CODEWITHSUNDEEP
pythonimageopencvnumpytiff
Damuno
Damuno

Reputation: 171

OpenCV - Read 16 bit TIFF image in Python (sentinel-1 data)

I'm trying to read a 16 bit TIFF image (26446 x 16688) in Python. Using OpenCV only reads a black image (all the intensities reads 0) by:

    self.img = cv2.imread(self.filename, cv2.IMREAD_UNCHANGED)

Can openCV handle 16 bit or large images (~840mb)? Any workaround?

EDIT: Also cv2.imshow("output", self.img[0:600]) displays a black image.

Upvotes: 3

Views: 2687

Answers (3)

apatsekin
apatsekin

Reputation: 1624

Yes, OpenCV can read uint16 perfectly fine:

img_np = cv2.imread(path, cv2.IMREAD_ANYDEPTH | cv2.IMREAD_UNCHANGED)

Speaking of file size limitation, depends on number of pixels:

By default number of pixels must be less than 2^30. Limit can be set using system variable OPENCV_IO_MAX_IMAGE_PIXELS

https://docs.opencv.org/3.4/d4/da8/group__imgcodecs.html#ga288b8b3da0892bd651fce07b3bbd3a56

Upvotes: 1

Anonymous Analyst
Anonymous Analyst

Reputation: 11

Use image libraries specific to raster data (in your case, sentinel-1). For example you can use rasterio for reading and displaying satellite images.

Example:

import rasterio
from rasterio.plot import show

img = rasterio.open("image.tif")

show(img)

Upvotes: 1

Jeru Luke
Jeru Luke

Reputation: 21233

As suggested by Andrew Paxson a different library can be used. There is a library dedicated exclusively for playing with tiff images.

Use the following code for the same. Make sure you have tif installed in your system.

import tifffile as tiff
import matplotlib.pyplot as plt 

filename = 'Image.tif'
img = tiff.imread(filename)
plt.imshow(img)

Upvotes: 0

Related Questions