给点云添加不同类型的噪声

最新推荐文章于 2024-04-29 23:28:15 发布

Meet_csdn

最新推荐文章于 2024-04-29 23:28:15 发布

阅读量1.2k

点赞数 2

分类专栏：防忘记文章标签：计算机视觉

本文链接：https://blog.csdn.net/Meet_csdn/article/details/126736397

版权

1.对于点云，首先要归一化，即将点云最大值归一化为1，matlab代码如下：

path = which('normalization');
path = [path(1:end - size('normalization.m',2)) './'];
fpath       =   fullfile(path, '*.ply');
im_dir      =   dir(fpath);
im_num      =   length(im_dir);
for i=1:im_num
    ptr = pcread(fullfile(path, im_dir(i).name));
    center_x = ptr.XLimits(1) + (ptr.XLimits(2) -ptr.XLimits(1))/2;
    center_y = ptr.YLimits(1) + (ptr.YLimits(2) -ptr.YLimits(1))/2;
    center_z = ptr.ZLimits(1) + (ptr.ZLimits(2) -ptr.ZLimits(1))/2;
    center = [center_x center_y center_z];
    location = bsxfun(@minus,ptr.Location, center);%（0,0,0）中心在最中间
    new_ptr = pointCloud(location);
    max_XYZ = max([new_ptr.XLimits new_ptr.YLimits new_ptr.ZLimits]);%除以XYZ中最大值，保证不变形
    new_ptr = pointCloud(location./max_XYZ,'Normal',ptr.Normal);
%     pcshow(new_ptr);
    pcwrite(new_ptr,[im_dir(i).name(1:end - size('.pcd',2)) '.ply']);
end

2.归一化后即可添加不同类型（如高斯，指数，伽马等类型）的噪声，matlab代码如下：

%给模型添加标准高斯噪声
% path = './gd_test/ply_normal';
% path = 'C:\Users\Administrator\Desktop\pointCloudGMM-all\point-cloud-GMM v0_relese\Test_model\20_PLY_GT';
path = './';
noise_level_all = [0.01];
fpath       =   fullfile(path, '*.ply');
im_dir      =   dir(fpath);
im_num      =   length(im_dir);
% r2 = random('Normal',0,2);
n