MATLAB如何保存当前视角与光照下的曲面着色对应矩阵？

Got it, let's break down how to capture the shaded intensity values matching your surf plot's lighting and view, then convert it into a 161×161 matrix ready for photometric stereo work. This works for MATLAB, which I assume you're using given the peaks, surf, and light functions mentioned.

Step 1: Generate Your 161×161 Surface Data #

First, create the high-resolution Z matrix using peaks (we'll scale it to 161×161 directly):

% Create 161x161 surface grid
[X, Y, Z] = peaks(161);

Step 2: Set Up Plot with Custom Lighting and View #

Replicate your existing plot setup, making sure to use smooth shading for accurate intensity capture:

figure;
surf(X, Y, Z);
shading interp; % Critical for smooth, continuous intensity values

% Add your specific lighting conditions (adjust parameters to match your setup)
light('Position', [5 5 10], 'Color', [1 1 1], 'Style', 'infinite'); % Example directional light
lighting phong; % Use the same lighting mode you originally used (phong/gouraud/flat)
view(2); % Set top-down 2D view
axis tight; % Ensure the plot fills the axis area to avoid empty borders
axis off; % Optional: hide axes to prevent artifacts in the captured image

Step 3: Capture and Map Rendered Intensities to 161×161 Matrix #

The key is to capture the rendered image and map it back to your original Z matrix dimensions. Here are two methods, with the second being more precise:

Method 1: Quick Resize (Simple but Approximate) #

If your plot is tightly aligned to the data, you can directly resize the captured grayscale image to match Z's size:

% Capture the axis content
frame = getframe(gca);
rendered_img = frame.cdata;

% Convert to grayscale (photometric stereo uses single-channel intensity)
gray_intensity = rgb2gray(rendered_img);

% Resize to match 161x161 Z matrix
shaded_intensity_matrix = imresize(gray_intensity, size(Z));

Method 2: Precise Interpolation (Avoids Resizing Artifacts) #

For better accuracy, map the image pixels to your original X/Y grid using interpolation:

% Capture the axis content
frame = getframe(gca);
rendered_img = frame.cdata;
gray_intensity = rgb2gray(rendered_img);

% Get axis limits to map image coordinates to data coordinates
[x_min, x_max] = xlim(gca);
[y_min, y_max] = ylim(gca);

% Create image coordinate grid
[img_cols, img_rows] = meshgrid(1:size(rendered_img, 2), 1:size(rendered_img, 1));

% Normalize image coordinates to match your surface's X/Y range
norm_x = (img_cols - 1)/(size(rendered_img, 2)-1)*(x_max - x_min) + x_min;
norm_y = (img_rows - 1)/(size(rendered_img, 1)-1)*(y_max - y_min) + y_min;

% Interpolate the intensity values to your original 161x161 grid
shaded_intensity_matrix = interp2(norm_x, norm_y, gray_intensity, X, Y);

Step 4: Save the Matrix #

Store the final intensity matrix for your photometric stereo work:

% Save as MAT file (preserves matrix dimensions and precision)
save('photometric_stereo_intensity.mat', 'shaded_intensity_matrix');

Key Notes #

• Shading Mode: Always use shading interp if you want continuous intensity values—flat or faceted will give blocky results that aren't useful for photometric stereo.
• Lighting Consistency: Make sure the light parameters (position, color, style) exactly match what you used in your original plot. This ensures the captured intensity matches your intended lighting setup.
• View Alignment: view(2) gives a true top-down orthographic view, which is essential for photometric stereo (avoid perspective distortion). Pair it with axis tight to ensure no empty space in the captured image.

内容的提问来源于stack exchange，提问作者SabreSword