module snowpyt.nirpy

Collection of NIR processing tools S. Filhol, March 2022

Inpired by the paper by Matzl and Schneebeli 2016 for more info

A calibration profile was derived from the camera with Micmac. This calibration profile is for correcting vigneting. correct image for vignetting (calib profile is of the size of Raw images. Crop from center to use with jpegs detrend if needed the luminosity, as it can vary linearly from top to bottom of the snowpit sample targets for absolute reflectance calibration (white= 99%, and grey=50%). Fit a linear model Convert reflectance image to SSA with the conversion equation 𝑆𝑆𝐴=𝐴𝑒𝑟/𝑡

Finally, use the ruler (or other object of know size) in image to scale image dimension to metric system.

TODO:

  • write function to extract SSA profile to import in niviz.org

  • Raw images are in 12 bit. Find a way to convert to BW from original while maintaining the 12bit resolution. Rawpy might be useful. Then make sure the processing pipeline can accept 12bit data (i.e. all skimage functions)

  • wrap micmac function to extract profile ‘mm3d vodka’. At least provide the method on how to do it.

function kernel_square

kernel_square(nPix)

Function to defin a square kernel of equal value for performing averaging

Args:

  • nPix (int): size of the kernel in pixel

Returns:

  • array: kernel matrix

function smooth

smooth(mat, kernel)

Function that produce a smoothed version of the 2D array

Args:

  • mat: 2D Array to smooth

  • kernel: kernel array (output) from the function kernel_square()

Returns:

  • 2D array: smoothed array

function micmac_radiometric

micmac_radiometric()

List of commands to run for deriving a radiometric calibratino profile for the camera

class nir

Class to process NIR snowpit photograph.

method __init__

__init__(
    fname_nir,
    fname_calib,
    highpass=True,
    kernel_size=2000,
    rotate_calib=False
)

Class initialization

Args:

  • fname_nir (str): path to NIR image

  • fname_calib (str): path to radiometric calibration image

  • highpass (bool): perform high pass filtering to remove luminosity gradient across the pit

  • kernel_size (int): size of the kernel for the highpass filter

method apply_calib

apply_calib(crop_calib=False)

Function to apply calibration profile to the NIR image.

Args:

  • crop_calib (bool): if calibration and image are of slightly different size, crop calib and align the two with center.

method convert_all

convert_all()

Function to convert pixel values to physical values using the targets

method convert_to_SSA

convert_to_SSA()

Function to convert reflectance to SSA

method convert_to_doptic

convert_to_doptic()

Function to convert SSA to optical diameter

method convert_to_reflectance

convert_to_reflectance()

Function to convert image to reflectance using at minimum 2 sets of reflectance targets previously picked

method extract_profile

extract_profile(
    imgs=['SSA', 'reflectance', 'd_optical'],
    param={'method': <module 'scipy' from '/home/simonfi/miniconda3/envs/dataAna/lib/python3.8/site-packages/scipy/__init__.py'>, 'n_samples': 1000}
)

Function to extract profile of values for a list of images

Args:

  • imgs (list): images from which to sample profile param (dict):

  • method (str): method to sample the profile. Avail: numpy, scipy, and skimage.

  • n_sample (int, numpy and scipy method): number of samples along profile

  • linewidth (int, skimage method): width of the profile

  • reduce_func (func, skimage method): function to agglomerate the pixels perpendicular to the line

  • spline_order (int, 0-5, skimage method): order of the spline applied to the sampled profile

examples: {’method’: scipy, ‘n_samples’:1000}, {’method’: numpy, ‘n_samples’:1000}, {’method’: skimage, ‘linewidth’:5, ‘reduce_func’:np.median, ‘spline_order’:1}

method load_calib

load_calib()

Function to load radiometrci calibration file

method load_nir

load_nir()

Function to load jpeg NIR images, and convert them to BW

method pick_targets

pick_targets(reflectances=[99, 50])

Function to pick reflectance targets

Args:

  • reflectances (list of int): List of reflectance targets to pick

method scale_spatially

scale_spatially()

Function to bring real spatial coordinate

Method: 1. click two points 2. provide corresponding length 3. option to provide geometrical correction

This file was automatically generated via lazydocs.