Learning Opencv_1: Interpretation of channels and grayscale

On the interpretation of the channel:

1, the more understandable explanation is: gray-scale channel number of 1, color map of the channel is 3. Basically, the description of a pixel, if it is grayscale, then only need a number to describe it, is a single channel. If a pixel, there are RGB three colors to describe it, is three channels.

2, 4 channels are usually rgba and may be used in some processing. 2-channel images are uncommon, usually used in program processing, such as Fourier transform, may be used, a channel is real, a channel for imaginary numbers, mainly programming convenience. There is also a situation is 16-bit image, is 3 channels, but in order to reduce the amount of data, compressed to 16 bits, just two channels, common format has RGB555 or RGB565, that is, R accounted for 5, G accounted for 5 or 6, B accounted for 5, there are RGBA5551 format. Old format, no need to worry about.

3, mainly some cameras often use some more "odd" format, there is no way. Add a situation, the current common some of the cameras like the use of YUV2 format, the format below Yuyv, in the process can be used 4 channels or 2 channels to deal with. If the original format is: Y1UY2V, the interpolation becomes y1uv,y2uv into two colored dots. YCRCB also has similar compression.

Interpretation of Grayscale:

Description of a pixel, if it is grayscale, then only need a number to describe it, is a single channel. If a pixel, there are RGB three colors to describe it, is the three channels
Windows BMP is sometimes a four-channel image, R, G, B plus a channel, representing transparency (vblittleboy)