位图 - 在MATLAB中的矢量图形导出期间渲染绘图的一部分

时间:2017-05-31 13:55:41

标签: matlab graphics vector-graphics bitmapimage

我有一个相当复杂的函数,我使用fsurf绘制了一个相当高的MeshDensity(我不能低于默认值,即35)。导出这个数字(saveas(gcf, 'file.pdf', 'pdf');)会产生一个质量非常好的20个MB的pdf文件,但是它的速度非常慢。我想减小文件大小,最重要的是,这个pdf文件的复杂性,而不是将整个绘图(我的意思是整个MATLAB图)导出为位图。我怎么能这样做?

完美的答案将解释我如何将表面图(我的意思是,只是白色背景上的彩色函数表面)转换为位图,同时保持轴和标签的矢量化特性。

更新:以下是此类情节的示例。

enter image description here

1 个答案:

答案 0 :(得分:0)

这是我的函数BitmapRender,它是Bitmap渲染图的一部分:

%% Test Code
clc;clf;
Objects = surf(-4-2*peaks);
hold('on');
Objects(2 : 50) = plot(peaks);
Objects(51) = imagesc([20 40], [0, 5], magic(100));
hold('off');
ylim([0 10]);
zlim([-10 15]);
Objects(1).Parent.GridLineStyle = 'none';
view(45, 45);
set(gcf, 'Color', 'white');
rotate3d on

saveas(gcf, 'pre.pdf');
BitmapRender(gca, Objects(2 : 3 : end));
% BitmapRender(gca, Objects(2 : 3 : end), [0.25 0.25 0.5 0.5], false);
saveas(gcf, 'post.pdf');

除了(重新)处理可见性之外,函数本身非常简单,因为按空格键(旋转,缩放等之后)会重新渲染图形。

function BitmapRender(Axes, KeepObjects, RelativePosition, Draft, Key)

if nargin < 2
    KeepObjects = [];
end
if nargin < 3
    RelativePosition = [0 0 1 1];
end
if nargin < 4
    Draft = false;
end
if nargin < 5
    Key = '';
end

Figure = Axes.Parent;
FigureInnerWH = Figure.InnerPosition([3 4 3 4]);
PixelPosition = round(RelativePosition .* FigureInnerWH);

if isempty(Key)
    OverlayAxes = axes(Figure, 'Units', 'Normalized', 'Position', PixelPosition ./ FigureInnerWH);
    if Draft
        OverlayAxes.Box = 'on';
        OverlayAxes.Color = 'none';
        OverlayAxes.XTick = [];
        OverlayAxes.YTick = [];
        OverlayAxes.HitTest = 'off';
    else
        uistack(OverlayAxes, 'bottom');
        OverlayAxes.Visible = 'off';
    end
    setappdata(Figure, 'BitmapRenderOriginalVisibility', get(Axes.Children, 'Visible'));

    Axes.CLimMode = 'manual';
    Axes.XLimMode = 'manual';
    Axes.YLimMode = 'manual';
    Axes.ZLimMode = 'manual';

    hManager = uigetmodemanager(Figure);
    [hManager.WindowListenerHandles.Enabled] = deal(false);
    set(Figure, 'KeyPressFcn', @(f, e) BitmapRender(gca, KeepObjects, RelativePosition, Draft, e.Key));
elseif strcmpi(Key, 'space')
    OverlayAxes = findobj(Figure, 'Tag', 'BitmapRenderOverlayAxes');
    delete(get(OverlayAxes, 'Children'));
    OriginalVisibility = getappdata(Figure, 'BitmapRenderOriginalVisibility');
    [Axes.Children.Visible] = deal(OriginalVisibility{:});
else
    return;
end

if Draft
    return;
end

Axes.Visible = 'off';

KeepObjectsVisibility = get(KeepObjects, 'Visible');
[KeepObjects.Visible] = deal('off');

drawnow;
Frame = getframe(Figure, PixelPosition);

[Axes.Children.Visible] = deal('off');
Axes.Visible = 'on';
Axes.Color = 'none';
if numel(KeepObjects) == 1
    KeepObjects.Visible = KeepObjectsVisibility;
else
    [KeepObjects.Visible] = deal(KeepObjectsVisibility{:});
end

Image = imagesc(OverlayAxes, Frame.cdata);
uistack(Image, 'bottom');
OverlayAxes.Tag = 'BitmapRenderOverlayAxes';
OverlayAxes.Visible = 'off';

end

显然,该解决方案在屏幕像素方面是像素完美的。两个pdf文件(prepost)看起来像这样。请注意,曲面,图像和一些绘图线是位图渲染的,但是其他一些绘图线以及轴和标签仍然是矢量化的。

enter image description here enter image description here