Tags: python3 opencv video
Rating:
Welcome to SEETF! Submit the teaser video flag here.
https://www.youtube.com/watch?v=0GVC30jiwJs
Download: https://drive.google.com/file/d/1Dlfp6vC1pto7vW5WUKqvX-tbIt6tfGfg/view
## Got the flag
The welcome video uploaded to YouTube even before the CTF start. After watched the video, we noticed from 00:25, there are white dot in a square box on the top right of the video till the end. The square seems nicely fit a QRCode, so the task would be get all the white dot to form a QRcode picture.
Extracts frames from video downloaded from Google drive link with ffmpeg
`ffmpeg -i input.mp4 '%04d.png'`
this creates ~ 4000 png files and the dot starts appears at ~ 1620.png.
The idea is to use opencv to crop each image in a big-enough square, as each of them is a numpy matrix, they can be added together, in the end, we will have all the white-dots and that is the QRcode.
The naive implementation returned a noise image:
```
import os
import cv2
images = sorted(os.listdir())[1620:]
acc = cv2.imread(images[0], cv2.IMREAD_GRAYSCALE)[:256,1620:]
for img in images:
tmp = cv2.imread(x, cv2.IMREAD_GRAYSCALE)[:256,1620:]
acc += tmp
cv2.imwrite("result.png", acc)
```
After examines several images, we found that black areas is not absolute black (value 0), so all the gray pixels add up (1, 2...), we end up with a white pixel (since we exams ~ 2400 images). From 00:31 to 00:35, there are even pretty white background video. We convert all values < 200 to 0, then remain only real white dots. Add up, after ~ 1 min, we got the QRcode, scanned it, got the flag.
```
import os
import cv2
images = sorted(os.listdir("welcome"))[1620:]
i = cv2.imread("welcome/4391.png", cv2.IMREAD_GRAYSCALE)
i = i[:256, 1620:]
i[i < 200] = 0
for x in images:
print(x, end=" ")
tmp = cv2.imread("welcome/" + x, cv2.IMREAD_GRAYSCALE)[:256, 1620:]
tmp[tmp < 200] = 0
i += tmp
cv2.imwrite("welcome.png", i)
```
