i/rtabmap
1
0
Fork 0
mirror of https://github.com/introlab/rtabmap.git synced 2026-10-06 01:57:45 +08:00
Code Issues Packages Projects Releases Wiki Activity

Added GFTT detector which can be used for Odometry

Browse Source 
Modified how depth is computed: Now the mean of neighbors is used as the value instead of interpolation... averaging the Kinect noise.

git-svn-id: http://rtabmap.googlecode.com/svn/trunk/rtabmap@1662 f169173b-cf89-36c8-b27e-44dbe73f0c83
This commit is contained in:
matlabbe
2014-08-20 23:05:06 +00:00
parent b0af55699d
commit f251b64310
11 changed files with 546 additions and 113 deletions
Download Patch File Download Diff File Expand all files Collapse all files
corelib/src/Odometry.cpp
+66 -16
View File
@@ -88,9 +88,15 @@ void Odometry::reset()
bool Odometry::isLargeEnoughTransform(const Transform & transform)
{
return fabs(transform.x()) > _linearUpdate ||
fabs(transform.y()) > _linearUpdate ||
fabs(transform.z()) > _linearUpdate;
float x,y,z, roll,pitch,yaw;
transform.getTranslationAndEulerAngles(x,y,z, roll,pitch,yaw);
return (_linearUpdate == 0.0f && _angularUpdate == 0.0f) ||
fabs(x) > _linearUpdate ||
fabs(y) > _linearUpdate ||
fabs(z) > _linearUpdate ||
fabs(roll) > _angularUpdate ||
fabs(pitch) > _angularUpdate ||
fabs(yaw) > _angularUpdate;
}
Transform Odometry::process(SensorData & data, int * quality)
@@ -147,7 +153,8 @@ OdometryBOW::OdometryBOW(const ParametersMap & parameters) :
group.compare("BRIEF") == 0 ||
group.compare("FAST") == 0 ||
group.compare("ORB") == 0 ||
group.compare("FREAK") == 0)
group.compare("FREAK") == 0 ||
group.compare("GFTT") == 0)
{
customParameters.insert(*iter);
}
@@ -173,6 +180,34 @@ void OdometryBOW::reset()
localMap_.clear();
}
std::multimap<int,pcl::PointXYZ> OdometryBOW::getLocalMeansMap() const
{
std::multimap<int,pcl::PointXYZ> localMeansMap;
for(std::multimap<int, std::pair<int, pcl::PointXYZ> >::const_iterator iter=localMap_.begin();
iter!= localMap_.end();)
{
int id = iter->first;
pcl::PointXYZ sumPt = iter->second.second;
int count = 1;
++iter;
std::multimap<int, std::pair<int, pcl::PointXYZ> >::const_iterator jter=iter;
while(jter->first == id && jter!= localMap_.end())
{
sumPt.x += jter->second.second.x;
sumPt.y += jter->second.second.y;
sumPt.z += jter->second.second.z;
++count;
++jter;
}
iter = jter;
sumPt.x /= float(count);
sumPt.y /= float(count);
sumPt.z /= float(count);
localMeansMap.insert(std::make_pair(id, sumPt));
}
return localMeansMap;
}
// return not null transform if odometry is correctly computed
Transform OdometryBOW::computeTransform(const SensorData & data, int * quality)
@@ -207,11 +242,14 @@ Transform OdometryBOW::computeTransform(const SensorData & data, int * quality)
pcl::PointCloud<pcl::PointXYZ>::Ptr inliers1(new pcl::PointCloud<pcl::PointXYZ>); // previous
pcl::PointCloud<pcl::PointXYZ>::Ptr inliers2(new pcl::PointCloud<pcl::PointXYZ>); // new
//Create the local map with mean of all features
std::multimap<int, pcl::PointXYZ> localMeansMap = getLocalMeansMap();
// No need to set max depth here, it is already applied in extractKeypointsAndDescriptors() above.
// Also! the localMap_ have points not in camera frame anymore (in local map frame), so filtering
// by depth here is wrong!
util3d::findCorrespondences(
localMap_,
localMeansMap,
newSignature->getWords3(),
*inliers1,
*inliers2,
@@ -274,7 +312,7 @@ Transform OdometryBOW::computeTransform(const SensorData & data, int * quality)
if(pcl::isFinite(pt))
{
pcl::PointXYZ pt2 = util3d::transformPoint(pt, transform);
localMap_.insert(std::make_pair<int, pcl::PointXYZ>(*iter, pt2));
localMap_.insert(std::make_pair(*iter, std::make_pair(newSignature->id(), pt2)));
}
}
}
@@ -290,25 +328,35 @@ Transform OdometryBOW::computeTransform(const SensorData & data, int * quality)
std::list<int> uniques = uUniqueKeys(newSignature->getWords3());
for(std::list<int>::iterator iter = uniques.begin(); iter!=uniques.end(); ++iter)
{
if(newSignature->getWords3().count(*iter) == 1 &&
uniqueCorrespondences.find(*iter) == uniqueCorrespondences.end())
if(newSignature->getWords3().count(*iter) == 1)
{
const pcl::PointXYZ & pt = newSignature->getWords3().find(*iter)->second;
if(pcl::isFinite(pt))
{
pcl::PointXYZ pt2 = util3d::transformPoint(pt, transform);
localMap_.insert(std::make_pair<int, pcl::PointXYZ>(*iter, pt2));
localMap_.insert(std::make_pair(*iter, std::make_pair(newSignature->id(), pt2)));
}
}
}
while(localMap_.size() && (int)localMap_.size() > this->getLocalHistory() && _memory->getStMem().size()>1)
while(localMap_.size() && (int)uUniqueKeys(localMap_).size() > this->getLocalHistory() && _memory->getStMem().size()>1)
{
std::list<int> deletedWords;
_memory->deleteLocation(*_memory->getStMem().begin(), &deletedWords);
for(std::list<int>::iterator iter = deletedWords.begin(); iter!=deletedWords.end(); ++iter)
int nodeId = *_memory->getStMem().begin();
std::list<int> removedPts = uUniqueKeys(_memory->getSignature(nodeId)->getWords3());
for(std::list<int>::iterator iter = removedPts.begin(); iter!=removedPts.end(); ++iter)
{
localMap_.erase(*iter);
bool removed = false;
for(std::multimap<int, std::pair<int, pcl::PointXYZ> >::iterator jter=localMap_.lower_bound(*iter);
jter->first == *iter && !removed;
++jter)
{
if(jter->second.first == nodeId)
{
localMap_.erase(jter);
removed = true;
}
}
}
_memory->deleteLocation(*_memory->getStMem().begin());
}
}
else
@@ -331,7 +379,7 @@ Transform OdometryBOW::computeTransform(const SensorData & data, int * quality)
const pcl::PointXYZ & pt = newSignature->getWords3().find(*iter)->second;
if(pcl::isFinite(pt))
{
localMap_.insert(std::make_pair<int, pcl::PointXYZ>(*iter, pt));
localMap_.insert(std::make_pair(*iter, std::make_pair(newSignature->id(), pt)));
}
}
}
@@ -340,11 +388,13 @@ Transform OdometryBOW::computeTransform(const SensorData & data, int * quality)
_memory->emptyTrash();
}
UINFO("Odom update time = %fs features=%d inliers=%d/%d dict=%d nodes=%d",
UINFO("Odom update time = %fs features=%d inliers=%d/%d local_map=%d[%d] dict=%d nodes=%d",
timer.elapsed(),
nFeatures,
inliers,
correspondences,
(int)uUniqueKeys(localMap_).size(),
(int)localMap_.size(),
(int)_memory->getVWDictionary()->getVisualWords().size(),
(int)_memory->getStMem().size());
return output;