我正在使用鱼眼镜头拍摄深度图像。我曾尝试使用StereoSGBM（）方法获得深度图，但没有给出理想的结果，但现在我正在尝试使用StereoBM（）方法。
stereoMatcher = cv2.StereoBM() preset = 2 ndisparities = 32 WinSize = 5 if not left.grab() or not right.grab(): print("No more frames") break _, leftFrame = left.retrieve() #leftFrame = cropHorizontal(leftFrame) leftHeight, leftWidth = leftFrame.shape[:2] _, rightFrame = right.retrieve() #rightFrame = cropHorizontal(rightFrame) rightHeight, rightWidth = rightFrame.shape[:2] if (leftWidth, leftHeight) != imageSize: print("Left camera has different size than the calibration data") break if (rightWidth, rightHeight) != imageSize: print("Right camera has different size than the calibration data") break fixedLeft = cv2.remap(leftFrame, leftMapX, leftMapY, REMAP_INTERPOLATION) fixedRight = cv2.remap(rightFrame, rightMapX, rightMapY, REMAP_INTERPOLATION) grayLeft = cv2.cvtColor(fixedLeft, cv2.COLOR_BGR2GRAY) grayRight = cv2.cvtColor(fixedRight, cv2.COLOR_BGR2GRAY)
计算时，StereoBM（）的函数还需要第三个参数作为视差。我使用
进行了计算grayLeft = cv2.cvtColor(fixedLeft, cv2.COLOR_BGR2GRAY) grayRight = cv2.cvtColor(fixedRight, cv2.COLOR_BGR2GRAY) disparity = stereoMatcher.compute(grayLeft, grayRight) depth = stereoMatcher.compute(grayLeft, grayRight) #cv2.imshow('left', fixedLeft) #cv2.imshow('right', fixedRight) cv2.imshow('depth', depth / DEPTH_VISUALIZATION_SCALE) if cv2.waitKey(1) & 0xFF == ord('q'): break left.release() right.release() cv2.destroyAllWindows()
但是它抛出了一个错误：
Traceback (most recent call last):
File "fisheye_depth.py", line 104, in <module>
disparity = stereoMatcher.compute(grayLeft, grayRight)
TypeError: Incorrect type of self (must be 'StereoMatcher' or its derivative)
如果有人对此有任何想法，那我在做什么错。让我知道我在计算出算术operator（）的差异时犯了什么错误。