当我尝试合并零通道以创建具有单个通道的图像时出错

Question

我正在尝试读取图像，然后在输出图像中仅保留red通道。这是我所做的：

color = cv2.imread("lohri.jpg")
b,g,r = cv2.split(color)
zeros = np.zeros([b.shape[0], b.shape[1]])
# Make other 2 channels as zeros
only_red = cv2.merge((zeros, zeros, r))

但是当我这样做时，我得到一个错误说：

OpenCV(3.4.1) Error: Assertion failed (mv[i].size == mv[0].size && mv[i].depth() == depth) in merge, file /io/opencv/modules/core/src/merge.cpp, line 458
Traceback (most recent call last):
File "inspect.py", line 23, in <module>
  only_red = cv2.merge((zeros, zeros, r))
  cv2.error: OpenCV(3.4.1) /io/opencv/modules/core/src/merge.cpp:458: 
  error: (-215) mv[i].size == mv[0].size && mv[i].depth() == depth in 
  function merge

我不明白这其中的原因。为什么我会收到此错误？

Answer 1

这是因为虽然您的零形状是正确的，但很可能存在数据类型不匹配。Numpy Zeros文档的默认数据类型为numpy.float64。只需为您的案例传递 dtype 参数即可。

color = cv2.imread("lohri.jpg")
b,g,r = cv2.split(color)
zeros = np.zeros(b.shape[0], b.shape[1]], dtype = b.dtype)
#Note that the dtype is most likely np.uint8, and you can use that instead too if you prefer
#zeros = np.zeros([b.shape[0], b.shape[1]], dtype = np.uint8)
#Also note that you can pass shape tuple directly.
#zeros = np.zeros(b.shape, dtype = np.uint8)
# Make other 2 channels as zeros
only_red = cv2.merge((zeros, zeros, r))

编辑：您还可以使用np.zeros_like使用正确的形状和数据类型创建数组，这也使代码更加干净简洁。谢谢马克！

color = cv2.imread("lohri.jpg")
b,g,r = cv2.split(color)
zeros = np.zeros_like(b)
only_red = cv2.merge((zeros, zeros, r))