MATLAB中Hough Transform绘图不符预期，如何调整设置实现目标效果？

Hey there! Let's work through getting your Hough transform plot looking exactly how you want it. I’ve dealt with this exact head-scratcher before when plotting Hough accumulators in MATLAB, so here are the key tweaks to fix up your current code:

Key Adjustments to Try #

• Fix the color mapping & intensity range
  The default grayscale colormap often makes Hough peaks hard to spot, and unadjusted intensity limits can flatten out important details. Add these lines right after imshow:

  colormap(gca, 'jet'); % Swap to a high-contrast colormap (hot, parula also work)
  clim([0 max(h(:))]); % Lock color range to match your accumulator's min/max values
  

• Tweak axis settings for clarity
  axis normal can sometimes stretch or misalign your theta/rho axes. Replace it with:

  axis tight; % Removes extra whitespace around the accumulator
  xlabel('Theta (degrees)');
  ylabel('Rho (pixels)');
  

  This ensures your axes fit the exact range of theta and rho values, plus adds clear labels to make the plot interpretable.

• Optional: Mark Hough peaks (if your target style includes this)
  If you want to highlight the strongest lines detected, add peak markers after enabling hold on:

  peaks = houghpeaks(h, 5); % Grab top 5 peaks (adjust number as needed)
  plot(theta(peaks(:,2)), rho(peaks(:,1)), 'rs', 'MarkerSize', 8, 'MarkerFaceColor', 'r');
  

Full Updated Code Example #

Here’s how all these pieces fit together:

% Assuming you already ran: [h, theta, rho] = hough(edge_detected_image);
figure;
imshow(h, [], 'XData', theta, 'YData', rho, 'InitialMagnification', 'fit');
colormap(gca, 'jet');
clim([0 max(h(:))]);
axis on;
axis tight;
xlabel('Theta (degrees)');
ylabel('Rho (pixels)');
title('Hough Transform Accumulator');
hold on;

% Add peak markers (optional)
peaks = houghpeaks(h, 5);
plot(theta(peaks(:,2)), rho(peaks(:,1)), 'rs', 'MarkerSize', 8, 'MarkerFaceColor', 'r');
hold off;

These changes should make your Hough accumulator plot match the typical, easy-to-read style—with clear peaks, properly aligned axes, and intuitive color coding.

内容的提问来源于stack exchange，提问作者Ravi Shankar