Vjylku

What is the tangent at a sharp point on a curve?

Print last inputted byte

"Marked down as someone wanting to sell shares." What does that mean?

Hackerrank All Women's Codesprint 2019: Name the Product

Do I need an EFI partition for each 18.04 ubuntu I have on my HD?

Air travel with refrigerated insulin

pipe commands inside find -exec?

Why is participating in the European Parliamentary elections used as a threat?

Someone scrambled my calling sign- who am I?

Is "inadequate referencing" a euphemism for plagiarism?

How to remove space in section title at KOMA-Script

How much propellant is used up until liftoff?

Determine voltage drop over 10G resistors with cheap multimeter

Isn't the word "experience" wrongly used in this context?

Why do I have a large white artefact on the rendered image?

Does fire aspect on a sword, destroy mob drops?

Triple Trouble Tribond

Turning a hard to access nut?

Why I don't get the wanted width of tcbox?

Extraneous elements in "Europe countries" list

How can I create URL shortcuts/redirects for task/diff IDs in Phabricator?

How to read string as hex number in bash?

Exposing a company lying about themselves in a tightly knit industry: Is my career at risk on the long run?

How to find the largest number(s) in a list of elements?

rasterio: converting from tif to png and appropriate pixel ranges

How to use GDAL to convert Int16 data to ByteGeoTiff convert to PNG with transparency?Can GeoWebCache response GeoTiff format imageCalculating NDVI with RasterioGdal Python Problem : Black strips in geotiffTranslating GeoTIFF to 16-bit PNG with GDALVectorise binary GeoTiff Raster (GeoTiff to Shapefile)QGIS rendering TIFF color with wrong colorConvert TIFF raster layer to PNGNumPy array to Raster file (GeoTIFF)

0

I am trying to convert some GeoTIFF files to PNG (or JPEG) format in order to apply deep learning algorithms. I need to import the GeoTIFF into python and then convert the image to a PNG format numpy array. I don't want to write the PNG file to disk, as most of the examples I have found suggest.

So I can read a GeoTIFF file into rasterio easy enough:

 import rasterio

 img = rasterio.open('/../../filepath')

This will give me a numpy array with values ranging from a minimum of 282 to a maximum of 1560. So I think the values in that range are a bit out of whack from a regular PNG (or JPEG) format.

The question is how would I convert pixels in this larger range to a standard PNG format, which I imagine is more in line with a range of like [0, 255] or [0,1]? I just know know the correct algorithm to use to implement the transformation--meaning I don't know the PNG standard or the TIFF standards well enough to know what the proper transformation are to apply to the data, and what the true ranges are for their pixel values. Is there a function to implement this transformation in rasterio or GDAL or can someone just indicate the algorithm and I can code it myself.

The thing that actually clued me into this issue, was that I could take the rasterio img variable--coded in this big [282, 1560] range--and pass it totensorflow`. But when I went to plot the converted image in a jupyter notebook--to make sure the data came in okay--I was getting a warning that said:

Clipping input data to the valid range for imshow with RGB 

data ([0..1] for floats or [0..255] for integers).

So I was getting a blank image because all of my pixels were above this range.

python gdal geotiff-tiff rasterio png

share

asked 5 mins ago

krishnabkrishnab

241214

add a comment | 

0

I am trying to convert some GeoTIFF files to PNG (or JPEG) format in order to apply deep learning algorithms. I need to import the GeoTIFF into python and then convert the image to a PNG format numpy array. I don't want to write the PNG file to disk, as most of the examples I have found suggest.

So I can read a GeoTIFF file into rasterio easy enough:

 import rasterio

 img = rasterio.open('/../../filepath')

This will give me a numpy array with values ranging from a minimum of 282 to a maximum of 1560. So I think the values in that range are a bit out of whack from a regular PNG (or JPEG) format.

The question is how would I convert pixels in this larger range to a standard PNG format, which I imagine is more in line with a range of like [0, 255] or [0,1]? I just know know the correct algorithm to use to implement the transformation--meaning I don't know the PNG standard or the TIFF standards well enough to know what the proper transformation are to apply to the data, and what the true ranges are for their pixel values. Is there a function to implement this transformation in rasterio or GDAL or can someone just indicate the algorithm and I can code it myself.

The thing that actually clued me into this issue, was that I could take the rasterio img variable--coded in this big [282, 1560] range--and pass it totensorflow`. But when I went to plot the converted image in a jupyter notebook--to make sure the data came in okay--I was getting a warning that said:

Clipping input data to the valid range for imshow with RGB 

data ([0..1] for floats or [0..255] for integers).

So I was getting a blank image because all of my pixels were above this range.

python gdal geotiff-tiff rasterio png

share

asked 5 mins ago

krishnabkrishnab

241214

add a comment | 

I am trying to convert some GeoTIFF files to PNG (or JPEG) format in order to apply deep learning algorithms. I need to import the GeoTIFF into python and then convert the image to a PNG format numpy array. I don't want to write the PNG file to disk, as most of the examples I have found suggest.

So I can read a GeoTIFF file into rasterio easy enough:

 import rasterio

 img = rasterio.open('/../../filepath')

This will give me a numpy array with values ranging from a minimum of 282 to a maximum of 1560. So I think the values in that range are a bit out of whack from a regular PNG (or JPEG) format.

The question is how would I convert pixels in this larger range to a standard PNG format, which I imagine is more in line with a range of like [0, 255] or [0,1]? I just know know the correct algorithm to use to implement the transformation--meaning I don't know the PNG standard or the TIFF standards well enough to know what the proper transformation are to apply to the data, and what the true ranges are for their pixel values. Is there a function to implement this transformation in rasterio or GDAL or can someone just indicate the algorithm and I can code it myself.

The thing that actually clued me into this issue, was that I could take the rasterio img variable--coded in this big [282, 1560] range--and pass it totensorflow`. But when I went to plot the converted image in a jupyter notebook--to make sure the data came in okay--I was getting a warning that said:

Clipping input data to the valid range for imshow with RGB 

data ([0..1] for floats or [0..255] for integers).

So I was getting a blank image because all of my pixels were above this range.

python gdal geotiff-tiff rasterio png

share

asked 5 mins ago

krishnabkrishnab

241214

 import rasterio

 img = rasterio.open('/../../filepath')

Clipping input data to the valid range for imshow with RGB 

data ([0..1] for floats or [0..255] for integers).

share

asked 5 mins ago

krishnabkrishnab

241214

share

asked 5 mins ago

krishnabkrishnab

241214

asked 5 mins ago

krishnabkrishnab

241214

asked 5 mins ago

krishnabkrishnab

241214