OpenCV-Python StereoBM error, thanks

Python
,
1

I need to derive disparity map from two images. I am using Python3.7 on Win10, PyCharm, and opencv-python=4.5.1.48. The following is my code.

import cv2

image_file1 = r'cam1.jpg'
image_file2 = r'cam2.jpg'
image_left = cv2.imread(image_file1, 0)
image_right = cv2.imread(image_file2, 0)

image_left = cv2.resize(image_left, (0,0), fx=0.25, fy=0.25)
image_right = cv2.resize(image_right, (0,0), fx=0.25, fy=0.25)

stereo = cv2.StereoBM(numDisparities=16, blockSize=15)
disparity = stereo.compute(image_left,image_right)

However, I got no result, and only the information as following:
Process finished with exit code -1073741819 (0xC0000005).

No error information, and no exception information. Can anybody help me on this? Thank you very much.

2

The problem is from the last line, i.e., stereo.compute.

3

https://docs.opencv.org/master/d9/dba/classcv_1_1StereoBM.html#a04fdf00525f82fe708d556e2cd359004

meaning you need to call StereoBM_create or StereoBM.create

when scaling images down by a power of two, consider using pyrDown (repeatedly if needed). that function applies a lowpass so there won’t be any sampling artefacts. if you stick with cv.resize, consider passing the INTER_AREA flag.

1 Like
4

python sample for stereo image matching and point cloud generation

1 Like
5

Thank you very much to all of you.