There is a various range of type in openCV with a specific name representing a code. Here is the explanation. Each of them are composed as follow : Type + Tuple Size + Encoding
Ex: Vec3d = Vector of tuple of 3 elements encoded as double
| Letter | Encoding |
|---|---|
| b | uchar |
| d | double |
| f | float |
| i | int |
| s | short |
| w | ushort |
| I | int64 |
We can find data type for image encoding. There are splited in several parts as follow : CV_ + size + Sign + C + n channels
| Letter | Sign |
|---|---|
| U | Unsigned |
| S | Signed |
| F | Floating |
Ex: CV_8UC3 = Matrix of 3 channels encoded on 8 bit, unsigned
The size are limited to these formats : 8, 16, 32 and 64
The number of channel is unlimited however a number between 1 to 4 (included) is preferred.