Types of Image Data

The type of the image defines the location and availability of data accessible over the three address pointers, st, ccmp1 and ccmp2. ccmp1 and ccmp2 are the start addresses of the additional color components. The following table gives an overview of the various image types and color components. More...

Collaboration diagram for Types of Image Data:

Macros
#define	IMAGE_GREY   (0)
	8 Bit Values per Pixel (0-255), One Channel (st).
#define	IMAGE_BAYER   (1)
	8 Bit Values per Pixel (0-255), One Channel (st), Patterned (RGRG..GBGB..RGRG...).
#define	IMAGE_RGB   (2)
	8 Bit Values per Pixel (0-255), Three Channels (st: R, ccmp1: G, ccmp2: B).
#define	IMAGE_CBCR444   (3)
	8 Bit Values per Pixel (0-255), Three Channels (st: Y, ccmp1: U, ccmp2: V).
#define	IMAGE_CBCR411   (4)
	8 Bit Values per Pixel (0-255), Three Channels (st: Y, ccmp1: U, ccmp2: V); U and V Channels have dx, dy and pitch divided by 2, which means that each 2×2 Y pixel block share the same U and V pixel.
#define	IMAGE_YUVNORM   (5)
	8 Bit Values per Pixel (0-255), Three Channels (st: Y, ccmp1: U*, ccmp2: V*).
#define	IMAGE_IHS   (6)
	8 Bit Values per Pixel (0-255), Three Channels (st: I, ccmp1: H, ccmp2: S).
#define	IMAGE_RGBO   (7)
	Only used Internally.
#define	IMAGE_VECTOR   (8)
	8 Bit Values per Pixel (0-255), Two Channels (st: Magnitude, ccmp1: Direction (f. ex. gradient_2x2())).
#define	IMAGE_GREY16   (9)
	16 Bit Values per Pixel (0-65535), One Channel (st: Always cast to (U16*)).
#define	IMAGE_GREY32   (10)
	32 Bit Values per Pixel (0-232), One Channel (st: Always cast to (U32*)).
#define	IMAGE_CMPLX16   (11)
	16 Bit Values per Pixel (0-65535), Two Interleaving Channels (st: [Real1, Imaginary1, Real2, ...]; Always cast to (I16*)), See vc_fft().
#define	IMAGE_CBCR422   (13)
	8 Bit Values per Pixel (0-255), Three Channels (st: Y, ccmp1: U, ccmp2: V); U and V Channels have dx and pitch (not dy) divided by 2, which means that each two Y pixel in a row share the same U and V pixel.
#define	IMAGE_YCBYCR   (14)
	Only used Internally.
#define	IMAGE_UNSET   (0xFFFF)
	Unset/Unused Information.

Detailed Description

type	Definition	Image Type	Memory Requirement	st	ccmp1	ccmp2
0	IMAGE_GREY	Grey-Scale Image (U8)	1 * dy * pitch	grey	—	—
1	IMAGE_BAYER	Bayer Pattern	1 * dy * pitch	bayer	—	—
2	IMAGE_RGB	Color Image RGB	3 * dy * pitch	red	green	blue
3	IMAGE_CBCR444	Color Image YCbCr 4:4:4	3 * dy * pitch	y	u	v
4	IMAGE_CBCR411	Color Image YCbCr 4:1:1	3/2 * dy * pitch	y	u	v
5	IMAGE_YUVNORM	Normalized YcbCr 4:4:4	3 * dy * pitch	y	u*	v*
6	IMAGE_IHS	Color Image IHS (HSI)	3 * dy * pitch	i	h	s
7	IMAGE_RGBO	Reserved	—	—	—	—
8	IMAGE_VECTOR	Vector Image	2 * dy * pitch	mag	dir	—
9	IMAGE_GREY16	Grey-Scale Image (U16)	2 * dy * pitch	grey	—	—
10	IMAGE_GREY32	Grey-Scale Image (U32)	4 * dy * pitch	grey	—	—
11	IMAGE_CMPLX16	Complex Valued Image	4 * dy * pitch	cmplx	—	—
13	IMAGE_CBCR422	Color Image YCbCr 4:2:2	2 * dy * pitch	y	u	v
14	IMAGE_YCBYCR	Reserved	—	—	—	—

For a grey scale image, the upper left pixel is stored at address st. Going 1 to the right (x-direction), 1 must be added to this address. Going down (y-direction), pitch must be added. For a color image, two additional memory blocks are needed to store the color information. It makes sense to store these 2 blocks tightly behind the first block and behind each other, but this is not mandatory.