How can I convert a array of integers of a JPG image to a JPG image?

Question

I have a camera on a robot set up. Long story short, the camera outputs the binary array, which I believe is a JPG encoded stream. I say this since a lot of the methods in the library I'm using give hint to it being JPG-encoded, and the first and last 2 bytes are 255 216 and 255 217 respectively, which are the magic numbers of the JPG file format.

How can I convert this array of integers to a JPG file in Python?

I tried writing the array of integers to a .JPG file, but the file is labeled as corrupt.

Long story short, I have an ArduCam hooked to a esp8266 (literally an Arduino board that can connect to WiFi). I have a script that captures a photo, and uploads the raw image data to a server hosted on the esp8266. Here is the main method I am using:

void camCapture(ArduCAM myCAM) {

 WiFiClient client = server.client(); // ignore this stuff
 uint32_t len  = myCAM.read_fifo_length();
 if (len >= MAX_FIFO_SIZE) //8M
 {
   Serial.println(F("Over size."));
 }
 if (len == 0 ) //0 kb
 {
   Serial.println(F("Size is 0."));
 }
 myCAM.CS_LOW();
 myCAM.set_fifo_burst();
 if (!client.connected()) return;
 String response = "[";
 i = 0;

 while ( len-- ) // this is where the raw image is collection
 {
   temp_last = temp;
   temp =  SPI.transfer(0x00);
   //Read JPEG data from FIFO
   if ( (temp == 0xD9) && (temp_last == 0xFF) ) //If find the end ,break while,
   {
     buffer[i++] = temp;  //save the last  0XD9
     response += String(int(temp)) + "], ";
     response += String(i);
     if (!client.connected()) break;
     Serial.print(String(i));
     server.send(200,"text/plain",response); // this is where the image is uploaded
     is_header = false;
     i = 0;
     myCAM.CS_HIGH();
     break;
   }
   if (is_header == true)
   {
     //Write image data to buffer if not full
     if (i < bufferSize)
       buffer[i++] = temp;
     else
     {
       //Write bufferSize bytes image data to file
       if (!client.connected()) break;
       i = 0;
       buffer[i++] = temp;
     }
     response += String(int(temp)) + ",";
   }
   else if ((temp == 0xD8) & (temp_last == 0xFF))
   {
     is_header = true;
     buffer[i++] = temp_last;
     buffer[i++] = temp;
     response += String(int(temp_last)) + "," + String(int(temp)) + ",";
   }
 }
}

The raw image data looks like this:

255,216,255,224,0,16,74,70,73,70,0,1,1,1,0,0,0,0,0,0,255,219,74,74,74,74,74,74,74,74,74,74,74,74,74,74 [...] 64,143,255,217

Now I know jpeg files start and end with 255,216 and 255,217 respectively. So I see this as a good sign. Furthermore, when I use the jpeg html code, it actually converts the raw image data into an actual image.

Here is my python script where I attempt to decode the raw image data into a .jpg file:

import cv2 

with open('img.txt') as file: # this is where I keep the raw image data
    img = file.read()


img = img.split(',')
img = [int(i) for i in img]
tmpfile = open("tmp.jpg", "wb")
for i in img:
    tmpfile.write(bytes(i))
tmpfile.close()

img = cv2.imread('tmp.jpg')
print(img)
cv2.imshow('img',img)
cv2.waitKey(0)

Answer 1

You are dumping the string representation of the values (i.e. "2", "5", and "5" instead of 0xFF) into the file since bytes() will convert your int to a byte string. You could have noticed that from the file size as it won't match the number of values in tmp.txt.

The correct way to write bytes to a file is as follows:

import struct

...
for i in img:
    tmpfile.write(struct.pack("B", i))
...

Since you are working with 8-bit values in this case, you can also use chr(i).