how to extract features from 350 clusters | extractEigenFeatures

2 次查看（过去 30 天）

Pavan Kumar B N 2021-10-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1546426-how-to-extract-features-from-350-clusters-extracteigenfeatures

编辑： Pavan Kumar B N 2021-10-11

Hello All,

I have around 350 clusters and I am trying to retain only matching clusters and remove unmatched ones. How to use this example in this case https://www.mathworks.com/help/lidar/ref/pcshowmatchedfeatures.html

2 个评论
显示无隐藏无

Venkata Ram Prasad Vanguri 2021-10-11

Hi,

We can extract matched cluster segments and corresponding features from the matched index pairs as below:

matchedSegments1 = segments1(indexPairs(:,1));

matchedSegments2 = segments2(indexPairs(:,2));

matchedFeatures1 = eigFeatures1(indexPairs(:,1));

matchedFeatures2 = eigFeatures2(indexPairs(:,2));

Pavan Kumar B N 2021-10-11

编辑：Pavan Kumar B N 2021-10-11

在 MATLAB Online 中打开

@Venkata Ram Prasad Vanguri Thank you so much for your response.

I tried as you suggested on pcap data. When I run the following code it is generating 50 figures of matched segments. I tired visualize these results using pcplayer. But it is giving a error like "Too many input arguments". Could you please help on how to visulize these matching features using pcplayer.?

Eagerly looking forward for your kind help.

veloReader = velodyneFileReader('test.pcap', 'VLP16');
for i = 1:50
    ptCloud1 = veloReader.readFrame(i);
    ptCloud2 = veloReader.readFrame(i+1);
    %remove the ground plane from the scans
    maxDistance = 1;
    referenceVector = [0 0 1];
    [~,~,selectIdx] = pcfitplane(ptCloud1, maxDistance, referenceVector); 
    ptCloud1 = select(ptCloud1,selectIdx, 'OutputSize', 'full');
    [~,~,selectIdx] = pcfitplane(ptCloud2, maxDistance, referenceVector);
    ptCloud2 = select(ptCloud2,selectIdx, 'OutputSize', 'full');
    %cluster the point cloud with 10 points per cluster
    minDistance = 2;
    minPoints = 10;
    lables1 = pcsegdist(ptCloud1, minDistance, 'NumClusterPoints', minPoints);
    lables2 = pcsegdist(ptCloud2, minDistance, 'NumClusterPoints', minPoints);
    [eigFeatures1, segments1] = extractEigenFeatures(ptCloud1,lables1);
    [eigFeatures2, segments2] = extractEigenFeatures(ptCloud2,lables2);
    %create matrices of the features and centroids extracted from each point
    %cloud
    features1 = vertcat(eigFeatures1.Feature);
    features2 = vertcat(eigFeatures2.Feature);
    centroids1 = vertcat(eigFeatures1.Centroid);
    centroids2 = vertcat(eigFeatures2.Centroid);
    %find future matches
    indexPairs = pcmatchfeatures(features1,features2, pointCloud(centroids1),pointCloud(centroids2),'MatchThreshold',0.1,'RejectRatio',0.9);
    matchedSegments1 = segments1(indexPairs(:,1));
    matchedSegments2 = segments2(indexPairs(:,2));
    matchedFeatures1 = eigFeatures1(indexPairs(:,1));
    matchedFeatures2 = eigFeatures2(indexPairs(:,2));
    figure
    pcshowMatchedFeatures(matchedSegments1,matchedSegments2,matchedFeatures1,matchedFeatures2)
    title('Matched Segments')
    
    xlimits = [-35 118];
    ylimits = [-55 75];
    zlimits = [-10 10];
    
    player = pcplayer(xlimits,ylimits,zlimits);
    %while(hasFrame(ptCloudPlayer))
    % ptCloudObj = readFrame(ptCloudPlayer);
    view(player,matchedSegments2.Location,matchedSegments2.Intensity);
    %end    
end