reference resources: http://stackoverflow.com/questions/13572448/change-values-in-a-numpy-array

###################################################################3

Get the video stream from the raspberry pie camera and convert it to opencv format:

http://blog.csdn.net/u012005313/article/details/51482994

At the same time, you want to operate on each frame, but there is an error:

ValueError：assignment destination is read-only

Images cannot be manipulated because they are read-only.

Find a way on stackhover: because in Python, the opencv image format is Numpy.ndarray You can modify the properties of ndarray by:

img.flags.writeable = True

####################################################################

Numpy.ndarray It has the following attributes:

C_ CONTIGUOUS（c_ contiguous）

F_ CONTIGUOUS（f_ contiguous）

OWNDATA（owndata）

WRITEABLE（writeable）

ALIGNED（aligned）

UPDATEIFCOPY（updateifcopy）

import numpy as np
help(np.ndarray.flags)

The flags property is information about the array memory layout

Among them, the flags attribute can be modified in the form of a dictionary, for example:

a.flags['WRITEABLE'] = True

You can also use lowercase attribute names:

a.flags.writeable = True

Abbreviations (‘c ‘/’f’, etc.) are only used in dictionary form

Only the attributes updateifcopy, writeable and aligned can be modified by users in three ways

1. Direct assignment:

a.flags.writeable = True

2. Dictionary input:

a.flags['WRITEABLE'] = True

3. Use function ndarray.setflags ：

help(np.ndarray.setflags)