How to download the fixed size [64,64] Image chip one by one , no random?

[Dushuai12138]

I want to download image chips like this:

how could i achieve it? please give me some advice. Thanks a lot!

[Dushuai12138]

@giswqs
Respected Prof. Wu, do you have any idea? thanks.

[giswqs]

Download the image using geemap, then use leafmap.split_raster() to split the image.
https://leafmap.org/notebooks/77_split_raster/

[Dushuai12138]

wow! Thanks so so so much! Downloading big size image is always a problem for me, now I re-learn geemap and solve it. How helpful geemap is! And spliting on local computer is not a problem for me.

[Huanglh1001]

I am also very interested in the solution to this problem. Could you please share the relevant implementation code?

[Dushuai12138]

I think you can try command clipToBoundsAndScale, which can clip the image into fixed size. Then download image to local computer using geemap, then use leafmap.split_raster() to split the image as Prof. Wu mentioned. Or you can use gdal package to deal with tiff images.

[Dushuai12138]

I am also very interested in the solution to this problem. Could you please share the relevant implementation code?

I think you can try command clipToBoundsAndScale, which can clip the image into fixed size. Then download image to local computer using geemap, then use leafmap.split_raster() to split the image as Prof. Wu mentioned. Or you can use gdal package to deal with tiff images. I have not finished this work because I am busy lately. So I don't have any code to share with you.

[Dushuai12138]

I am also very interested in the solution to this problem. Could you please share the relevant implementation code?

image = ee.Image('your image') 
out = r'image.tif'
geemap.download_ee_image(image, out, shape=[4320,7560])
outdir = r'tiles'
leafmap.split_raster(out, out_dir=outdir, tile_size=(64,64), overlap=0)

[Huanglh1001]

Thanks!!!

[Dushuai12138]

It seems this command leafmap.split_raster face failure when dealing with big size images of about 100G.

[giswqs]

Why do you want to have an image > 100 GB? You can try using GDAL to split it.

[Dushuai12138]

yeah, GDAL is a good choice. I directly download daily temperature_2m data from era5 of one year, and it shows 94G in total in jupyternotebook during downloading. However, it only occupy 1.2G storage in local. I tried to use leafmap.split_raster to split it and failed. So I turn to GDAL. But leafmap is more convienient.