jureviciusr / particle-match Goto Github PK

Hello,

I enabled CUDA
CMakeLists.txt
option(USE_OPENCV_CUDA= "Enable this flag to compile features requiring GPU support" ON)

ParticuleFastMatch.cpp
//cv::Mat bestView = Utilities::extractWarpedMapPart(imageGray, templ.size(), transform);
cv::cuda::GpuMat bestView = Utilities::extractWarpedMapPart(imageGrayGpu,

But it does not seem to use any GPU:

Hi,

I build on UBUNTU 20 with OpenCV version: 3.4.14
and I had to include these headers

ParticuleFastMatch.cpp  
#include <fstream>

Utilities.cpp
#include <random>

I am trying to reproduce the results in "Robust GNSS Denied Localization for UAV Using Particle Filter and Visual Odometry" and for Visual Odometry, I need to know what were the parameters of the camera used in PX4 simulation for producing the aerial images?

Hi,
I really appreciate the research and code, I think I have learned a lot about visual odometry, particle filter and computer vision concepts.This was very insightful for me but I have something on my mind: I would like to go further and try this project with custom map and data. However, when I try to change the .tif file with another .tif example which I have found, I got some errors. How can I basically customize this project for my map area and image data? Would you like to summarize generally the steps for this custom configuration? (I am aware of that I will get inconsistent result since simulations has done with different maps, but I just want to change the map.)