openvslam (1) 运行和增大跟踪效果
执行命令
0 程序编译 mkdir build && cd build cmake \-DBUILD_WITH_MARCH_NATIVE=ON \-DUSE_PANGOLIN_VIEWER=ON \-DUSE_SOCKET_PUBLISHER=OFF \-DUSE_STACK_TRACE_LOGGER=ON \-DBOW_FRAMEWORK=DBoW2 \-DBUILD_EXAMPLES=ON \.. make -j41 运行SLAM接收图像开启./run_image_slam -h 输出============================================== Allowed options:-h, --help produce help message-v, --vocab arg vocabulary file path-i, --img-dir arg directory path which contains images-c, --config arg config file path--mask arg mask image path--frame-skip arg (=1) interval of frame skip--no-sleep not wait for next frame in real time--auto-term automatically terminate the viewer--debug debug mode--eval-log store trajectory and tracking times for evaluation-p, --map-db arg store a map database at this path after SLAM==================================================# ./run_image_slam_mydata \ -c /media/dongdong/Elements1/项目/ehang_/2TestData/2TestData/ehang4/openvslam_config/GNSS_config.yaml图像文件夹建立地图 ./run_image_slam_resize \ -v /media/dongdong/新加卷/0ubuntu20/1slam/数据/2RTK/hf1_youtian/hunhe/slam_config/orb_vocab.dbow2 \ -i /media/dongdong/新加卷/0ubuntu20/1slam/数据/2RTK/hf1_youtian/hunhe/images \ -c /media/dongdong/新加卷/0ubuntu20/1slam/数据/2RTK/hf1_youtian/hunhe/slam_config/GNSS_config.yaml \ --map-db /media/dongdong/新加卷/0ubuntu20/1slam/数据/2RTK/hf1_youtian/hunhe/slam_config/MapResult.msg图像文件夹在线定位 ./run_image_localization \ -v /home/nuc/v2_Project/openvslam/v2_Data/data/orb_vocab.dbow2 \ -i /home/nuc/v2_Project/openvslam/v2_Data/data/img \ -c /home/nuc/v2_Project/openvslam/v2_Data/data/camera.yaml \ --map-db /home/nuc/v2_Project/openvslam/v2_Data/data/MapResult.msg测试视频建图 ./run_video_slam_resize \ -v ../data/orb_vocab.dbow2 \ -m /home/nuc/桌面/data/角度数据/采集数据/rtk学院楼门前畸变80米/DJI_0014.MOV \ -c ../data/camera_RTK_JIBIAN.yaml \ --frame-skip 1 \ --map-db ../data/DJI_00014_270.msg测试视频定位./run_video_localization_resize \ -v ../data/orb_vocab.dbow2 \ -m /home/nuc/桌面/data/角度数据/采集数据/rtk学院楼门前畸变80米/DJI_0007.MOV \ -c ../data/camera_RTK_JIBIAN.yaml \ --frame-skip 1 \ --map-db ../data/DJI_0000_00.msg \ --mapping
内参文件
关键点 提高跟踪
# Feature.max_num_keypoints: 3000 Feature.num_levels: 8 Feature.scale_factor: 1.2# ORB Parameters Feature.max_num_keypoints: 3000 # 大幅增加！默认1000不够用 # Feature.scale_factor: 1.4 # 增大金字塔尺度因子，适应大尺度变化 1.2 # Feature.num_levels: 10 # 增加金字塔层数，适应远景 8 # Feature.ini_fast_threshold: 10 # 降低初始阈值，在弱纹理区也能提点 # Feature.min_fast_threshold: 4 # 降低最小阈值，保证特征数量 Tracking.num_tracked_points: 15 # 降低跟踪点阈值（默认20），允许在特征少时继续跟踪 Tracking.min_num_tracked_points: 8 # 重定位/初始化最小点数，适当降低 Tracking.relocalization_threshold: 0.7 # 降低重定位相似度阈值，更容易触发重定位 Tracking.enable_temporal_keyframe_only: false # 确保插入正常关键帧，而非仅临时帧 # # Initialization (部分参数可能在代码中，需在初始化器里设置) # Initializer.max_num_ransac_iters: 200 # 增加RANSAC迭代次数 # Initializer.min_num_triangulated: 40 # 降低成功初始化所需的最小三角化点数 # Initializer.parallax_deg_thr: 0.5 # 降低视差阈值（度），适应平移小的俯拍 # Keyframe Insertion Keyframe.max_num_frms: 30 # 最大关键帧间隔（按帧数），可适当增大 Keyframe.min_num_frms: 1 # 最小间隔，防止插入过快 Keyframe.num_reliable_keypoints: 100 # 关键帧最小可靠点数，可适当降低
完整
Camera.color_order: RGB Camera.cols: 1805 Camera.cx: 910.8785687265895 Camera.cy: 602.174293145834 Camera.fps: 10 Camera.fx: 1193.7076128098686 Camera.fy: 1193.1735265967602 Camera.k1: 0 Camera.k2: 0 Camera.k3: 0 Camera.k4: 0 Camera.model: perspective Camera.name: NWPU monocular Camera.p1: 0 Camera.p2: 0 Camera.rows: 1203 Camera.setup: monocular Config_PATH: /home/dongdong/2project/0data/RTK/data_1_nwpuUp/data3_1130_13pm//300_location_14pm/ # Feature.max_num_keypoints: 3000 Feature.num_levels: 8 Feature.scale_factor: 1.2# ORB Parameters Feature.max_num_keypoints: 3000 # 大幅增加！默认1000不够用 # Feature.scale_factor: 1.4 # 增大金字塔尺度因子，适应大尺度变化 1.2 # Feature.num_levels: 10 # 增加金字塔层数，适应远景 8 # Feature.ini_fast_threshold: 10 # 降低初始阈值，在弱纹理区也能提点 # Feature.min_fast_threshold: 4 # 降低最小阈值，保证特征数量 Tracking.num_tracked_points: 15 # 降低跟踪点阈值（默认20），允许在特征少时继续跟踪 Tracking.min_num_tracked_points: 8 # 重定位/初始化最小点数，适当降低 Tracking.relocalization_threshold: 0.7 # 降低重定位相似度阈值，更容易触发重定位 Tracking.enable_temporal_keyframe_only: false # 确保插入正常关键帧，而非仅临时帧 # # Initialization (部分参数可能在代码中，需在初始化器里设置) # Initializer.max_num_ransac_iters: 200 # 增加RANSAC迭代次数 # Initializer.min_num_triangulated: 40 # 降低成功初始化所需的最小三角化点数 # Initializer.parallax_deg_thr: 0.5 # 降低视差阈值（度），适应平移小的俯拍 # Keyframe Insertion Keyframe.max_num_frms: 30 # 最大关键帧间隔（按帧数），可适当增大 Keyframe.min_num_frms: 1 # 最小间隔，防止插入过快 Keyframe.num_reliable_keypoints: 100 # 关键帧最小可靠点数，可适当降低Fixed.altitude: 400.0 Fixed.altitude_flag: 0 GNSS_Have: 1 GNSS_PSTH: slam_config/gnss.txt GNSS_USE: 1 Initial.alt: 415 Initial.lat: 34.033727638888884 Initial.lon: 108.76406197222222 Loop_Th: 80.0 Map.Ininterval: 10 Map.InsertFps: 20 Map.alt: 715 Marker.Num: 0 PangolinViewer.camera_line_width: 3 PangolinViewer.camera_size: 0.08 PangolinViewer.graph_line_width: 1 PangolinViewer.keyframe_line_width: 2 PangolinViewer.keyframe_size: 0.07 PangolinViewer.point_size: 2 PangolinViewer.viewpoint_f: 400 PangolinViewer.viewpoint_x: 0 PangolinViewer.viewpoint_y: -0.65 PangolinViewer.viewpoint_z: -1.9 Relocalize_KF_Th: 3.0 Relocalize_Th: 80.0 Save.data: 1 Save.newmap: 1 Tracking_CF_Th: 10.0 V_Instant_Th: 200.0 WorkMode: map_onlynewData WorkMode_1: map_all_keep_allKfms WorkMode_2: map_all WorkMode_3: location_only WorkMode_4: map_onlynewData hAcc: 1.0 image_dir_path: images image_dir_type: JPG map_db_path: slam_out/Map_GNSS.msg mask_img_path: none.jpg op.Global_Optimize_Th: 1.0 op.Remove_Kayframe_Th: 6.0 op.Second_Optimize_Th: 0 op.is_Second_Optimize: 0 sRt_colmap2gnss: sparse/srt_colmap2gnss.yaml sRt_slam2gnss: slam_out/srt_slam2gnss.yaml vAcc: 1.0 vocab_file_path: slam_config/orb_vocab.dbow2
主文件
#include "util/image_util.h"#ifdef USE_PANGOLIN_VIEWER
#include "pangolin_viewer/viewer.h"
#elif USE_SOCKET_PUBLISHER
#include "socket_publisher/publisher.h"
#endif#include "openvslam/system.h"
#include "openvslam/config.h"#include <iostream>
#include <chrono>
#include <numeric>#include <opencv2/core/core.hpp>
#include <opencv2/imgcodecs.hpp>
#include <spdlog/spdlog.h>
#include <popl.hpp>
#include<opencv2/opencv.hpp>using namespace std;
using namespace cv;#ifdef USE_STACK_TRACE_LOGGER
#include <glog/logging.h>
#endif#ifdef USE_GOOGLE_PERFTOOLS
#include <gperftools/profiler.h>
#endifvoid mono_tracking(const std::shared_ptr<openvslam::config>& cfg,const std::string& vocab_file_path, const std::string& image_dir_path, const std::string& mask_img_path,const unsigned int frame_skip, const bool no_sleep, const bool auto_term,const bool eval_log, const std::string& map_db_path) {// load the mask imageconst cv::Mat mask = mask_img_path.empty() ? cv::Mat{} : cv::imread(mask_img_path, cv::IMREAD_GRAYSCALE);const image_sequence sequence(image_dir_path, cfg->camera_->fps_);const auto frames = sequence.get_frames();// build a SLAM systemopenvslam::system SLAM(cfg, vocab_file_path);// startup the SLAM processSLAM.startup();#ifdef USE_PANGOLIN_VIEWERpangolin_viewer::viewer viewer(cfg, &SLAM, SLAM.get_frame_publisher(), SLAM.get_map_publisher());
#elif USE_SOCKET_PUBLISHERsocket_publisher::publisher publisher(cfg, &SLAM, SLAM.get_frame_publisher(), SLAM.get_map_publisher());
#endifstd::vector<double> track_times;track_times.reserve(frames.size());// run the SLAM in another threadstd::thread thread([&]() {for (unsigned int i = 0; i < frames.size(); ++i) {const auto& frame = frames.at(i);const auto img = cv::imread(frame.img_path_, cv::IMREAD_UNCHANGED);const auto tp_1 = std::chrono::steady_clock::now();if (!img.empty() && (i % frame_skip == 0)) {//要缩小图像，通常使用CV_INTER_AREA插值效果最好，而要放大图像，通常使用CV_INTER_CUBIC（慢）或CV_INTER_LINEAR（更快但看起来还可以）效果最好。cv::Size dsize = cv::Size(1805, 1203);cv::Mat imgResize;cv::resize(img, imgResize, dsize, 0, 0, INTER_AREA);// input the current frame and estimate the camera poseSLAM.feed_monocular_frame(imgResize, frame.timestamp_, mask);}const auto tp_2 = std::chrono::steady_clock::now();const auto track_time = std::chrono::duration_cast<std::chrono::duration<double>>(tp_2 - tp_1).count();if (i % frame_skip == 0) {track_times.push_back(track_time);}// wait until the timestamp of the next frameif (!no_sleep && i < frames.size() - 1) {const auto wait_time = frames.at(i + 1).timestamp_ - (frame.timestamp_ + track_time);if (0.0 < wait_time) {std::this_thread::sleep_for(std::chrono::microseconds(static_cast<unsigned int>(wait_time * 1e6)));}}// check if the termination of SLAM system is requested or notif (SLAM.terminate_is_requested()) {break;}}// wait until the loop BA is finishedwhile (SLAM.loop_BA_is_running()) {std::this_thread::sleep_for(std::chrono::microseconds(5000));}// automatically close the viewer
#ifdef USE_PANGOLIN_VIEWERif (auto_term) {viewer.request_terminate();}
#elif USE_SOCKET_PUBLISHERif (auto_term) {publisher.request_terminate();}
#endif});// run the viewer in the current thread
#ifdef USE_PANGOLIN_VIEWERviewer.run();
#elif USE_SOCKET_PUBLISHERpublisher.run();
#endifthread.join();// shutdown the SLAM processSLAM.shutdown();if (eval_log) {// output the trajectories for evaluationSLAM.save_frame_trajectory("frame_trajectory.txt", "TUM");SLAM.save_keyframe_trajectory("keyframe_trajectory.txt", "TUM");// output the tracking times for evaluationstd::ofstream ofs("track_times.txt", std::ios::out);if (ofs.is_open()) {for (const auto track_time : track_times) {ofs << track_time << std::endl;}ofs.close();}}if (!map_db_path.empty()) {// output the map databaseSLAM.save_map_database(map_db_path);}std::sort(track_times.begin(), track_times.end());const auto total_track_time = std::accumulate(track_times.begin(), track_times.end(), 0.0);std::cout << "median tracking time: " << track_times.at(track_times.size() / 2) << "[s]" << std::endl;std::cout << "mean tracking time: " << total_track_time / track_times.size() << "[s]" << std::endl;
}int main(int argc, char* argv[]) {
#ifdef USE_STACK_TRACE_LOGGERgoogle::InitGoogleLogging(argv[0]);google::InstallFailureSignalHandler();
#endif// create optionspopl::OptionParser op("Allowed options");auto help = op.add<popl::Switch>("h", "help", "produce help message");auto vocab_file_path = op.add<popl::Value<std::string>>("v", "vocab", "vocabulary file path");auto img_dir_path = op.add<popl::Value<std::string>>("i", "img-dir", "directory path which contains images");auto config_file_path = op.add<popl::Value<std::string>>("c", "config", "config file path");auto mask_img_path = op.add<popl::Value<std::string>>("", "mask", "mask image path", "");auto frame_skip = op.add<popl::Value<unsigned int>>("", "frame-skip", "interval of frame skip", 1);auto no_sleep = op.add<popl::Switch>("", "no-sleep", "not wait for next frame in real time");auto auto_term = op.add<popl::Switch>("", "auto-term", "automatically terminate the viewer");auto debug_mode = op.add<popl::Switch>("", "debug", "debug mode");auto eval_log = op.add<popl::Switch>("", "eval-log", "store trajectory and tracking times for evaluation");auto map_db_path = op.add<popl::Value<std::string>>("p", "map-db", "store a map database at this path after SLAM", "");try {op.parse(argc, argv);}catch (const std::exception& e) {std::cerr << e.what() << std::endl;std::cerr << std::endl;std::cerr << op << std::endl;return EXIT_FAILURE;}// check validness of optionsif (help->is_set()) {std::cerr << op << std::endl;return EXIT_FAILURE;}if (!vocab_file_path->is_set() || !img_dir_path->is_set() || !config_file_path->is_set()) {std::cerr << "invalid arguments" << std::endl;std::cerr << std::endl;std::cerr << op << std::endl;return EXIT_FAILURE;}// setup loggerspdlog::set_pattern("[%Y-%m-%d %H:%M:%S.%e] %^[%L] %v%$");if (debug_mode->is_set()) {spdlog::set_level(spdlog::level::debug);}else {spdlog::set_level(spdlog::level::info);}// load configurationstd::shared_ptr<openvslam::config> cfg;try {cfg = std::make_shared<openvslam::config>(config_file_path->value());}catch (const std::exception& e) {std::cerr << e.what() << std::endl;return EXIT_FAILURE;}#ifdef USE_GOOGLE_PERFTOOLSProfilerStart("slam.prof");
#endif// run trackingif (cfg->camera_->setup_type_ == openvslam::camera::setup_type_t::Monocular) {mono_tracking(cfg, vocab_file_path->value(), img_dir_path->value(), mask_img_path->value(),frame_skip->value(), no_sleep->is_set(), auto_term->is_set(),eval_log->is_set(), map_db_path->value());}else {throw std::runtime_error("Invalid setup type: " + cfg->camera_->get_setup_type_string());}#ifdef USE_GOOGLE_PERFTOOLSProfilerStop();
#endifreturn EXIT_SUCCESS;
}
#自己添加 resize图像建图的文件夹例子 add_executable(run_image_slam_resize run_image_slam_resize.cc util/image_util.cc) list(APPEND EXECUTABLE_TARGETS run_image_slam_resize)#自己添加 resize图像定位文件夹例子 add_executable(run_image_localization_resize run_image_localization_resize.cc util/image_util.cc) list(APPEND EXECUTABLE_TARGETS run_image_localization_resize)