollection code ?

@spatialelearning Год назад

@@ishanweerakoon9838 Ok, will do.

@spatialelearning Год назад

@@harshelsuares3194 You can upload a shape file to asset manager.

@asmaaalzahrani6387 Год назад

Please can you share the code link

@donaldandrade2037 Год назад

Great work! So informative and detailed thank you so much for this!!!

@leidycastro9519 Год назад

Gracias!! :)

@theforester_ 2 года назад

awesome video. i got a question, i've got an preprocessed shape with possible areas of pine plantation in brazil, i just i'd like to run a classification to verify inside these polygons wethers it's pine or not. Is there a way to transform a shapefile in FeatureCollection for then use it as my ROI? thanks again

@user-kp6jn7cp4h Год назад

Hi dear admin, Please may I ask you: when I zoom the google earth engine interface imagery, the ground fearures look like pretty clear and visible, but when I use your algorithm to make true color composite imagery to download, but at that time the ground fearures could not clearly visible? What is this problem and how to solve it ? I will be thankful for your valuable guidance about it.

@user-kp6jn7cp4h Год назад

Hi dear Admin. When I run this code: //Overlay the point on the imagery to get trainingDataset var trainImage = imput.sampleRegion({ collcetion : training, properties: [lable], scale: 30 }); then this error appeared: "Line 90: imput is not defined" Please tell me how to handle this issue?

@aarizahmed2422 3 года назад

Very nice video! I an having some query I want to extract only urban class but while creating the different classes the Urban class mixed with the Barren class.kindly help me for the same.

@sanjanadutt2941 Год назад

Hi, can someone please help me with this code error: Cannot read property 'sampleRegions' of undefined. 34th minute of the video. I tried exactly as it was given but its not working.

@allkindofknowledge1400 2 года назад

Best tutorial so far with detailed working. Sometimes people hide some of the steps in the video, and that's create confusion when you try to implement the same thing. I have one question why you have choose band b4, b3, b2, b8 for classification? Is it true for all kindof classification using sentinel images. Which bands we can use in case of landsat8 data

@spatialelearning Год назад

Thanks. I chose bands 4, 3, 2 and 8 since they are the highest resolution (10 meter). You can apply similar band combinations for Landsat 8.

@surabhisoni3460 2 года назад

Great video sir, how to export confusion matrix?

@rahelsintayehu9070 Год назад

Thank you again for wonderful tutorial. I have one question is it possible to get the satellite map in Google earth engine in historic period like google earth pro?

@ounassabdelaziz7349 2 года назад

i try to do the same but i have this message ! (Jsloader error (code #1): Timeout reached for loading script /javascript/acorn.js) please can some one help me to fix this ?

@gaaty1954 Год назад

Why when i use "Sentinel-2: Cloud Probability" to mask the clouds, my overall accuracy went down from 0.7 to 0.578. 1) How to improve overall accuracy 2) How many Geometry Imports per class is need/best/sufficient

@Bev16 Год назад

This is great!! Can you please share the code in the comments?

@fatimaezzahraallam438 Год назад

I can't get the confusion matrix scrip to work, it doesn't give me any error, only that both the confusion matrix and the Overall Accuracy give me 0 value. What can it be?

@HarishPandey-py3iz Год назад

My classification is appearing very decent with 85% accuracy; However, the area is very large and while exporting, in task a multiple run options are appearing, can anyone help its urgent, Thanking you, I want to export in a single frame.

@fatematuzzohora3154 2 года назад

Could you please explain the reason for using only Band 2, 3, 4 and 8 for classification?. It's on line 22, under the comment //Creating Training Data.

@rajithanuwantha3313 2 года назад

They are the ones having spatial resolution of 10m .. Using other bands may reduce the resolution..

@salamander5969 Год назад

I CANT BELIEVE THAT THIS IS WORKKKKKKKK, CAN I ASK YOU QUESTION WHY YOU USE ONLY 4 BANDS, MEANWHILE MAYBE THER IS SOME CLASSIFICATION THAT THE OTHER BAND

@UlukbekBurganakov 4 месяца назад

Hi! Can you share me with your codes

@amenechelias3858 2 года назад

Please can you help me, I followed all steps and worked. But Finally it shows ''Classification_CART: Layer error: Property 'class' of feature '1_1_1_1_1_0_0' is missing''. How can I fix this problem . I am binger

@Google_Earth_Engine 2 года назад

check featureCollection of your landcover

@atuldiwaker8597 Год назад

@@Google_Earth_Engine facing same problem, please help me out

@gaaty1954 Год назад

var trainImage = input.sampleRegions({ collection: training, properties: [label], scale:30 }); Why is it at the trainImage, the scale is selected to be 30 instead of 10?

@gaaty1954 Год назад

hmmm, when i adjust scale to 10, the water classification output seems a bit off, using 30 then. Using the scale of 30, also improved the overall accuracy from 0.7 to 0.75

@leonardoscarpa8784 2 года назад

Very good video. Congratulations! Just one doubt, I can't get the confusion matrix scrip to work, it doesn't give me any error, only that both the confusion matrix and the Overall Accuracy give me 0 value. What can it be?

@binyamhailu1771 2 года назад

I am having the same problem.

@sanabasheer7650 2 года назад

Hi... I'm having the same problem. have you find any solution related to that?

@sanabasheer7650 2 года назад

@@binyamhailu1771 Have you find any solution?

@rahelsintayehu9070 Год назад

@@sanabasheer7650 you edit the script actual: 'Class' then you will get the result

@rahelsintayehu9070 Год назад

@@binyamhailu1771 Hey edit the script actual: 'Class" then you will get the result

@sandigis8147 2 года назад

WOULD YOU LIKE TO SHARE THE SCRIPT?

@michaire5999 Год назад

is it possible to convert the result in shp?

@spatialelearning Год назад

Yes. You can use raster to vector function.

@mattiabarsanti9336 2 года назад

I have an error: line 14, Map.addLayer(image.clip(roi), visParamsTrue, 'Sentinel 2020'); -> Error: image.clip is not a function Someone can help me? Thanks in advance

@spatialelearning Год назад

Check the roi if it matches to image coverage.

@mattiabarsanti9336 Год назад

@@spatialelearning solved. thank you.

@frevaboy298 Год назад

@@mattiabarsanti9336 sorry can you share to me the code you Used

@asimrizwan3960 2 года назад

Share this video code please...

@christy_v_7394 Год назад

Error: Image.clipToBoundsAndScale, argument 'input': Invalid type. Expected type: Image. Actual type: ImageCollection. (Error code: 3) This is the error that occured to me

@alexmurphy5015 2 года назад

ERROR - I got this error right at the end (heartbreaking!)😭 can anyone help?????? "Layer error: Property 'Class' of feature '2_0_0': Invalid type. Expected type: Float. Actual type: String. Actual value: C"

@spatialelearning Год назад

Can you share the ee code link?