如何使用MATLAB查看OV7670输出的YUV422单帧图像数据？

Hey there! Since you've already got a Java program up and running to visualize your OV7670's YUV422 output, let's get MATLAB to do the same for your single-frame data. Here's a step-by-step guide tailored to your setup:

Step 1: Load Your Raw YUV422 Data #

First, we need to turn that space-separated numeric data (like "128 74 83 93...") into a usable MATLAB array.

• If your data is saved in a text file:

  % Replace 'ov7670_frame.txt' with your actual file path
  raw_data = readmatrix('ov7670_frame.txt', 'Delimiter', ' ');
  

• If you're pasting the data directly as a string:

  raw_data = str2num('128 74 83 93...'); % Paste your full data string here
  

Step 2: Parse YUV422 into Y, U, V Components #

OV7670 typically outputs YUV422 in YUYV format, which follows this repeating pattern:
Y0 U0 Y1 V0 Y2 U1 Y3 V1 ...
Every 4 values correspond to 2 Y pixels, 1 U, and 1 V component.

Let's split the raw data into separate channels:

% Reshape data into a column vector for easy indexing
yuv_col = reshape(raw_data, [], 1);

% Extract individual components
Y = yuv_col(1:2:end);       % All odd indices: Y0, Y1, Y2, Y3...
U = yuv_col(2:4:end);       % Indices 2,6,10...: U0, U1, U2...
V = yuv_col(4:4:end);       % Indices 4,8,12...: V0, V1, V2...

Step 3: Reshape and Upsample to Match Image Resolution #

OV7670's common resolutions are 640x480 or 320x240—replace the values below with your sensor's actual output size:

img_width = 640;  % Adjust to your sensor's width
img_height = 480; % Adjust to your sensor's height

% Reshape Y to full image dimensions (transpose fixes MATLAB's column-major order)
Y_matrix = reshape(Y, img_width, img_height)';

% Reshape U/V (they're half the width of Y) then upsample to match Y's size
U_matrix = reshape(U, img_width/2, img_height)';
U_matrix = imresize(U_matrix, 2, 'nearest'); % Upsample 2x to match Y resolution

V_matrix = reshape(V, img_width/2, img_height)';
V_matrix = imresize(V_matrix, 2, 'nearest');

Step 4: Convert YUV to RGB and Display #

MATLAB's image tools work best with RGB, so let's convert and show the frame:

% Combine Y/U/V into a 3-channel matrix (cast to uint8 since sensor outputs 0-255 values)
YUV_matrix = cat(3, uint8(Y_matrix), uint8(U_matrix), uint8(V_matrix));

% Convert YUV to RGB
RGB_matrix = yuv2rgb(YUV_matrix);

% Display the final image
imshow(RGB_matrix);
title('OV7670 YUV422 Frame (Converted to RGB)');

Quick Troubleshooting Tips #

• Tinted/wrong colors? Your sensor might be using UYVY format instead of YUYV. Adjust component extraction like this:

  U = yuv_col(1:4:end);
  Y_part1 = yuv_col(2:4:end);
  Y_part2 = yuv_col(4:4:end);
  Y = [Y_part1; Y_part2]; % Merge two Y segments into one full vector
  V = yuv_col(3:4:end);
  

• Distorted image? Double-check that img_width and img_height match your sensor's output—incorrect resolution will mess up the image layout during reshaping.

内容的提问来源于stack exchange，提问作者Jiji Gaming