HDR - High Dynamic Ranging

Prerequisites

• Python 3
• Python packages: OpenCV, numpy, matplotlib.pyplot, scipy, os, time
• MAC OS or Windows 10 (Ubuntu Windows' Bash)

Installation

• Clone the repo:

git clone https://github.com/msouppe/HDR-Computer-Vision.git

• Navigate to main.py

• Run the program:

python3 main.py 

Project Structure

• main.py: Main program to ob
• gamma.py: Obtain the gamma value to raise the brightness to the power of gamma
• hdr.py: Creates composite images
• util/analyzer.py: Plots all different graphs and histogram
• util/image_process.py: Preprocessing images such as obtaining region of interest, splitting channels, merging channels, and calculating the average brightness of an
• output/: All of the output graphs and images
• hdr/: Selected three images to create HDR images
• hdr_fullstack/: All images to create HDR images

Output

Part 1
For every color channel; blue, green, and red, we plot the graphs for the following:

• Exposure Time vs Brightness
• Log of Exposure Time vs Log of Brightness with Regression
• Exposure Time vs Brightness'G

Part 2
For every color channel; blue, green, and red, we plot the graphs for the following:

• Histogram B'g (a0 * T)
• Histogram B'g (a1 * T)
• Histogram B'g (a2 * T)
• Histogram B'g (a1 * T) / a1
• Histogram B'g (a2 * T) / a2

Part 3
For every color channel; blue, green, and red, we plot the graphs the composite images for two different algorithms:

• HDR1 Histogram
  • Algorithm 1: Takes the best non-saturated pixels
• HDR2 Histogram
  • Algorithm 2: Takes the average non-saturated pixels per color channel across the three images

Part 4
Tone-mapped composite HDR final image

• HDR1 image using algorithm 1
• HDR2 image using algorithm 2