VXmakeimage, VXsetimage, VXembedimage, VXbitimage, VXidximage, VXresetimage − VisionX image access

int VXmakeimage(image, type, bbx, chan);
VisXimage_t *image;
float bbx[4];
int type, chan;
int VXsetimage(image, a, f);
VisXimage_t *image;
VisXelem_t *a;
VisXfile_t *f;
int VXembedimage(i, j ,x1,x2,y1,y2);
VisXimage_t *i, *j;
float x1,x2,y1,y2;
VXbitimage(i);
VisXimage_t *i;
int VXresetimage(i);
VisXimage_t *i;

Image data in VisionX is managed by a set of programming tools that record image array accessing information in a single structure (VisXimage_t). The image is considered as a two dimensional array. An image in a data list consists of two consecutive data elements, a bounding box followed by a conforming pixel data element. The image tools provide convenient structured access to the data element under the constraints of the bounding box.

The image toolkit contains the following functions: VXmakeimage − creates an image structure, VXsetimage − creates an image reference structure for a list VXembedimage − creates an image copy with a border, VXfloatimage − creates a float image copy with an optional border, VXresetimage − free memory associated with an image reference structure, VXbitimage − update a bitplane image structure, VXsbufimage − set up a buffer for n image frames, VXrbufimage − read the next image frame into the buffer, and VXidximage − sets the image index convention

VXmakeimage Initializes an image structure. Its arguments are: the image structure to be initialized, the base data type e.g. VX_PBYTE, the bounding box for the image, and the number of channels in the image. A list structure is allocated that contains the image information and a reference structure is allocated that will be used for convenient image data access. For example consider the following statement:

VisXimage_t foo;

VXmakeimage ( &foo, VX_PFLOAT, (float){0, 5, 0, 8}, 3);

A (5 x 8) pixel image is created in which each pixel consists of three floating point numbers. The bounding box is specified by an array of four floating point numbers (xmin, xmax, ymin, ymax) and the number of channels is specified by an integer. Only a 4 element bounding box is created; this may be extended to a six element bounding box by a call to VXpzset(foo.list, n).

VXembedimage creates a copy of an image in a second image structure. The additional arguments are positive integers that specify a border region of zeros for each of the four sides of the image. This feature is very useful for local filtering operations since it permits simple access to elements outside the boundary of the original image. To make a direct copy of an image simply set the border values to zero.

VXbitimage Bitplane data (packed 8-bits/byte) is expanded into a unsigned int buffer (1-byte /pixel) for convenient manipulation (zero is false anything else is true). Before writing the image list it is necessary to repack the bits; this is achieved by call to VXbitimage. If the image type is not bitplane (VX_PBIT) then VXbitimage has no effect.

VXsetimage scans a list for a valid image pixel type and if one is found it constructs a reference structure to provide convenient access to it and returns the value one. If no pixel type is found VXmakeimage returns a zero and does nothing. The first argument to Vmakeimage is the structure to be created, the second argument is the list with image data and the third argument is the file structure for that list. The list is scanned and information is obtained using the following algorithm.

   1. check to see it the  list item is at an image data type.

  2. If not search FORWARD from the given position to
      locate a numeric image data type (not color index images ).

  3. If no image is found by the end of the list then return zero

  4. If an image is found the critical parameters are obtained
      (bounding box and number of channels)

  5. The BBX should precede the pixel data element

  6. if no BBX found then return zero.

  7. Computes the number of channels in the image

  8. Compute the reference structure.

  9. If all OK return 1. imitem points to data elem.
                          flist points to entry pt.
                          glist points to any global list

The base pixel types supported by the image tools are as follows:

NAME, C-Type, EXTENSION, DESCRIPTION

VX_PBYTE, unsigned char, .u, usual byte pixel format

VX_PFLOAT, float, .f, 32-bit floating point

VX_PDOUBLE, double, .d,

64-bit floating point

VX_PBIT, unsigned char, .b, 1-bit (packed 8/byte)

VX_PBYTE, char, .c, 8-bit signed integer

VX_PSHORT, short, .s, 32-bit signed integer

VX_PINT

int

.i

32-bit signed integer

VX_PIDX

unsigned char

.?

8-bit color index

The EXTENSION field specifies the element of the VisXimage data structure which will point to the two dimensional array.

VXresetimage deallocates the dynamic memory associated with the image structure that has been created by either VXmakeimage(3), VXsetimage, VXembedimage, or VXfloatimage(3). VXresetimage also deallocates the image list associated with the image structure unless either the image structure was created by VXsetimage or the image list has been moved somewhere else e.g. by VXlknlst. It is important to deallocate image structures with VXresetimage before reusing them with a different image.

VXfloatimage(3), VXrbufimage(3)

A. P. Reeves