Hi everyone! I need help transferring an image via TCP from a python program on 
my raspberry pi to an android application.

I have set up a client-server architecture such that my raspberry pi 3 records 
audio, performs some analysis on it, and then sends the data (via TCP) to the 
android app to display on the app screen. The recording and analysis is done 
and I am able to make the connection and transfer string data that displays on 
the app with no problem. However, I have been unsuccessful in transferring an 
image from rpi to android app. So basically, the image is stored on the rpi and 
I an attempting to transfer the image to the app to display it. I have been 
working on this for over a week with no luck so any help would be greatly 
appreciated!

My current implementation (code snippets provided below):

On rpi (python): Like I said, sending strings and displaying them on the 
android app is done without any problem. When I am sending the image portion of 
the audio analysis, I send a string first that says "?start" so that the 
android side knows that an image instead of a string is about to be sent (and 
will wait to update the GUI until it receives the entire image). Then, I open 
the image stored on rpi and read the entire image as a byte array (typically 
about 40-50k bytes). I get the length of the byte array and send that as a 
string to android app. Finally, I send the byte array to the android and it 
waits for an OK message from the app. All of this works without reporting any 
errors.

On android app (java): When the app receives the "?start" string, it then uses 
a Buffered Reader (which is what I used to read the string data I had 
transferred to the app successfully earlier) to read the size of the image byte 
array. Then, I create 2 buffers, msg_buff and img_buff. msg_buff will read in 
1024 bytes at a time while img_buff will hold the entire byte array of the 
image. In the infinite while loop, I have a DataInputStream, called in, read 
bytes into msg_buff and returns the number of bytes read. Then, I concatenate 
the copied contents of msg_buff into img_buff. Once the bytes read from in is 
-1 or the img_offset (which is just the total number of bytes read) is greater 
than or equal to the size of the image bytes array, the while loop is broken. 
Then, I would attempt to save the image to android internal storage and then 
load it later to an imageView to display it. This code does successfully read 
in the bytes until there are around 2000-3000 bytes left to be read and
  then it seems to freeze on the int bytes_read = in.read(msg_buff, 0, 
msg_buff.length) line. I have not been able to get past that point so I do not 
know if saving the image to internal storage and then loading it to imageview 
that way will work either.

I have also tried using base64 encoding/decoding but that also kept producing 
errors. I have tried rpi only sending 1024 bytes of the image at a time but 
that also did not work. I have tried several implementations of this approach 
but nothing has worked so far. If anyone sees anything wrong or has another 
approach, I am all ears!

Android Studio (app side):

//receives the message which the server sends back 
  InputStream sin = socket.getInputStream(); 
  OutputStream sout = socket.getOutputStream(); 
  DataInputStream in = new DataInputStream(sin);  
  mBufferIn = new BufferedReader(new                                            
          InputStreamReader(socket.getInputStream()));   

//in this while the client listens for the messages sent by the server 
while (mRun) {  
     mServerMessage = mBufferIn.readLine();  
     if (mServerMessage != null && mMessageListener != null) {  
          //Check if data is image 
          if(mServerMessage.equals("?start")) {  
               // Get length of image byte array 
               int size = Integer.parseInt(mBufferIn.readLine());   
         
               // Create buffers 
               byte[] msg_buff = new byte[1024]; 
               byte[] img_buff = new byte[size];
               int img_offset = 0; 

               while(true){ 
                    int bytes_read = in.read(msg_buff, 0, msg_buff.length);  
                    if(bytes_read == -1){ break; }  

                    //copy bytes into img_buff 
                    System.arraycopy(msg_buff, 0, img_buff, img_offset, 
bytes_read); 
                    img_offset += bytes_read;       
                    if( img_offset >= size) { break; }             
               }  
               ContextWrapper cw = new 
ContextWrapper(ApplicationContextProvider.getContext());      
               File directory = cw.getDir("imageDir", Context.MODE_PRIVATE); 
               File mypath = new File(directory, "signal.jpeg");  
               Bitmap bitmap = BitmapFactory.decodeByteArray(img_buff, 0, 
img_buff.length); 
               FileOutputStream fos = null;  
               try{      
                    fos = new FileOutputStream(mypath);                 
                    //Use compress method on Bitmap object to write image to 
OutputStream                
                    bitmap.compress(Bitmap.CompressFormat.JPEG, 100, fos);      
 
                    //Send OK byte[]           
                    OK = new byte[] {0x4F, 0x4B};      
                    sout.write(OK); 
                }   catch (Exception e) { 
                    e.printStackTrace(); 
               }                                             
     }
}

Raspberry Pi (python):

def image_to_byte_array(image_file, conn): 
     with open(image_file, 'rb') as imageFile: 
          content = imageFile.read() 
          conn.sendall("?start\n".encode('utf-8')) 
          size = len(content) strSize = str(size) + "\n" 
          conn.sendall(strSize.encode('utf-8')) 
          conn.sendall(content)

Note that conn is the connection between the app and the rpi and the images are 
PNG.

If anyone knows why this isn't working or has a better way for me to do this, I 
would greatly appreciate it!! Thank you in advance!!
-- 
https://mail.python.org/mailman/listinfo/python-list

Reply via email to