i/rtabmap
1
0
Fork 0
mirror of https://github.com/introlab/rtabmap.git synced 2026-09-03 01:50:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

SuperPoint: updated descriptor normalization process to match exactly the implementation in SuperGlue. generateWords3DMono: use variance at 25% instead of 50% of the sqrd variance vector (to do the same than with PnP), also fixed return camera transform to include local transform of the camera model. rtabmap-matcher: updated default parameters when no depth images are used, added reproj error in title bar.

Browse Source
This commit is contained in:
matlabbe
2020-05-10 14:13:41 -04:00
parent fbdc3780c7
commit cc9fd439fe
4 changed files with 35 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
corelib/src/superpoint_torch/SuperPoint.cc
View File

@@ -218,30 +218,35 @@ cv::Mat SPDetector::compute(const std::vector<cv::KeyPoint> &keypoints)
{
cv::Mat kpt_mat(keypoints.size(), 2, CV_32F); // [n_keypoints, 2] (y, x)
// Based on sample_descriptors() of SuperPoint implementation in SuperGlue:
// https://github.com/magicleap/SuperGluePretrainedNetwork/blob/45a750e5707696da49472f1cad35b0b203325417/models/superpoint.py#L80-L92
float s = 8;
for (size_t i = 0; i < keypoints.size(); i++) {
kpt_mat.at<float>(i, 0) = (float)keypoints[i].pt.y;
kpt_mat.at<float>(i, 1) = (float)keypoints[i].pt.x;
kpt_mat.at<float>(i, 0) = (float)keypoints[i].pt.y - s/2 + 0.5;
kpt_mat.at<float>(i, 1) = (float)keypoints[i].pt.x - s/2 + 0.5;
}
auto fkpts = torch::from_blob(kpt_mat.data, {(long int)keypoints.size(), 2}, torch::kFloat);
float w = desc_.size(3); //W/8
float h = desc_.size(2); //H/8
torch::Device device(cuda_?torch::kCUDA:torch::kCPU);
auto grid = torch::zeros({1, 1, fkpts.size(0), 2}).to(device); // [1, 1, n_keypoints, 2]
grid[0][0].slice(1, 0, 1) = 2.0 * fkpts.slice(1, 1, 2) / prob_.size(1) - 1; // x
grid[0][0].slice(1, 1, 2) = 2.0 * fkpts.slice(1, 0, 1) / prob_.size(0) - 1; // y
grid[0][0].slice(1, 0, 1) = 2.0 * fkpts.slice(1, 1, 2) / (w*s - s/2 - 0.5) - 1; // x
grid[0][0].slice(1, 1, 2) = 2.0 * fkpts.slice(1, 0, 1) / (h*s - s/2 - 0.5) - 1; // y
auto desc = torch::grid_sampler(desc_, grid, 0, 0, true); // [1, 256, 1, n_keypoints]
desc = desc.squeeze(0).squeeze(1); // [256, n_keypoints]
// normalize to 1
auto dn = torch::norm(desc, 2, 1);
desc = desc.div(torch::unsqueeze(dn, 1));
desc = torch::nn::functional::normalize(desc.reshape({1, desc_.size(1), -1})); //[1, 256, n_keypoints]
desc = desc.squeeze(); //[256, n_keypoints]
desc = desc.transpose(0, 1).contiguous(); //[n_keypoints, 256]
desc = desc.transpose(0, 1).contiguous(); // [n_keypoints, 256]
if(cuda_)
desc = desc.to(torch::kCPU);
cv::Mat desc_mat(cv::Size(desc.size(1), desc.size(0)), CV_32FC1, desc.data<float>());
cv::Mat desc_mat(cv::Size(desc.size(1), desc.size(0)), CV_32FC1, desc.data_ptr<float>());
return desc_mat.clone();
}