KalmanTracker.cpp

Go to the documentation of this file.

// © OpenShot Studios, LLC
//
// SPDX-License-Identifier: LGPL-3.0-or-later
 
// KalmanTracker.cpp: KalmanTracker Class Implementation Declaration
 
#include "KalmanTracker.h"
#include <ctime>
#include <algorithm>
 
using namespace std;
using namespace cv;
 
// initialize Kalman filter
void KalmanTracker::init_kf(
    StateType stateMat)
{
    int stateNum = 8;
    int measureNum = 4;
    kf = KalmanFilter(stateNum, measureNum, 0);
 
    measurement = Mat::zeros(measureNum, 1, CV_32F);
 
    kf.transitionMatrix = (Mat_<float>(8, 8) << 1, 0, 0, 0, 1, 0, 0, 0,
 
                           0, 1, 0, 0, 0, 1, 0, 0,
                           0, 0, 1, 0, 0, 0, 1, 0,
                           0, 0, 0, 1, 0, 0, 0, 1,
                           0, 0, 0, 0, 1, 0, 0, 0,
                           0, 0, 0, 0, 0, 1, 0, 0,
                           0, 0, 0, 0, 0, 0, 1, 0,
                           0, 0, 0, 0, 0, 0, 0, 1);
 
    setIdentity(kf.measurementMatrix);
    setIdentity(kf.processNoiseCov, Scalar::all(1e-1));
    kf.processNoiseCov.at<float>(2, 2) = 1e0; // higher noise for area (s) to adapt to size changes
    kf.processNoiseCov.at<float>(3, 3) = 1e0; // higher noise for aspect ratio (r)
    setIdentity(kf.measurementNoiseCov, Scalar::all(1e-4));
    setIdentity(kf.errorCovPost, Scalar::all(1e-2));
 
    // initialize state vector with bounding box in [cx,cy,s,r] style
    kf.statePost.at<float>(0, 0) = stateMat.x + stateMat.width / 2;
    kf.statePost.at<float>(1, 0) = stateMat.y + stateMat.height / 2;
    kf.statePost.at<float>(2, 0) = stateMat.area();
    kf.statePost.at<float>(3, 0) = stateMat.width / stateMat.height;
    kf.statePost.at<float>(4, 0) = 0.0f;
    kf.statePost.at<float>(5, 0) = 0.0f;
    kf.statePost.at<float>(6, 0) = 0.0f;
    kf.statePost.at<float>(7, 0) = 0.0f;
}
 
// Predict the estimated bounding box.
StateType KalmanTracker::predict()
{
    // predict
    Mat p = kf.predict();
    m_age += 1;
 
    if (m_time_since_update > 0)
        m_hit_streak = 0;
    m_time_since_update += 1;
 
    StateType predictBox = get_rect_xysr(p.at<float>(0, 0), p.at<float>(1, 0), p.at<float>(2, 0), p.at<float>(3, 0));
 
    m_history.push_back(predictBox);
    return m_history.back();
}
 
StateType KalmanTracker::predict2()
{
    // predict
    Mat p = kf.predict();
 
    StateType predictBox = get_rect_xysr(p.at<float>(0, 0), p.at<float>(1, 0), p.at<float>(2, 0), p.at<float>(3, 0));
 
    return predictBox;
}
 
// Update the state vector with observed bounding box.
void KalmanTracker::update(
    StateType stateMat)
{
    m_time_since_update = 0;
    m_history.clear();
    m_hits += 1;
    m_hit_streak += 1;
 
    // measurement
    measurement.at<float>(0, 0) = stateMat.x + stateMat.width / 2;
    measurement.at<float>(1, 0) = stateMat.y + stateMat.height / 2;
    measurement.at<float>(2, 0) = stateMat.area();
    measurement.at<float>(3, 0) = stateMat.width / stateMat.height;
 
    // update
    kf.correct(measurement);
    // time_t now = time(0);
    // convert now to string form
 
    // detect_times.push_back(dt);
}
 
void KalmanTracker::update_class_scores(const std::vector<ClassScore>& classScores, int fallbackClassId, float fallbackConfidence)
{
    const double decay = 0.82;
    const double update_weight = 1.0 - decay;
    const bool had_history = !classScoreHistory.empty();
 
    for (auto it = classScoreHistory.begin(); it != classScoreHistory.end();) {
        it->second *= decay;
        if (it->second < 0.0001)
            it = classScoreHistory.erase(it);
        else
            ++it;
    }
 
    const double candidate_weight = had_history ? update_weight : 1.0;
    if (classScores.empty()) {
        classScoreHistory[fallbackClassId] += fallbackConfidence * candidate_weight;
    } else {
        for (const auto& candidate : classScores) {
            if (candidate.classId < 0 || candidate.score <= 0.0f)
                continue;
            classScoreHistory[candidate.classId] += candidate.score * candidate_weight;
        }
    }
 
    if (classScoreHistory.empty()) {
        classId = fallbackClassId;
        return;
    }
 
    auto best = std::max_element(
        classScoreHistory.begin(), classScoreHistory.end(),
        [](const auto& a, const auto& b) { return a.second < b.second; });
    classId = best->first;
}
 
// Return the current state vector
StateType KalmanTracker::get_state()
{
    Mat s = kf.statePost;
    return get_rect_xysr(s.at<float>(0, 0), s.at<float>(1, 0), s.at<float>(2, 0), s.at<float>(3, 0));
}
 
// Convert bounding box from [cx,cy,s,r] to [x,y,w,h] style.
StateType KalmanTracker::get_rect_xysr(
    float cx,
    float cy,
    float s,
    float r)
{
    float w = sqrt(s * r);
    float h = s / w;
    float x = (cx - w / 2);
    float y = (cy - h / 2);
 
    if (x < 0 && cx > 0)
        x = 0;
    if (y < 0 && cy > 0)
        y = 0;
 
    return StateType(x, y, w, h);
}