DDA4310 Assignment 01 Solution

Implement Image Processing Pipeline with rawpy
Download Data: data.rar
Part 0- Setting up your machine and environment

This assignment is required to use of python, define your python file name as main.py.
1. For this assignment, we require you to use the latest conda to manage your developing environment. It will help you to achieve a great experience for coding.
2. Install rawpy · PyPI and opencv-python · PyPI in conda.

3. Implement the given blank functions using "rawpy". Read online document Params class rawpy 0.18.0a1 documentation
classrawpy.Params(self, demosaic_algorithm=None, half_size=False, four_color_rgb=False, dcb_iterations=0, dcb_enhance=False, fbdd_noise_reduction=FBDDNoiseReductionMode.Off, noise_thr=None,
median_filter_passes=0, use_camera_wb=False, use_auto_wb=False, user_wb=None, output_color=ColorSpace.sRGB, output_bps=8, user_flip=None, user_black=None, user_sat=None, no_auto_bright=False, auto_bright_thr=None, adjust_maximum_thr=0.75, bright=1.0, highlight_mode=HighlightMode.Clip, exp_shift=None, exp_preserve_highlights=0.0, no_auto_scale=False, gamma=None, chromatic_aberration=None, bad_pixels_path=None)¶
Try your best to implement a better visual-looking result with a faster speed and less memory consumption.
Part 1 - Implement ISP using rawpy

Note: Here you do not need to fill the corresponding functions, just comments the functions and implement this steps using rawpy's functions.
BTW, pay attention to the metadata, it's just the example metadata and you need to read and replace the metadata using rawpy instead of directly using the given one.
Step 1. Bad Pixel Correction (10pts)
Implement it using rawpy, and save the corresponding result using opencv in .jpg format. Show and explain your result in your report.
Step 2.'Black Level Compensation' (10pts)
Implement it using rawpy, and save the corresponding result using opencv in .jpg format. Show and explain your result in your report.
user_black (int) – custom black level
Example: read RawPy
XX.postprocess(param=rawpy.Paramers) Disable all other unrelated parameters and **only enable** parameters this and previous steps. You can use the same way to implement the following steps.
Step 3.'lens shading correction (Skipped)
Step 4. Anti Aliasing Filter (Skipped)
### Step 5. Auto White Balance and Gain Control (10pts)
Implement it using rawpy, and save the corresponding result using opencv in .jpg format. Show and explain your result in your report.
use_auto_wb (bool) – whether to try automatically calculating the white balance
Step 6. Test the following functions (10pts)
no_auto_scale (bool) – Whether to disable pixel value scaling no_auto_bright (bool) – whether to disable automatic increase of brightness Implement it using rawpy, and save the corresponding result using opencv in .jpg format. Show and explain your result in your report. Test different algorithms and compare the results.
Step 7. Color Space Conversion (10pts)
For now, you have got the no-auto-scale and no_auto_bright RGB image.
Implement color space conversion using opencv, and save the corresponding result using opencv in .jpg format. Show and explain your result in your report.

Step 8. Test Gamma Correction (10pts)
gamma (tuple) – pair (power,slope), default is (2.222, 4.5) for rec. BT.709
Also do the gamma directly in RGB space using BT.709's standard. Compare the difference of your results and the gamma provided by rawpy.
Submission

To Grade submission for this homework. It would help if you tried committing and tagging your code as specified in the given code structure. We highly recommend that you finish the homework to prevent any issues when you start on the following projects for this class.
Your grades = correct code (70%) + good report (30%)
Code submission （60%）
The code is required with a simple run, and then the TAs can see the results of each steps (name you saved results as partxx-stepxx.jpg)
Report （40%）
For each function you implemented, show the results crops , describe what you have done, and explain your results.
Write your report using Markdown (like MarkText, a free markdown editor) and export it to PDF format. Markdown is easy to use, just learn to use.