hbVPImage

typedef struct hbVPImage {
  uint8_t imageFormat;
  uint8_t imageType;
  int32_t width;
  int32_t height;
  int32_t stride;
  void *dataVirAddr;
  uint64_t dataPhyAddr;
  void *uvVirAddr;
  uint64_t uvPhyAddr;
  int32_t uvStride;
} hbVPImage;

Picture Structure. Only BPU memory is supported.

Picture channel description: take HB_VP_IMAGE_FORMAT_RGB as an example, when the number of channels is 3, the imageType is HB_VP_IMAGE_TYPE_U8C3. When the number of channels is 4, the data exists an extra memory due to alignment requirements, at this time it is necessary to specify the imageType as HB_VP_IMAGE_TYPE_U8C4 (RRR* GGG* BBB*).

Member

Member Name	Description
`imageFormat`	Picture format.
`imageType`	Picture type.
`width`	The pixel width of the picture.
`height`	The pixel height of the picture.
`stride`	The pixel offset of the picture, expressed as a number of bytes.
`dataVirAddr`	The logical address of the picture on the arm side.
`dataPhyAddr`	The physical address of the picture.
`uvVirAddr`	The uv data logical address when the picture format is NV12.
`uvPhyAddr`	The uv data physical address when the picture format is NV12.
`uvStride`	The offset of the uv data when the picture format is NV12, expressed as the number of bytes.

Offset description: stride is in the range [bytes_per_pixel * width, bytes_per_pixel * MAX_IMAGE_WIDTH] and satisfies the bytes_per_pixel alignment. The bytes_per_pixel indicate the number of bytes per pixel, the size of which depends on the picture format imageType, width indicates the pixel width of the picture, MAX_IMAGE_WIDTH indicates the maximum width of the picture pixels. When the picture format is NV12, uvStride ranges from [2*uv_width, MAX_IMAGE_WIDTH] and must be even. When using pyramid and gdc hardware, the values of stride and uvStride should be greater than or equal to the width and meet the 16-byte alignment requirement.

When the image format is NV12, if it is 8-bit, the imageType should be set to HB_VP_IMAGE_TYPE_U8C1. If it is 10-bit or 12-bit, the imageType should be set to HB_VP_IMAGE_TYPE_U16C1, indicating that the channel data is aligned to 16-bit storage. Specifically, if the imageFormat is set to HB_VP_IMAGE_FORMAT_NV12_Y10C8_MSB or HB_VP_IMAGE_FORMAT_NV12_Y12C8_MSB, it means only the Y component has a pixel width of 10-bit or 12-bit and is aligned to 16-bit storage, while the UV components still have an 8-bit pixel width. When the image is in multi-plane formats such as RGB_P, BGR_P, or YUV444_P, the imageType should be set according to the actual data type of the elements to the corresponding single-channel type. For example, for 3-channel multi-plane data in S16 format, choose HB_VP_IMAGE_TYPE_S16C1.

#hbVPImage

hbVPImage