在OpenCV中使用Python中的findContours

Question

I'm using OpenCV on the raspberry pi and building with Python. Trying to make a simple object tracker that uses color to find the object by thresholding the image and finding the contours to locate the centroid. When I use the following code:

image=frame.array
imgThresholded=cv2.inRange(image,lower,upper)    
_,contours,_=cv2.findContours(imgThresholded,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
cnt=contours[0]
Moments = cv2.moments(cnt)
Area = cv2.contourArea(cnt)

I get the following error.

Traceback (most recent call last):
 File "realtime.py", line 122, in <module>
  cnt=contours[0]
IndexError: list index out of range

I've tried a few other settings and get the same error or

ValueError: too many values to unpack

I'm using the PiCamera. Any suggestions for getting centroid position?

Thanks

Z

Answer 1

Error 1:

Traceback (most recent call last):
 File "realtime.py", line 122, in <module>
  cnt=contours[0]
IndexError: list index out of range

Simply stands that the cv2.findContours() method didn't found any contours in the given image, so it is always suggested to do a sanity checking before accessing the contour, as:

if len(contours) > 0:
    # Processing here.
else:
    print "Sorry No contour Found."

Error2

ValueError: too many values to unpack

This error is raised due to _,contours,_ = cv2.findContours , since the cv2.findContours returns only 2 values, contours and hierarchy, So obviously when you try to unpack 3 values from 2 element tuple returned by the cv2.findContours , it would raise the above mentioned error.

Also the cv2.findContours changes the input mat in place, so it is suggested to call the cv2.findContours as:

contours, hierarchy = cv2.findContours(imgThresholded.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
if len(contours) > 0:
    # Processing here.
else:
    print "Sorry No contour Found."