MATLAB 视频测平面移动距离 图像二值化 特征点提取 代码实现

以下是基于 MATLAB 的视频测平面移动距离、图像二值化和特征点提取的代码实现：

1. 视频读取和显示

% 读取视频
v = VideoReader('video.mp4');

% 显示第一帧
frame = readFrame(v);
imshow(frame);

2. 图像二值化

% 将图像转换为灰度图
gray_frame = rgb2gray(frame);

% 图像二值化
threshold = graythresh(gray_frame);
bw_frame = imbinarize(gray_frame, threshold);
imshow(bw_frame);

3. 特征点提取

% 使用 SURF 算法提取特征点和特征描述符
points = detectSURFFeatures(gray_frame);
[features,valid_points] = extractFeatures(gray_frame,points);

% 可视化特征点
imshow(gray_frame); hold on;
plot(valid_points.selectStrongest(50),'showOrientation',true);

4. 平面移动距离测量

% 循环读取视频帧
while hasFrame(v)
    % 读取一帧
    frame = readFrame(v);
    gray_frame = rgb2gray(frame);

    % 计算当前帧与上一帧的光流向量
    flow = estimateFlow(opticFlow, gray_frame);
    dx = mean(flow.Vx(:));
    dy = mean(flow.Vy(:));

    % 计算平面移动距离
    distance = sqrt(dx^2 + dy^2);
    total_distance = total_distance + distance;

    % 显示当前帧和移动距离
    imshow(frame);
    title(sprintf('Total distance: %.2f', total_distance));
    drawnow;

    % 更新上一帧图像
    prev_gray_frame = gray_frame;
end

完整代码参考：

% 读取视频
v = VideoReader('video.mp4');

% 显示第一帧
frame = readFrame(v);
imshow(frame);

% 将第一帧转换为灰度图
gray_frame = rgb2gray(frame);

% 创建光流估计器
opticFlow = opticalFlowLK;

% 初始化平面移动距离
total_distance = 0;

% 循环读取视频帧
while hasFrame(v)
    % 读取一帧
    frame = readFrame(v);
    gray_frame = rgb2gray(frame);

    % 计算当前帧与上一帧的光流向量
    flow = estimateFlow(opticFlow, gray_frame);
    dx = mean(flow.Vx(:));
    dy = mean(flow.Vy(:));

    % 计算平面移动距离
    distance = sqrt(dx^2 + dy^2);
    total_distance = total_distance + distance;

    % 显示当前帧和移动距离
    imshow(frame);
    title(sprintf('Total distance: %.2f', total_distance));
    drawnow;

    % 更新上一帧图像
    prev_gray_frame = gray_frame;
end

% 使用 SURF 算法提取特征点和特征描述符
points = detectSURFFeatures(gray_frame);
[features,valid_points] = extractFeatures(gray_frame,points);

% 可视化特征点
imshow(gray_frame); hold on;
plot(valid_points.selectStrongest(50),'showOrientation',true);

% 图像二值化
threshold = graythresh(gray_frame);
bw_frame = imbinarize(gray_frame, threshold);
imshow(bw_frame);