geometric image manipulation.htm

是一部关于java高级图像处理的的一本入门书

          vertical resampling. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>width</CODE> 
          <P></P>
        <TD>The width of a horizontal resampling kernel. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>height</CODE> 
          <P></P>
        <TD>The height of a vertical resampling kernel. Ignored if 
          <CODE>dataV</CODE> is null. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>subsample-BitsH</CODE> 
          <P></P>
        <TD>The log<SUB>2</SUB> of the number of horizontal subsample bins. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>subsample-BitsV</CODE> 
          <P></P>
        <TD>The log<SUB>2</SUB> of the number of vertical subsample bins. 
          Ignored if <CODE>dataV</CODE> is null. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>precision-Bits</CODE> 
          <P></P>
        <TD>The number of bits of fractional precision to be used when 
          resampling integral sample values. The same value is used for both 
          horizontal and vertical resampling. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>dataH</CODE> 
          <P></P>
        <TD>The horizontal table entries, as a float array of 
          2<SUP>subsampleBitsH</SUP> entries each of length <CODE>width</CODE>. 
          <P></P>
      <TR vAlign=top>
        <TD>
          <P></P>
        <TD><CODE>dataV</CODE> 
          <P></P>
        <TD>The vertical table entries, as a float array of 
          2<SUP>subsampleBitsV</SUP> entries each of length <CODE>height</CODE>, 
          or null. If null, the <CODE>dataH</CODE> table is used for vertical 
          interpolation as well and the <CODE>topPadding</CODE>, 
          <CODE>height</CODE>, and <CODE>subsampleBitsV</CODE> parameters are 
          ignored. 
          <P></P></TR></TBODY></TABLE></A>
    <P></P></DD></DL><PRE><UL>
<LI>InterpolationTable(int leftPadding, int topPadding, int width, 
       int height, int subsampleBitsH, int subsampleBitsV, 
       int  precisionBits, double[] dataH, double[] dataV)
<P></P></LI></UL></PRE>
  <DL><A name=71012>
    <DT>
    <DD>constructs an <CODE>InterpolationTable</CODE> with specified horizontal 
    and vertical extents (support), number of horizontal and vertical bins, 
    fixed-point fractional precision, and kernel entries. </A>
    <P></P></DD></DL><PRE><UL>
<LI>InterpolationTable(int leftPadding, int topPadding, int width, 
       int height, int subsampleBitsH, int subsampleBitsV, 
       int  precisionBits, int[] dataH, int[] dataV)
<P></P></LI></UL></PRE>
  <DL><A name=73497>
    <DT>
    <DD>constructs an InterpolationTable with specified horizontal and vertical 
    extents (support), number of horizontal and vertical bins, fixed-point 
    fractional precision, and int kernel entries. </A>
    <P></P></DD></DL><A name=71057>
  <H4>8.2.5.6 <IMG src="Geometric Image Manipulation.files/space.gif">Additional 
  Interpolation Table-related Methods</H4></A>The 
  <CODE>InterpolationTable</CODE> class provides several methods for retrieving 
  an interpolation table's kernel data values, subsample size, and precision. 
  <P>
  <TABLE border=0>
    <TBODY>
    <TR>
      <TD><IMG src="Geometric Image Manipulation.files/cistine.gif"></TD>
      <TD>
        <HR>
        <B>API:</B> <CODE>javax.media.jai.InterpolationTable </CODE>
        <HR>
      </TD></TR></TBODY></TABLE><PRE><UL>
<LI>int getSubsampleBitsH()
<P></P></LI></UL></PRE>
  <DL><A name=71075>
    <DT>
    <DD>returns the number of bits used to index subsample positions in the 
    horizontal direction. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int getSubsampleBitsV()
<P></P></LI></UL></PRE>
  <DL><A name=71105>
    <DT>
    <DD>returns the number of bits used to index subsample positions in the 
    vertical direction. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int getPrecisionBits()
<P></P></LI></UL></PRE>
  <DL><A name=71132>
    <DT>
    <DD>returns the number of bits of fractional precision used to store the 
    fixed-point table entries. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int getLeftPadding()
<P></P></LI></UL></PRE>
  <DL><A name=71148>
    <DT>
    <DD>returns the number of bits of fractional precision used to store the 
    fixed-point table entries. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int getTopPadding()
<P></P></LI></UL></PRE>
  <DL><A name=71163>
    <DT>
    <DD>returns the number of samples required above the center. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int getWidth()
<P></P></LI></UL></PRE>
  <DL><A name=71173>
    <DT>
    <DD>returns the number of samples required for horizontal resampling. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int getHeight()
<P></P></LI></UL></PRE>
  <DL><A name=71187>
    <DT>
    <DD>returns the number of samples required for vertical resampling. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int[] getHorizontalTableData()
<P></P></LI></UL></PRE>
  <DL><A name=71199>
    <DT>
    <DD>returns the integer (fixed-point) horizontal table data. </A>
    <P></P></DD></DL><PRE><UL>
<LI>int[] getVerticalTableData()
<P></P></LI></UL></PRE>
  <DL><A name=71227>
    <DT>
    <DD>returns the integer (fixed-point) vertical table data. </A>
    <P></P></DD></DL><PRE><UL>
<LI>float[] getHorizontalTableDataFloat()
<P></P></LI></UL></PRE>
  <DL><A name=71243>
    <DT>
    <DD>returns the floating-point horizontal table data. </A>
    <P></P></DD></DL><PRE><UL>
<LI>float[] getVerticalTableDataFloat()
<P></P></LI></UL></PRE>
  <DL><A name=71257>
    <DT>
    <DD>returns the floating-point vertical table data. </A>
    <P></P></DD></DL><PRE><UL>
<LI>double[] getHorizontalTableDataDouble()
<P></P></LI></UL></PRE>
  <DL><A name=71271>
    <DT>
    <DD>returns the double horizontal table data. </A>
    <P></P></DD></DL><PRE><UL>
<LI>double[] getVerticalTableDataDouble()
<P></P></LI></UL></PRE>
  <DL><A name=71281>
    <DT>
    <DD>returns the double vertical table data. </A>
    <P></P></DD></DL><A name=56707>
  <H2>8.3 <IMG src="Geometric Image Manipulation.files/space.gif">Geometric 
  Transformation</H2></A>Geometric transformations provide the ability to 
  reposition pixels within an image. Pixels may be relocated from their 
  (<EM>x</EM>,<EM>y</EM>) spatial coordinates in the source image to new 
  coordinates in the destination. Geometric transformations are used, for 
  example, to move (translate), rotate, and scale the geometry of an image. A 
  general type of geometric transformation, warp, is discussed later in this 
  chapter (see <A 
  href="http://java.sun.com/products/java-media/jai/forDevelopers/jai1_0_1guide-unc/Geom-image-manip.doc.html#53798">Section 
  8.7, "Warping</A>"). 
  <P>Geometric transformations are used to register multiple images, correct 
  geometric distortions introduced in the image acquisition process, or to add 
  visual effects. The geometric transformation operations discussed here 
  include: 
  <P>
  <UL>
    <LI>Translation (<CODE>Translate</CODE>) - moves an image up, down, left, or 
    right
    <P></P></LI></UL>
  <UL>
    <LI>Scaling (<CODE>Scale</CODE>) - enlarges or shrinks an image
    <P></P></LI></UL>
  <UL>
    <LI>Rotation (<CODE>Rotate</CODE>) - rotates an image about a given point
    <P></P></LI></UL>
  <UL>
    <LI>Affine (<CODE>Affine</CODE>) - includes translation, scaling, and 
    rotation in one operation
    <P></P></LI></UL>All transformation operations are performed by moving pixel 
  values from their original spatial coordinates to new coordinates in the 
  destination image. Every pixel in the source image is passed through this 
  transformation, creating a geometrically-transformed output pixel location. 
  Each pixel of the source image is transformed, pixel by pixel, to its new 
  location in the destination image. 
  <P>With a very few exceptions, all transformations result in some output pixel 
  locations being missed because no input pixels were transformed there. The 
  missed locations will be devoid of any pixel values and result in a black hole 
  in the destination image. To overcome this problem, intermediate pixel values 
  are estimated through interpolation (See <A 
  href="http://java.sun.com/products/java-media/jai/forDevelopers/jai1_0_1guide-unc/Geom-image-manip.doc.html#51290">"Interpolation" 
  on page 249</A>). One of four <CODE>interpolation</CODE> methods may be 
  selected:
  <P>
  <TABLE cellPadding=3 border=3>
    <CAPTION><FONT size=-1><B></B></FONT></CAPTION>
    <TBODY>
    <TR vAlign=top>
      <TH><A name=64125>interpolation Methods </A>
      <TH><A name=64127>Description </A>
    <TR vAlign=top>
      <TD><A name=64129>INTERP_NEAREST</A><BR>
      <TD><A name=64131>Use nearest-neighbor interpolation</A><BR>
    <TR vAlign=top>
      <TD><A name=64133>INTERP_BILINEAR</A><BR>
      <TD><A name=64135>Use bilinear interpolation</A><BR>
    <TR vAlign=top>
      <TD><A name=64137>INTERP_BICUBIC</A><BR>
      <TD><A name=64139>Use bicubic interpolation</A><BR>
    <TR vAlign=top>
      <TD><A name=64141>INTERP_BICUBIC2</A><BR>
      <TD><A name=64143>Use bicubic2 interpolation (uses a different 
        polynomial function)</A><BR></TR></TBODY></TABLE>
  <P><A name=60866>
  <H3>8.3.1 <IMG src="Geometric Image Manipulation.files/space.gif">Translation 
  Transformation</H3></A>Image translation is the spatial shifting of an image 
  up, down, left, or right. The relationships between the source and destination 
  image coordinates are given by the following equation: 
  <P>
  <P><A name=60873>
  <UL>
    <TABLE>
      <TBODY>
      <TR>
        <TD width=430><IMG 
          src="Geometric Image Manipulation.files/Geom-image-manip.doc.anc.gif"> 
        </TD>
        <TD>(8.1)</TD></TR></TBODY></TABLE></UL></A>
  <DL><A name=60874>
    <DT>
    <DD>where: </A>
    <P>
    <DL><A name=60875>
      <DT>
      <DD><EM>x</EM><SUB>D</SUB> and <EM>y</EM><SUB>D</SUB> are the integer 
      pixel coordinates of the destination image </A>
      <P><A name=60876></P>
      <DT>
      <DD><EM>t</EM><SUB>x</SUB> and <EM>t</EM><SUB>y</SUB> are the translation 
      values </A>
      <P><A name=60877></P>
      <DT>
      <DD><EM>x</EM>'S and <EM>y</EM>'S denote the source image point from which 
      the pixel estimate is computed. </A>
      <P></P></DD></DL></DD></DL>Translation is often used to register multiple