OpenCV, red ball detection

[4] We are trying to detect red balls on the Nao. First we did it on our computer, here is the code:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
#!/usr/bin/env python 
import cv2
import numpy as np

#cap = cv2.VideoCapture(0)
c = cv2.VideoCapture(0)
detected = False
while(detected is not True):

    # Take each frame
    _, img = c.read()
    img = cv2.flip(img,5)

    # Convert BGR to HSV
    hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)

    #cv2.imshow("b", hsv)
    # define range of blue color in HSV
    lower_color = np.array([110,50,50], dtype=np.uint8)
    upper_color = np.array([130,255,255], dtype=np.uint8)

    # Threshold the HSV image to get only blue colors
    mask = cv2.inRange(hsv, lower_color, upper_color)
    #cv2.imshow('mask',mask)  
    # Bitwise-AND mask and original image
    res = cv2.bitwise_and(img,img, mask= mask)
    #cv2.imshow("b", res)

    imgray = cv2.cvtColor(res,cv2.COLOR_BGR2GRAY)
    ret,thresh = cv2.threshold(imgray,127,255,0)
    contours, hierarchy = cv2.findContours(imgray,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
    

    #erode = cv2.erode(res,None,iterations = 3)
    #dilate = cv2.dilate(erode,None,iterations = 10)
    #contours,hierarchy = cv2.findContours(res,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)
    for cnt in contours:

        x,y,w,h = cv2.boundingRect(cnt)
        print w,h
        if(w < 60 or h < 60):
            continue

        cx,cy = x+w/2, y+h/2

        if 20 < res.item(cy,cx,0) < 30:
             #cv2.rectangle(f,(x,y),(x+w,y+h),[0,255,255],2)
            print "yellow :", x,y,w,h
        elif 100 < res.item(cy,cx,0) < 120:
             cv2.rectangle(img,(x,y),(x+w,y+h),[255,0,0],2)
             print "blue :", x,y,w,h
             detected = True
        cv2.imshow('img',img)  

    k = cv2.waitKey(5) & 0xFF
    if k == 27:
        break

cv2.destroyAllWindows()

On the computer, this code work fine, but on the Nao, it seems that the image processing is too slow. Our project manager contacted Aldebaran to know more about this issue.
They recommended us to use Nao library to import the camera image insted of using cv2 library.

We are know trying to adapt our code to these guidelines to make all that work.