MATLAB：如何基于相机视图创建存储在 3D 修补对象的 FaceVertexCData 中的数据的 2D 像素图？

Question

最小可行示例：

% Define a simple 6 sided cube with dimensions 0-1 in X, Y, and Z
vertices = [0,0,1;  % 1 Front-Top-Left (FTL)
            1,0,1;  % 2 Front-Top-Right (FTR)
            1,0,0;  % 3 Front-Bottom-Right (FBoR)
            0,0,0;  % 4 Front-Bottom-Left (FBoL)
            0,1,1;  % 5 Back-Top-Left (BaTL)
            1,1,1;  % 6 Back-Top-Right (BaTR)
            1,1,0;  % 7 Back-Bottom-Right (BaBoR)
            0,1,0]; % 8 Back-Bottom-Left (BaBoL)
faces = [1,2,3,4;  % 1 Front
         2,6,7,3;  % 2 Right
         6,5,8,7;  % 3 Back
         5,1,4,8;  % 4 Left
         1,5,6,2;  % 5 Top
         4,3,7,8]; % 6 Bottom
temps = [5;10;15;20;25;30]; % One temperature per face
% Plot the cube
fig1 = figure;
ax1 = axes(fig1);
patch('Faces',faces,'Vertices',vertices,'FaceColor','flat','EdgeColor','none','FaceVertexCData',temps,'Parent',ax1);
% Rotate to an isometric type of view angle
ax1.View = [-42.5 28];
% Turn off axes
ax1.XAxis.Visible = 'off';
ax1.YAxis.Visible = 'off';
ax1.ZAxis.Visible = 'off';
% Set data aspect ratio to 1 1 1
daspect(ax1,[1 1 1]);

然后我需要生成一个看起来像这样的矩阵的代码（对于上面的代码）：

% -------------------------------------------------------------------------
% Code goes here that should produce a 2D matrix that looks something like:
% (15x15 as example, but would need to be more like 1024x1024 up to 
%  4096x4096 in real world use cases)
%       1  2  3  4  5  6  7  8  9 10 11 12 13 14 15
% 1  % 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 
% 2  % 00 00 00 00 00 00 25 00 00 00 00 00 00 00 00 
% 3  % 00 00 00 00 00 25 25 25 00 00 00 00 00 00 00 
% 4  % 00 00 00 00 25 25 25 25 25 00 00 00 00 00 00 
% 5  % 00 00 00 25 25 25 25 25 25 25 00 00 00 00 00 
% 6  % 00 00 20 25 25 25 25 25 25 10 00 00 00 00 00 
% 7  % 00 00 20 20 20 25 25 25 10 10 00 00 00 00 00 
% 8  % 00 00 20 20 20 20 25 10 10 10 00 00 00 00 00 
% 9  % 00 00 20 20 20 20 20 10 10 10 00 00 00 00 00 
% 10 % 00 00 20 20 20 20 20 10 10 10 00 00 00 00 00 
% 11 % 00 00 00 20 20 20 20 10 10 00 00 00 00 00 00 
% 12 % 00 00 00 00 20 20 20 10 00 00 00 00 00 00 00 
% 13 % 00 00 00 00 00 20 20 10 00 00 00 00 00 00 00 
% 14 % 00 00 00 00 00 00 20 00 00 00 00 00 00 00 00 
% 15 % 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

现实世界的用例包括从多个不同角度投影到包含超过 100k 多边形的复杂 3D 几何图形。还存在几何体的一部分与几何体的其他部分重叠的情况，并且解决方案需要仅考虑特定视角下的可见几何体，并忽略“最接近相机”面后面的所有其他看不见的面。当前的绘图解决方案将每个主体绘制一个面片对象作为 hgtransform 的子对象，并且模型中可以有数百个主体。

如有任何帮助，我们将不胜感激！

Answer 1

假设您可以访问 MATLAB 图窗并且它位于所需的视图中。这个想法是捕获图中显示的图像，并使用

colormap

及其关联的

clim

将颜色映射到初始值。

% Extract colormap limits
h = gca;
lims = h.CLim;

% Extract colormap RGB values and create correspondence vector from lims
map = colormap();
vals = linspace(lims(1),lims(2),size(map,1));

% Get the figure color content from the current view (convert to double
% precision)
F = getframe(fig1);
Fdouble = im2double(F.cdata);

% Map each pixel RGB value in F.cdata to the corresponding value, using the map
% and the corresponding vals
% To do so, for each pixel color (F.cdata(ii,jj,:)) get idx of the closest color in map
out = zeros(size(F.cdata,1),size(F.cdata,2));

for ii = 1:size(F.cdata,1)
    for jj = 1:size(F.cdata,2)
        
        [M,I] = min(sqrt((map(:,1)-Fdouble(ii,jj,1)).^2+(map(:,2)-Fdouble(ii,jj,2)).^2+(map(:,3)-Fdouble(ii,jj,3)).^2));
        % if the color is close enough, get the corresponding value in
        % vals. if not keep 0;

        if M < 0.1

            out(ii,jj) = vals(I);

        end

    end
end

% Flip the rows in out to match the input:
out = out(end:-1:1,:);

输出

有趣的是，表示输出的最简单方法是将其