Thanks for the great sharing.
Before porting to android system, I am trying to port VINS-Mobile to linux system and make some optimization and testing. But I face some problem on reconstruction.
I changed the top img_msg->header = systemtime(my own timestamp).
Q: Is it any problem to change the time_stamp from
float lowPart = image.at(0,0);
float highPart = image.at(0,1);
to common system time? [system time updating every 0.01sec on my own thread]
I gave a virtual IMU data which is random and update every 0.01sec. Does it cause any problem?
Now what i'm facing is that, the feature detection is working now, but the ID is always updating, and the track_num and f_manager.getFeatureCount() is always get 0.
I think it is problem on f_manager.addFeatureCheckParallax() or solveVinsPnP(), because the Ps, Vs, Rs, Bas and Bgs is also 0. Do anybody have idea....
Thanks for the great sharing. Before porting to android system, I am trying to port VINS-Mobile to linux system and make some optimization and testing. But I face some problem on reconstruction.
I changed the top img_msg->header = systemtime(my own timestamp). Q: Is it any problem to change the time_stamp from float lowPart = image.at(0,0);(0,1);
to common system time? [system time updating every 0.01sec on my own thread]
float highPart = image.at
I gave a virtual IMU data which is random and update every 0.01sec. Does it cause any problem?
Now what i'm facing is that, the feature detection is working now, but the ID is always updating, and the track_num and f_manager.getFeatureCount() is always get 0. I think it is problem on f_manager.addFeatureCheckParallax() or solveVinsPnP(), because the Ps, Vs, Rs, Bas and Bgs is also 0. Do anybody have idea....