在這裡，對matlab的demo進行更改，測試了雙目視覺！具體的**如下：
%%
% 雙目立體視覺
% 對比實驗
%%% 清空工作區
clc;
clear;
close all;
%%% 匯入影象資料
i1 = imread('pattern_cam1_im1.png');
i2 = imread('pattern_cam2_im1.png');
figure
imshowpair(i1, i2, 'montage');
title('original images');
% 匯入相機引數
load cameraparams.mat
%%% 校正
i1 = undistortimage(i1, camera1params);
i2 = undistortimage(i2, camera1params);
figure
imshowpair(i1, i2, 'montage');
title('undistorted images');
%%% 特徵點提取
imagepoints1 = detectmineigenfeatures(rgb2gray(i1), 'minquality', 0.1);
%%% 視覺化
figure
imshow(i1, 'initialmagnification', 50);
title('150 strongest corners from the first image');
hold on
plot(selectstrongest(imagepoints1, 150));
%%% create the point tracker
tracker = vision.pointtracker('maxbidirectionalerror', 1, 'numpyramidlevels', 5);
imagepoints1 = imagepoints1.location;
initialize(tracker, imagepoints1, i1);
% track the points
[imagepoints2, valididx] = step(tracker, i2);
matchedpoints1 = imagepoints1(valididx, :);
matchedpoints2 = imagepoints2(valididx, :);
%%% 特徵點匹配
figure
showmatchedfeatures(i1, i2, matchedpoints1, matchedpoints2);
title('tracked features');
%%% f矩陣估計
[fmatrix, epipolarinliers] = estimatefundamentalmatrix(...
matchedpoints1, matchedpoints2, 'method', 'msac', 'numtrials', 10000);
% 極線
inlierpoints1 = matchedpoints1(epipolarinliers, :);
inlierpoints2 = matchedpoints2(epipolarinliers, :);
% 顯示內點
figure
showmatchedfeatures(i1, i2, inlierpoints1, inlierpoints2);
title('epipolar inliers');
%%% r和t（也可以用ransac演算法）
r = [0.9455,-0.0096,0.3253;
0.0120,0.9999,-0.0053;
-0.3252,0.0090,0.9456];
t = [98.4069,0.1741,18.9018];
%%% 稠密的特徵點
imagepoints1 = detectmineigenfeatures(rgb2gray(i1), 'minquality', 0.001);
%%% create the point tracker
tracker = vision.pointtracker('maxbidirectionalerror', 1, 'numpyramidlevels', 5);
% initialize the point tracker
imagepoints1 = imagepoints1.location;
initialize(tracker, imagepoints1, i1);
% track the points
[imagepoints2, valididx] = step(tracker, i2);
matchedpoints1 = imagepoints1(valididx, :);
matchedpoints2 = imagepoints2(valididx, :);
%%% cameramatrix
cammatrix1 = cameramatrix(camera1params, eye(3), [0 0 0]);
cammatrix2 = cameramatrix(camera2params, r', -t*r');
% 三維點雲的計算
points3d = triangulate(matchedpoints1, matchedpoints2, cammatrix1, cammatrix2);
% 獲取顏色資訊
numpixels = size(i1, 1) * size(i1, 2);
allcolors = reshape(i1, [numpixels, 3]);
coloridx = sub2ind([size(i1, 1), size(i1, 2)], round(matchedpoints1(:,2)), ...
round(matchedpoints1(:, 1)));
color = allcolors(coloridx, :);
% 建立點雲
ptcloud = pointcloud(points3d, 'color', color);
%%% 視覺化
camerasize = 0.3;
figure
plotcamera('size', camerasize, 'color', 'r', 'label', '1', 'opacity', 0);
hold on
grid on
plotcamera('location', t, 'orientation', r, 'size', camerasize, ...
'color', 'b', 'label', '2', 'opacity', 0);
% 點雲的視覺化
pcshow(ptcloud, 'verticalaxis', 'y', 'verticalaxisdir', 'down', ...
'markersize', 45);
% rotate and zoom the plot
camorbit(0, -30);
camzoom(1.5);
% label the axes
xlabel('x-axis');
ylabel('y-axis');
zlabel('z-axis')
title('up to scale reconstruction of the scene');
第五講5 4 1雙目視覺生成點雲
cmake minimum required version 3.15 project stereovision find package pangolin required find package opencv required include directories include direc...
基於雙目視覺的汽車前向障礙物檢測
在我的設計中 1 測試案例中的左圖 左上圖 和右圖使用自己改進的sgm演算法進行計算得到視差圖，通過gpu的加速可以實時輸出視差圖 右中圖為視差圖的偽色彩圖 可以看到，改進版的sgm演算法不但速度滿足了實時計算的要求，而且效果明顯好於原版的sgm演算法。2 通過對視差資訊的計算，獲得三維點雲圖 右下...
基於結構光的立體視覺
1介紹 對立體視覺進行介紹，分析現有的缺陷。如何引進結構光，測量的原理和標定，效能測試與評估。二維的影象無法獲取深度資訊。雙目立體視覺 特徵提取 sift,suft,orb。特徵描述 特徵匹配 犧牲特徵點的數量。特徵點數目比較少，比較稀疏。無法實現識別。結構光視覺 視覺範圍 小於5m，精度 最大0....