lensfunpy

To correct both vignetting and TCA, you must follow a specific order: Vignetting must be corrected first.

Steps:

1. Initialize lensfunpy.Modifier with the flags parameter set to lensfunpy.ModifyFlags.VIGNETTING | lensfunpy.ModifyFlags.TCA.
2. Call mod.apply_color_modification(img) to handle vignetting.
3. Call mod.apply_subpixel_distortion() to get the subpixel distortion coordinates.
4. Use cv2.remap on each color channel (R, G, B) separately using the corresponding channel from the distortion coordinates.

Note: The input image must be in a linear state.

import imageio
import cv2
import lensfunpy

db = lensfunpy.Database()
cam = db.find_cameras('Canon', 'Canon EOS 5D Mark IV')[0]
lens = db.find_lenses(cam, 'Sigma', 'Sigma 8mm f/3.5 EX DG circular fisheye')[0]

img = imageio.imread('/path/to/image.tiff')

focal_length = 8.0
aperture = 11
distance = 10
width = img.shape[1]
height = img.shape[0]

mod = lensfunpy.Modifier(lens, cam.crop_factor, width, height)
mod.initialize(focal_length, aperture, distance, pixel_format=img.dtype, flags=lensfunpy.ModifyFlags.VIGNETTING | lensfunpy.ModifyFlags.TCA)

# 1. Vignette Correction
mod.apply_color_modification(img)

# 2. TCA Correction
undist_coords = mod.apply_subpixel_distortion()
img[..., 0] = cv2.remap(img[..., 0], undist_coords[..., 0, :], None, cv2.INTER_LINEAR)
img[..., 1] = cv2.remap(img[..., 1], undist_coords[..., 1, :], None, cv2.INTER_LINEAR)
img[..., 2] = cv2.remap(img[..., 2], undist_coords[..., 2, :], None, cv2.INTER_LINEAR)

imageio.imwrite('/path/to/image_corrected.tiff', img)

To correct geometric distortion, use lensfunpy.Modifier. This process typically involves generating undistorted coordinates and remapping the image using a library like OpenCV.

Steps:

1. Create a lensfunpy.Modifier instance with the lens, camera.crop_factor, width, and height.
2. Initialize the modifier with focal_length, aperture, distance, and the image's pixel_format.
3. Call mod.apply_geometry_distortion() to get the undistorted coordinates.
4. Use cv2.remap (OpenCV) or SciPy to apply the transformation to your image.

Note: The lensfunpy.util module provides convenience functions for RGB images that handle both OpenCV and SciPy automatically.

import cv2
import lensfunpy

focal_length = 28.0
aperture = 1.4
distance = 10
image_path = '/path/to/image.tiff'
undistorted_image_path = '/path/to/image_undist.tiff'

img = cv2.imread(image_path)
height, width = img.shape[0], img.shape[1]

# Assuming 'lens' and 'cam' were found via Database()
mod = lensfunpy.Modifier(lens, cam.crop_factor, width, height)
mod.initialize(focal_length, aperture, distance, pixel_format=img.dtype)

undist_coords = mod.apply_geometry_distortion()
img_undistorted = cv2.remap(img, undist_coords, None, cv2.INTER_LINEAR)
cv2.imwrite(undistorted_image_path, img_undistorted)

If you need a specific lensfun version or cannot use binary wheels, follow these steps:

1. Install liblensfun:
   • Ubuntu: sudo apt-get install liblensfun-dev
   • Or build from source:

     git clone https://github.com/lensfun/lensfun
     cd lensfun
     cmake .
     sudo make install

2. Install lensfunpy:

   git clone https://github.com/letmaik/lensfunpy
   cd lensfunpy
   pip install numpy cython
   pip install .

Troubleshooting ImportError: liblensfun.so.0 on Linux: If the library is not found, add /usr/local/lib to your dynamic linker configuration:

echo "/usr/local/lib" | sudo tee /etc/ld.so.conf.d/99local.conf
sudo ldconfig

Note: The installation path might vary (e.g., /usr/local/lib/x86_64-linux-gnu or /usr/local/lib64).

# Install liblensfun from source
git clone https://github.com/lensfun/lensfun
cd lensfun
cmake .
sudo make install

# Install lensfunpy
git clone https://github.com/letmaik/lensfunpy
cd lensfunpy
pip install numpy cython
pip install .

To correct vignetting, use lensfunpy.Modifier.apply_color_modification(img).

Important Requirements:

• The input image must be in a linear state (not gamma corrected).
• The method returns a boolean indicating if the correction was applied. If it returns False, calibration data for that lens/camera combination might be missing.

Steps:

1. Initialize lensfunpy.Modifier with lens, crop factor, width, and height.
2. Call mod.initialize(focal_length, aperture, distance, pixel_format=img.dtype).
3. Call mod.apply_color_modification(img) to modify the image in-place.

import lensfunpy
import imageio

db = lensfunpy.Database()
cam = db.find_cameras('NIKON CORPORATION', 'NIKON D3S')[0]
lens = db.find_lenses(cam, 'Nikon', 'Nikkor AF 20mm f/2.8D')[0]

img = imageio.imread('/path/to/image.tiff')

focal_length = 20
aperture = 4
distance = 10
width = img.shape[1]
height = img.shape[0]

mod = lensfunpy.Modifier(lens, cam.crop_factor, width, height)
mod.initialize(focal_length, aperture, distance, pixel_format=img.dtype)

did_apply = mod.apply_color_modification(img)
if did_apply:
    imageio.imwrite('/path/to/image_corrected.tiff', img)
else:
    print('vignetting not corrected, calibration data missing?')