本文主要是介绍1.8. 绘制图片 Drawing Images，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！
1.8. 绘制图片 Drawing Images
As mentioned previously, OpenGL has a great deal of support for drawing images in addition to
its support for drawing 3D geometry. In OpenGL parlance, images are called PIXEL RECTANGLES.
The values that define a pixel rectangle start out in application-controlled memory as shown in
Figure 1.1 (11). Color or grayscale pixel rectangles are rendered into the frame buffer with
glDrawPixels, and bitmaps are rendered into the frame buffer with glBitmap. Images that are
destined for texture memory are specified with glTexImage or glTexSubImage. Up to a point, the
same basic processing is applied to the image data supplied with each of these commands.
1.8.1. Pixel Unpacking
OpenGL reads image data provided by the application in a variety of formats. Parameters that
define how the image data is stored in memory (length of each pixel row, number of rows to
skip before the first one, number of pixels to skip before the first one in each row, etc.) can be
specified with glPixelStore. So that operations on pixel data can be defined more precisely, pixels
read from application memory are converted into a coherent stream of pixels by an operation
referred to as PIXEL UNPACKING (12). When a pixel rectangle is transferred to OpenGL by a call like
glDrawPixels, this operation applies the current set of pixel unpacking parameters to determine
how the image data should be read and interpreted. As each pixel is read from memory, it is
converted to a PIXEL GROUP that contains either a color, a depth, or a stencil value. If the pixel
group consists of a color, the image data is destined for the color buffer in the frame buffer. If
the pixel group consists of a depth value, the image data is destined for the depth buffer. If the
pixel group consists of a stencil value, the image data is destined for the stencil buffer. Color
values are made up of a red, a green, a blue, and an alpha component (i.e., RGBA) and are
constructed from the input image data according to a set of rules defined by OpenGL. The result
is a stream of RGBA values that are sent to OpenGL for further processing.
1.8.2. Pixel Transfer
After a coherent stream of image pixels is created, pixel rectangles undergo a series of
operations called PIXEL TRANSFER (13). These operations are applied whenever pixel rectangles are
transferred from the application to OpenGL (glDrawPixels, glTexImage, glTexSubImage), from OpenGL
back to the application (glReadPixels), or when they are copied within OpenGL (glCopyPixels,
glCopyTexImage, glCopyTexSubImage).
The behavior of the pixel transfer stage is modified with glPixelTransfer. This command sets state
that controls whether red, green, blue, alpha, and depth values are scaled and biased. It can
also set state that determines whether incoming color or stencil values are mapped to different
color or stencil values through the use of a lookup table. The lookup tables used for these
operations are specified with the glPixelMap command.
Some additional operations that occur at this stage are part of the OpenGL IMAGING SUBSET, which
is an optional part of OpenGL. Hardware vendors that find it important to support advanced
imaging capabilities will support the imaging subset in their OpenGL implementations, and other
vendors will not support it. To determine whether the imaging subset is supported, applications
need to call glGetString with the symbolic constant GL_EXTENSIONS. This returns a list of
extensions supported by the implementation; the application should check for the presence of
the string "ARB_imaging" within the returned extension string.
The pixel transfer operations that are defined to be part of the imaging subset are convolution,
color matrix, histogram, min/max, and additional color lookup tables. Together, they provide
powerful image processing and color correction operations on image data as it is being
transferred to, from, or within OpenGL.
1.8.3. Rasterization and Back-End Processing
Following the pixel transfer stage, fragments are generated through rasterization of pixel
rectangles in much the same way as they are generated from 3D geometry (14). This process,
along with the current OpenGL state, determines where the image will be drawn in the frame
buffer. Rasterization takes into account the current RASTER POSITION, which can be set with
glRasterPos or glWindowPos, and the current zoom factor, which can be set with glPixelZoom and
which causes an image to be magnified or reduced in size as it is drawn.
After fragments have been generated from pixel rectangles, they undergo the same set of
fragment processing operations as geometric primitives (6) and then go on to the remainder of
the OpenGL pipeline in exactly the same manner as geometric primitives, all the way until
pixels are deposited in the frame buffer (8, 9, 10).
Pixel values provided through a call to glTexImage or glTexSubImage do not go through rasterization
or the subsequent fragment processing but directly update the appropriate portion of texture
memory (15).
1.8.4. Read Control
Pixel rectangles are read from the frame buffer and returned to application memory with
glReadPixels. They can also be read from the frame buffer and written to another portion of the
frame buffer with glCopyPixels, or they can be read from the frame buffer and written into texture
memory with glCopyTexImage or glCopyTexSubImage. In all of these cases, the portion of the frame
buffer that is to be read is controlled by the READ CONTROL stage of OpenGL and set with the
glReadBuffer command (16).
The values read from the frame buffer are sent through the pixel transfer stage (13) in which
various image processing operations can be performed. For copy operations, the resulting pixels
are sent to texture memory or back into the frame buffer, depending on the command that
initiated the transfer. For read operations, the pixels are formatted for storage in application
memory under the control of the PIXEL PACKING stage (17). This stage is the mirror of the pixel
unpacking stage (12), in that parameters that define how the image data is to be stored in
memory (length of each pixel row, number of rows to skip before the first one, number of pixels
to skip before the first one in each row, etc.) can be specified with glPixelStore. Thus, application
developers enjoy a lot of flexibility in determining how the image data is returned from OpenGL
into application memory.
这篇关于1.8. 绘制图片 Drawing Images的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！
