SunFounder picar-x
(continued from previous page)
img = frame.array
img,img_2,img_3 = color_detect(img,'red') # Color detection function
cv2.imshow("video", img) # OpenCV image show
cv2.imshow("mask", img_2) # OpenCV image show
cv2.imshow("morphologyEx_img", img_3) # OpenCV image show
rawCapture.truncate(0) # Release cache
k = cv2.waitKey(1) & 0xFF
# 27 is the ESC key, which means that if you press the ESC key to exit
if k == 27:
break
print('quit ...')
cv2.destroyAllWindows()
camera.close()
How it works?
First, the range of H in the HSV color space is defined as a dictionary, which is convenient for the following color
judgment algorithm:
color_dict = {'red':[0,4],'orange':[5,18],'yellow':[22,37],'green':[42,85],'blue':[92,
˓→110],'purple':[115,165],'red_2':[165,180]}
Then, a convolution kernel of size 5x5 is defined, which will be used for morphological operations, like filtering.
kernel_5 = np.ones((5,5),np.uint8)
Next, the color_detect() function will processes pictures in four steps:
1. Extract the data of the target color as a new binary image (array).
2. Performs advanced morphological transformations.
3. Finds contours in a binary image.
4. Draws a frame for the recognized object on the image.
def color_detect(img,color_name):
# The blue range will be different under different lighting conditions and can be
˓→adjusted flexibly. H: chroma, S: saturation v: lightness
resize_img = cv2.resize(img, (160,120), interpolation=cv2.INTER_LINEAR) # In
˓→order to reduce the amount of calculation, the size of the picture is reduced to
˓→(160,120)
hsv = cv2.cvtColor(resize_img, cv2.COLOR_BGR2HSV) # Convert from BGR
˓→to HSV
color_type = color_name
mask = cv2.inRange(hsv,np.array([min(color_dict[color_type]), 60, 60]), np.
˓→array([max(color_dict[color_type]), 255, 255]) ) # inRange()Make the ones
˓→between lower/upper white, and the rest black
if color_type == 'red':
mask_2 = cv2.inRange(hsv, (color_dict['red_2'][0],0,0), (color_dict['red_2
˓→'][1],255,255))
mask = cv2.bitwise_or(mask, mask_2)
morphologyEx_img = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel_5,iterations=1)
˓→ # Perform an open operation on the image
(continues on next page)
66 Chapter 4. Play with Python
To Next Page
Loading...
