How to save an image with the same properties as a different image

Question

I want to use the properties of a specific image, for example:

t0 = cv2.cvtColor(cv2.imread('original.jpg'), cv2.COLOR_BGR2RGB)

and save a new image with the same properties.

Rephrasing my question:

I want to save 'new_image' with the same dpi, dimensions, color properties (RGB/BGR/GRAYSCALE), etc. as 'original'

original = cv2.cvtColor(cv2.imread('original.jpg'), cv2.COLOR_BGR2RGB)
new_image = cv2.cvtColor(cv2.imread('new.jpg'), cv2.COLOR_BGR2RGB)

How can I do it via opencv or matplotlib?

There are some related ideas here... https://stackoverflow.com/a/57555123/2836621 — Mark Setchell, Feb 08 '21 at 12:42
If you are using JPEG, rather than PNG as per my link above, I guess you could `cv2.imencode()` into memory and then search for the JPEG APP0 marker and replace those bytes with the ones from the original image easily enough. Yes, it is a hack. — Mark Setchell, Feb 08 '21 at 12:55

Mark Setchell · Accepted Answer · 2021-02-08T15:48:08.210

AFAIK, you cannot set the dpi when writing a JPEG with OpenCV, so you could maybe "shell out" to exiftool with Python's subprocess command after writing the image with cv2.imwrite():

exiftool -jfif:Xresolution=301 -jfif:Yresolution=302 result.jpg

As an alternative, this is a rather nasty hack to overwrite the dpi and x-resolution and y-resolution in a JPEG file generated by OpenCV:

#!/usr/bin/env python3

import struct
import numpy as np
import cv2

def writeJPEGwithdpi(im, filename, dpi=(72,72)):
   """Save the image as JPEG with embedded dpi"""

   # Encode as JPEG into memory
   retval, buffer = cv2.imencode(".jpg", im)
   s = bytearray(buffer)

   # APP0 segment looks like this
   # 0xFF, 0xE0,                     // APP0 segment
   # 0x00, 0x10,                     // size of segment, including these 2 bytes; 0x10 = 16 bytes
   # 0x4A, 0x46, 0x49, 0x46, 0x00,   // identifier string: "JFIF"
   # 0x01, 0x01,                     // JFIF version 1.01
   # 0x00,                           // density units (0=no units, 1=dpi)
   # 0x00, 0x01,                     // horizontal density
   # 0x00, 0x01,                     // vertical density
   # 0x00,                           // X thumbnail size
   # 0x00                            // Y thumbnail size

   # Find JFIF marker
   JFIF = s.find(b'JFIF\0')

   # Overwrite units, and x-resolution, and y-resolution 
   s[JFIF+7:JFIF+8]   = b'\x01'                                # density units = 1, i.e. dpi
   s[JFIF+8 :JFIF+10] = (dpi[0]).to_bytes(2, byteorder='big')  # 2 bytes of x-resolution
   s[JFIF+10:JFIF+12] = (dpi[1]).to_bytes(2, byteorder='big')  # 2 bytes of y-resolution

   with open(filename, "wb") as out:
      out.write(s)

################################################################################
# main
################################################################################

# Load sample image
im = cv2.imread('/Users/mark/sample/images/lena.png')

# Save at specific dpi
writeJPEGwithdpi(im, "result.jpg", (77,309))

Check result with exiftool:

exiftool result.jpg

ExifTool Version Number         : 12.00
File Name                       : result.jpg
Directory                       : .
File Size                       : 105 kB
File Modification Date/Time     : 2021:02:08 14:48:52+00:00
File Access Date/Time           : 2021:02:08 14:48:54+00:00
File Inode Change Date/Time     : 2021:02:08 14:48:52+00:00
File Permissions                : rw-r--r--
File Type                       : JPEG
File Type Extension             : jpg
MIME Type                       : image/jpeg
JFIF Version                    : 1.01
Resolution Unit                 : inches      <--- LOOKS GOOD
X Resolution                    : 77          <--- LOOKS GOOD
Y Resolution                    : 309         <--- LOOKS GOOD
Image Width                     : 512
Image Height                    : 512
Encoding Process                : Baseline DCT, Huffman coding
Bits Per Sample                 : 8
Color Components                : 3
Y Cb Cr Sub Sampling            : YCbCr4:2:0 (2 2)
Image Size                      : 512x512
Megapixels                      : 0.262

score 0 · Answer 2 · answered Feb 08 '21 at 11:39

Use imwrite method from opencv.

import cv2

img='original.jpg'
t0 = cv2.cvtColor(cv2.imread(img), cv2.COLOR_BGR2RGB)

# To save your processed image
cv2.imwrite('img_t0.jpg',t0)

#To save your original file with save properties 
cv2.imwrite('new_original.jpg',img)

How to save an image with the same properties as a different image

2 Answers2

Linked