HKUST VINSMONO ：香港科大开源VINSSLAM算法 part 2

int main(int argc, char **argv)
{

    ros::init(argc, argv, "vins_estimator");
    ros::NodeHandle n("~");
    ros::console::set_logger_level(ROSCONSOLE_DEFAULT_NAME, ros::console::levels::Info);
    readParameters(n);
    estimator.setParameter();
#ifdef EIGEN_DONT_PARALLELIZE
    ROS_DEBUG("EIGEN_DONT_PARALLELIZE");
#endif
    ROS_WARN("waiting for image and imu...");

    registerPub(n);
    
    //NOTE tcpNoDelay: a TCP transport is used, specifies whether or not to use TCP_NODELAY to provide a potentially lower-latency connection. 
    
    ros::Subscriber sub_imu = n.subscribe(IMU_TOPIC, 2000, imu_callback, ros::TransportHints().tcpNoDelay());
    ros::Subscriber sub_image = n.subscribe("/feature_tracker/feature", 2000, feature_callback);
    ros::Subscriber sub_raw_image = n.subscribe(IMAGE_TOPIC, 2000, raw_image_callback);

    std::thread measurement_process{process};
    std::thread loop_detection;
    if (LOOP_CLOSURE)
    {
        ROS_WARN("LOOP_CLOSURE true");
        loop_detection = std::thread(process_loop_detection);   
        m_camera = CameraFactory::instance()->generateCameraFromYamlFile(CAM_NAMES);
    }
    ros::spin();

    return 0;
}

作为estimator的主函数，首先进行了ros初始化，之后定义了三个subscriber分别为：

1. sub imu

2. sub image

3. sub raw image

其中sub imu订阅了/imu0 topic然后将imu数据push back到imu buffer，

sub image 订阅了feature tracker的publisher ，并将feature push back 到feature buffer

sub raw image 订阅了/cam0/raw_image topic并将图像通过cv：：bridge传入

之后创建了两个线程分别为：

std::thread measurement_process{process};

以及

std::thread loop_detection;

我们首先来看getMeasurements()函数：

std::vector<std::pair<std::vector<sensor_msgs::ImuConstPtr>, sensor_msgs::PointCloudConstPtr>>
getMeasurements()

代码定义了一个vector，里面的element 是std：：pair，这个pair由一个imu vector以及一个 特征点指针构成。

这个函数的作用是将imu buffer的数据以及feature buffer 的数据返回。

另外一个比较重要的是void send_imu(const sensor_msgs::ImuConstPtr &imu_msg)函数，

在process函数内，将measurement的返回值的first部分，也就是IMU vector内的每一个元素，调用send_imu（）；

而send imu（） 的目的就是将测量值的imu部分input到estimator.processIMU(dt, Vector3d(dx, dy, dz), Vector3d(rx, ry, rz))部分。

processIMU这个estimator的成员函数将measurement 的IMU部分进行了预积分处理。

在process（）内还有一个estimator.processImage()函数，我们来看一下这个函数

void Estimator::processImage(const map<int, vector<pair<int, Vector3d>>> &image, const std_msgs::Header &header)
{
    ROS_DEBUG("new image coming ------------------------------------------");
    ROS_DEBUG("Adding feature points %lu", image.size());
    if (f_manager.addFeatureCheckParallax(frame_count, image))
        marginalization_flag = MARGIN_OLD;
    else
        marginalization_flag = MARGIN_SECOND_NEW;

    ROS_DEBUG("this frame is--------------------%s", marginalization_flag ? "reject" : "accept");
    ROS_DEBUG("%s", marginalization_flag ? "Non-keyframe" : "Keyframe");
    ROS_DEBUG("Solving %d", frame_count);
    ROS_DEBUG("number of feature: %d", f_manager.getFeatureCount());
    Headers[frame_count] = header;

    ImageFrame imageframe(image, header.stamp.toSec());
    imageframe.pre_integration = tmp_pre_integration;
    all_image_frame.insert(make_pair(header.stamp.toSec(), imageframe));
    tmp_pre_integration = new IntegrationBase{acc_0, gyr_0, Bas[frame_count], Bgs[frame_count]};

    if(ESTIMATE_EXTRINSIC == 2)
    {
        ROS_INFO("calibrating extrinsic param, rotation movement is needed");
        if (frame_count != 0)
        {
            vector<pair<Vector3d, Vector3d>> corres = f_manager.getCorresponding(frame_count - 1, frame_count);
            Matrix3d calib_ric;
            if (initial_ex_rotation.CalibrationExRotation(corres, pre_integrations[frame_count]->delta_q, calib_ric))
            {
                ROS_WARN("initial extrinsic rotation calib success");
                ROS_WARN_STREAM("initial extrinsic rotation: " << endl << calib_ric);
                ric[0] = calib_ric;
                RIC[0] = calib_ric;
                ESTIMATE_EXTRINSIC = 1;
            }
        }
    }

    if (solver_flag == INITIAL)
    {
        if (frame_count == WINDOW_SIZE)
        {
            bool result = false;
            if( ESTIMATE_EXTRINSIC != 2 && (header.stamp.toSec() - initial_timestamp) > 0.1)
            {
               result = initialStructure();
               initial_timestamp = header.stamp.toSec();
            }
            if(result)
            {
                solver_flag = NON_LINEAR;
                solveOdometry();
                slideWindow();
                f_manager.removeFailures();
                ROS_INFO("Initialization finish!");
                last_R = Rs[WINDOW_SIZE];
                last_P = Ps[WINDOW_SIZE];
                last_R0 = Rs[0];
                last_P0 = Ps[0];
                
            }
            else
                slideWindow();
        }
        else
            frame_count++;
    }
    else
    {
        TicToc t_solve;
        solveOdometry();
        ROS_DEBUG("solver costs: %fms", t_solve.toc());

        if (failureDetection())
        {
            ROS_WARN("failure detection!");
            failure_occur = 1;
            clearState();
            setParameter();
            ROS_WARN("system reboot!");
            return;
        }

        TicToc t_margin;
        slideWindow();
        f_manager.removeFailures();
        ROS_DEBUG("marginalization costs: %fms", t_margin.toc());
        // prepare output of VINS
        key_poses.clear();
        for (int i = 0; i <= WINDOW_SIZE; i++)
            key_poses.push_back(Ps[i]);

        last_R = Rs[WINDOW_SIZE];
        last_P = Ps[WINDOW_SIZE];
        last_R0 = Rs[0];
        last_P0 = Ps[0];
    }
}

process（）这个函数主要完成了以下几件事：

1. 将测量值的IMU部分进行预积分处理；

2. processImage，将测量值的特征点部分进行处理并进行imu融合，在processImage函数中主要完成以下两项工作：

1. 滑窗 slideWindow（）；
2. 滑窗内位姿优化：位于solveOdometry 中的 optimization（）函数中，为VINS中IMU与视觉融合的重点部分。；