一般
PDF 规范允许 PDF 提供多个页面边界，参见 this answer。除此之外，内容边界可能仅来自页面内容，例如来自
-
表单 XObject：
表单 XObject 是 PDF 内容流，它是任何图形对象序列（包括路径对象、文本对象和采样图像）的自包含描述。一个表单 XObject 可以被绘制多次（在多个页面上或在同一页面上的多个位置）并且每次都产生相同的结果，仅受调用时的图形状态的影响。
-
剪切路径：
图形状态应包含一个当前剪切路径，它限制了受绘画操作符影响的页面区域。该路径的封闭子路径应定义可以绘制的区域。落在该区域内的标记应应用于页面；那些落在它外面的人不会。
...
要找到其中任何一个，必须解析页面内容，查找适当的操作，并计算结果边界。
在 OP 的情况下
每个示例 PDF 都明确定义了一个页面边界，即 MediaBox。因此，所有其他 PDF 页面边界（CropBox、BleedBox、TrimBox、ArtBox）默认为它。所以难怪在你的尝试中
each (PDRectangle) 返回页面的全高：842
它们都不包含表单 XObject，但都使用剪切路径。
-
如果是test-pdf4.pdf：
Start at: 28.31999969482422, 813.6799926757812
Line to: 565.9199829101562, 813.6799926757812
Line to: 565.9199829101562, 660.2196655273438
Line to: 28.31999969482422, 660.2196655273438
Line to: 28.31999969482422, 813.6799926757812
（这可能与您问题中的草图相匹配。）
-
如果是test-pdf5.pdf：
Start at: 23.0, 34.0
Line to: 572.0, 34.0
Line to: 572.0, -751.0
Line to: 23.0, -751.0
Line to: 23.0, 34.0
和
Start at: 23.0, 819.0
Line to: 572.0, 819.0
Line to: 572.0, 34.0
Line to: 23.0, 34.0
Line to: 23.0, 819.0
由于与草图的匹配，我假设 Illustrator 会考虑在一个重要的剪切路径生效时绘制的所有内容，一个以剪切路径为边框的 复合元素。
使用 PDFBox 查找剪切路径
我使用 PDFBox 来查找上面提到的剪切路径。我使用了目前正在开发的 2.0.0 版本的当前 SNAPSHOT，因为与当前版本 1.8.8 相比，所需的 API 已经有了很大改进。
我将PDFGraphicsStreamEngine 扩展为ClipPathFinder 类：
public class ClipPathFinder extends PDFGraphicsStreamEngine implements Iterable<Path>
{
public ClipPathFinder(PDPage page)
{
super(page);
}
//
// PDFGraphicsStreamEngine overrides
//
public void findClipPaths() throws IOException
{
processPage(getPage());
}
@Override
public void appendRectangle(Point2D p0, Point2D p1, Point2D p2, Point2D p3) throws IOException
{
startPathIfNecessary();
currentPath.appendRectangle(toFloat(p0), toFloat(p1), toFloat(p2), toFloat(p3));
}
@Override
public void drawImage(PDImage pdImage) throws IOException { }
@Override
public void clip(int windingRule) throws IOException
{
currentPath.complete(windingRule);
paths.add(currentPath);
currentPath = null;
}
@Override
public void moveTo(float x, float y) throws IOException
{
startPathIfNecessary();
currentPath.moveTo(x, y);
}
@Override
public void lineTo(float x, float y) throws IOException
{
currentPath.lineTo(x, y);
}
@Override
public void curveTo(float x1, float y1, float x2, float y2, float x3, float y3) throws IOException
{
currentPath.curveTo(x1, y1, x2, y2, x3, y3);
}
@Override
public Point2D.Float getCurrentPoint() throws IOException
{
return currentPath.getCurrentPoint();
}
@Override
public void closePath() throws IOException
{
currentPath.closePath();
}
@Override
public void endPath() throws IOException
{
currentPath = null;
}
@Override
public void strokePath() throws IOException
{
currentPath = null;
}
@Override
public void fillPath(int windingRule) throws IOException
{
currentPath = null;
}
@Override
public void fillAndStrokePath(int windingRule) throws IOException
{
currentPath = null;
}
@Override
public void shadingFill(COSName shadingName) throws IOException
{
currentPath = null;
}
void startPathIfNecessary()
{
if (currentPath == null)
currentPath = new Path();
}
Point2D.Float toFloat(Point2D p)
{
if (p == null || (p instanceof Point2D.Float))
{
return (Point2D.Float)p;
}
return new Point2D.Float((float)p.getX(), (float)p.getY());
}
//
// Iterable<Path> implementation
//
public Iterator<Path> iterator()
{
return paths.iterator();
}
Path currentPath = null;
final List<Path> paths = new ArrayList<Path>();
}
它使用这个帮助类来表示路径：
public class Path implements Iterable<Path.SubPath>
{
public static class Segment
{
Segment(Point2D.Float start, Point2D.Float end)
{
this.start = start;
this.end = end;
}
public Point2D.Float getStart()
{
return start;
}
public Point2D.Float getEnd()
{
return end;
}
final Point2D.Float start, end;
}
public class SubPath implements Iterable<Segment>
{
public class Line extends Segment
{
Line(Point2D.Float start, Point2D.Float end)
{
super(start, end);
}
//
// Object override
//
@Override
public String toString()
{
StringBuilder builder = new StringBuilder();
builder.append(" Line to: ")
.append(end.getX())
.append(", ")
.append(end.getY())
.append('\n');
return builder.toString();
}
}
public class Curve extends Segment
{
Curve(Point2D.Float start, Point2D.Float control1, Point2D.Float control2, Point2D.Float end)
{
super(start, end);
this.control1 = control1;
this.control2 = control2;
}
public Point2D getControl1()
{
return control1;
}
public Point2D getControl2()
{
return control2;
}
//
// Object override
//
@Override
public String toString()
{
StringBuilder builder = new StringBuilder();
builder.append(" Curve to: ")
.append(end.getX())
.append(", ")
.append(end.getY())
.append(" with Control1: ")
.append(control1.getX())
.append(", ")
.append(control1.getY())
.append(" and Control2: ")
.append(control2.getX())
.append(", ")
.append(control2.getY())
.append('\n');
return builder.toString();
}
final Point2D control1, control2;
}
SubPath(Point2D.Float start)
{
this.start = start;
currentPoint = start;
}
public Point2D getStart()
{
return start;
}
void lineTo(float x, float y)
{
Point2D.Float end = new Point2D.Float(x, y);
segments.add(new Line(currentPoint, end));
currentPoint = end;
}
void curveTo(float x1, float y1, float x2, float y2, float x3, float y3)
{
Point2D.Float control1 = new Point2D.Float(x1, y1);
Point2D.Float control2 = new Point2D.Float(x2, y2);
Point2D.Float end = new Point2D.Float(x3, y3);
segments.add(new Curve(currentPoint, control1, control2, end));
currentPoint = end;
}
void closePath()
{
closed = true;
currentPoint = start;
}
//
// Iterable<Segment> implementation
//
public Iterator<Segment> iterator()
{
return segments.iterator();
}
//
// Object override
//
@Override
public String toString()
{
StringBuilder builder = new StringBuilder();
builder.append(" {\n Start at: ")
.append(start.getX())
.append(", ")
.append(start.getY())
.append('\n');
for (Segment segment : segments)
builder.append(segment);
if (closed)
builder.append(" Closed\n");
builder.append(" }\n");
return builder.toString();
}
boolean closed = false;
final Point2D.Float start;
final List<Segment> segments = new ArrayList<Path.Segment>();
}
public class Rectangle extends SubPath
{
Rectangle(Point2D.Float p0, Point2D.Float p1, Point2D.Float p2, Point2D.Float p3)
{
super(p0);
lineTo((float)p1.getX(), (float)p1.getY());
lineTo((float)p2.getX(), (float)p2.getY());
lineTo((float)p3.getX(), (float)p3.getY());
closePath();
}
//
// Object override
//
@Override
public String toString()
{
StringBuilder builder = new StringBuilder();
builder.append(" {\n Rectangle\n Start at: ")
.append(start.getX())
.append(", ")
.append(start.getY())
.append('\n');
for (Segment segment : segments)
builder.append(segment);
if (closed)
builder.append(" Closed\n");
builder.append(" }\n");
return builder.toString();
}
}
public int getWindingRule()
{
return windingRule;
}
void complete(int windingRule)
{
finishSubPath();
this.windingRule = windingRule;
}
void appendRectangle(Point2D.Float p0, Point2D.Float p1, Point2D.Float p2, Point2D.Float p3) throws IOException
{
finishSubPath();
currentSubPath = new Rectangle(p0, p1, p2, p3);
finishSubPath();
}
void moveTo(float x, float y) throws IOException
{
finishSubPath();
currentSubPath = new SubPath(new Point2D.Float(x, y));
}
void lineTo(float x, float y) throws IOException
{
currentSubPath.lineTo(x, y);
}
void curveTo(float x1, float y1, float x2, float y2, float x3, float y3) throws IOException
{
currentSubPath.curveTo(x1, y1, x2, y2, x3, y3);
}
Point2D.Float getCurrentPoint() throws IOException
{
return currentPoint;
}
void closePath() throws IOException
{
currentSubPath.closePath();
finishSubPath();
}
void finishSubPath()
{
if (currentSubPath != null)
{
subPaths.add(currentSubPath);
currentSubPath = null;
}
}
//
// Iterable<Path.SubPath> implementation
//
public Iterator<SubPath> iterator()
{
return subPaths.iterator();
}
//
// Object override
//
@Override
public String toString()
{
StringBuilder builder = new StringBuilder();
builder.append("{\n Winding: ")
.append(windingRule)
.append('\n');
for (SubPath subPath : subPaths)
builder.append(subPath);
builder.append("}\n");
return builder.toString();
}
Point2D.Float currentPoint = null;
SubPath currentSubPath = null;
int windingRule = -1;
final List<SubPath> subPaths = new ArrayList<Path.SubPath>();
}
ClipPathFinder 类是这样使用的：
PDDocument document = PDDocument.load(PDFRESOURCE, null);
PDPage page = document.getPage(PAGENUMBER);
ClipPathFinder finder = new ClipPathFinder(page);
finder.findClipPaths();
for (Path path : finder)
{
System.out.println(path);
}
document.close();