Skip to content

Latest commit

 

History

History
130 lines (108 loc) · 4.2 KB

README.md

File metadata and controls

130 lines (108 loc) · 4.2 KB

Here’s the updated README.md for your project, OpenHDRISP, incorporating all the details you provided:

OpenHDRISP: RAW to SDR/HDR Image Signal Processor

Overview

OpenHDRISP is a lightweight and flexible Image Signal Processor (ISP) designed to convert RAW images into Standard Dynamic Range (SDR) or High Dynamic Range (HDR) images. It implements a complete pipeline from RAW data processing to SDR/HDR image generation, including black level correction, demosaicing, and color space transformations.

Features

  • Black Level Correction (BLC): Corrects sensor black level offsets for accurate color representation.
  • Demosaicing: Converts Bayer-pattern RAW data into full-color RGB images.
  • Two RAW-to-XYZ Conversion Methods:
    • AWB + CCM: Automatic White Balance (AWB) followed by Color Correction Matrix (CCM).
    • Direct Characterization Model: A model-based approach for direct RAW-to-XYZ conversion.
  • XYZ to SDR/HDR Processing: Converts XYZ color space data into SDR or HDR images using advanced tone mapping and gamma correction.
  • HDR Support: Generates HDR images using the Pillow-HEIF library.

Requirements

  • Python 3.11.0
  • Dependencies: 
    contourpy==1.3.1
cycler==0.12.1
ExifRead==3.0.0
fonttools==4.55.3
kiwisolver==1.4.7
matplotlib==3.10.0
numpy==2.2.0
packaging==24.2
piexif==1.1.3
pillow==11.0.0
pillow_heif==0.21.0
pyparsing==3.2.0
python-dateutil==2.9.0.post0
rawpy==0.23.2
six==1.17.0

Installation

  1. Clone the repository:

    git clone https://github.com/UltraMo114/OpenHDRISP.git
cd OpenHDRISP

  2. Install the required dependencies:

    pip install -r requirements.txt

Usage

Below is an example of how to use OpenHDRISP to process a RAW image into an SDR image:

import rawpy
import numpy as np
import exifread
import math
from modules.RawBlc import RawBlc
from modules.RawToRgb import RawToRgb
from modules.RgbToXyz import RgbToXyz
from modules.XyzToRgb import XyzToRgb
from modules.RgbToImg import RgbToImg

# Path to the RAW file
raw_file_path = "DSC04657.dng"

# Step 1: Read EXIF metadata
with open(raw_file_path, 'rb') as f:
    tags = exifread.process_file(f)

    # Extract relevant metadata
    shutter_speed = tags.get('EXIF ShutterSpeedValue')
    aperture = tags.get('EXIF ApertureValue')
    iso = tags.get('EXIF ISOSpeedRatings')

    # Print metadata
    print(f"Shutter Speed: {shutter_speed}")
    print(f"Aperture: {aperture}")
    print(f"ISO: {iso}")

# Step 2: Read RAW data using rawpy
raw = rawpy.imread(raw_file_path)

# Extract RAW data, Bayer Pattern, and bit depth
raw_data = raw.raw_image  # RAW data (H x W)
bayer_pattern = raw.raw_pattern  # Bayer Pattern (e.g., [[0, 1], [3, 2]] for RGGB)
bit_depth = math.log(raw.white_level + 1, 2)   # Bit depth of the RAW data
blc_param = raw.black_level_per_channel[0]  # Black level per channel (R, G1, G2, B)

# Step 3: Apply Black Level Correction (BLC)
blc_params = {
    "R": blc_param,  # Black level offset for R channel
    "G1": blc_param,  # Black level offset for G1 channel
    "G2": blc_param,  # Black level offset for G2 channel
    "B": blc_param   # Black level offset for B channel
}
raw_blc = RawBlc(bayer_pattern, blc_params)
blc_data = raw_blc.process(raw_data)

# Step 4: Convert RAW to RGB
raw_to_rgb = RawToRgb(blc_data, bayer_pattern, demosaic=True, bit_depth=bit_depth)
rgb_data = raw_to_rgb.process()

# Step 5: Convert RGB to XYZ
ccm = np.array([
    [0.4124, 0.3576, 0.1805],
    [0.2126, 0.7152, 0.0722],
    [0.0193, 0.1192, 0.9505]
])
rgb_to_xyz = RgbToXyz(method="greyworld", ccm=ccm)
xyz_data = rgb_to_xyz.process(rgb_data)

# Step 6: Convert XYZ to RGB
xyz_to_rgb = XyzToRgb(method="default", color_space="sRGB")
rgb_data_final = xyz_to_rgb.process(xyz_data)

# Step 7: Save the final RGB image as a JPEG
rgb_to_img = RgbToImg(mode="SDR")
output_path = "output_sdr_default_gamma.jpeg"
rgb_to_img.process(rgb_data_final, output_path)
print(f"SDR image saved to {output_path}")

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for details.