Reorder matching indinces in openCV + python

Question

I followed this code that works (see the question at the end)

import numpy as np
import cv2

img1 = cv2.imread('rectif/Camera1.tif',-1)
img2 = cv2.imread('rectif/Camera2.tif',-1)
img3 = cv2.imread('rectif/Camera3.tif',-1)

gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
gray2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
gray3 = cv2.cvtColor(img3, cv2.COLOR_BGR2GRAY)

orb = cv2.ORB(1000, 1.2)

#extract features
(kp1,des1) = orb.detectAndCompute(gray1, None)
(kp2,des2) = orb.detectAndCompute(gray2, None)
(kp3,des3) = orb.detectAndCompute(gray3, None)

# Create matcher
bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)

# Do matching
matches = bf.match(des1,des2)
matches2 = bf.match(des1,des3)

# Sort the matches based on distance.  Least distance
# is better
matches = sorted(matches, key=lambda val: val.distance)
matches2 = sorted(matches2, key=lambda val: val.distance)

But I do not have complete consciousness of what matches python structure is. I learned that I can extract the X,Y coordinates of feature by

lkp1 = [kp1[mat.queryIdx].pt for mat in matches] #X,Y of features img1 
lkp2 = [kp2[mat.trainIdx].pt for mat in matches] #X,Y of features img2

but I am definitely not an expert. I am wondering if there is a chance to order the matches, and extract the feature (their indices in matches structure) that are in common between img1 and img2 and img 3. Better, I would like to obtain 3 list of x,y coordinates, if they have the same index, they represent the same feature in image.

The ideal result (not so simple of course) would be something like

wannabe_matches = bf.match(des1,des2,des3)

There is a way to do something like this (PSEUDOCODE: this is not a working code!)

matches = bf.match(des1,des2)
matches2 = bf.match(des1,des3)

for mat in matches:
   for mat2 in matches2:

    if mat.descriptor == mat2.descriptor: #how it can be done?
      img1_idx = mat.queryIdx
      img2_idx = mat.trainIdx
      img3_idx = mat2.trainIdx

      (x1,y1) = kp1[img1_idx].pt
      (x2,y2) = kp2[img2_idx].pt
      (x3,y3) = kp2[img3_idx].pt

      # Append to each list
      list_kp1.append((x1, y1))
      list_kp2.append((x2, y2))
      list_kp3.append((x3, y3))

Or anything better that lead me to the result?

I am working with python 2.7 and openCV 2.4.13 (not openCV 3!)