在OpenCV Python中画线时出错

Question

So what I am trying to do is use hough transformations to find lines on an image and then draw the lines into the image.

But I am getting an error that says:

Traceback (most recent call last): 
    File "houghLines.py", line 31, in <module>
    main() File "houghLines.py", line 13, in main 
    a = hough(image)
    File "houghLines.py", line 29, in hough
    (0,0,255))
    TypeError: only length-1 arrays can be converted to Python scalars

This is my code:

import cv2
import os
import math
def main():

    directory = os.path.dirname(__file__)
    picLoc = os.path.join(directory, "../video-image/1m50s.png")

    image = cv2.imread(picLoc)
    print "sending image to houghLines.py"

    a = hough(image)
    cv2.imshow("", a)
    cv2.waitKey(0)
    cv2.destroyAllChildren()

def hough(image):
    canny = cv2.Canny(image, 50, 200)
    color_image = cv2.cvtColor(canny, cv2.COLOR_GRAY2BGR)
    houghLines = cv2.HoughLinesP(canny, 1, math.pi/180, 50)
    for x in range(len(houghLines)):
        print x
        pt1 = (houghLines[x][0], houghLines[x][1])
        pt2 = (houghLines[x][2], houghLines[x][3])
        cv2.line(color_image, pt1, pt2, (0,0,255), 3)
    return color_image

main()

Answer 1

This problem occurs because cv2.line() expects pt1 and pt2 to be tuples consisting of only a single element for each x and y coordinate. If you print pt1 in your example, you will quickly see that this is not the case.

cv2.HoughLinesP() returns a numpy array with only one element. This element contains lists of four points which are the start and end points of your lines. Knowing this, a correct implementation follows:

    for line in houghLines[0]:
        pt1 = tuple(line[:2])
        print pt1
        pt2 = tuple(line[-2:])
        cv2.line(color_image, pt1, pt2, (0,0,255), 3)