rtsensordataworker.cpp

Go to the documentation of this file.

//=============================================================================================================
 
//=============================================================================================================
// INCLUDES
//=============================================================================================================
 
#include "rtsensordataworker.h"
#include "core/rendertypes.h"
 
#include <disp/plots/helpers/colormap.h>
#include <QMutexLocker>
#include <QDebug>
#include <cmath>
 
using namespace DISP3DRHILIB;
 
//=============================================================================================================
// DEFINE MEMBER METHODS
//=============================================================================================================
 
RtSensorDataWorker::RtSensorDataWorker(QObject *parent)
    : QObject(parent)
{
}
 
//=============================================================================================================
 
void RtSensorDataWorker::addData(const Eigen::VectorXf &data)
{
    QMutexLocker locker(&m_mutex);
 
    // Cap queue size at sampling frequency (1 second of data)
    if (m_lDataQ.size() < static_cast<int>(m_dSFreq)) {
        m_lDataQ.append(data);
    }
 
    // Also store for looping
    if (m_bIsLooping) {
        m_lDataLoopQ.append(data);
        // Cap loop queue at 10 seconds of data
        if (m_lDataLoopQ.size() > static_cast<int>(m_dSFreq) * 10) {
            m_lDataLoopQ.removeFirst();
        }
    }
}
 
//=============================================================================================================
 
void RtSensorDataWorker::clear()
{
    QMutexLocker locker(&m_mutex);
    m_lDataQ.clear();
    m_lDataLoopQ.clear();
    m_vecAverage = Eigen::VectorXf();
    m_iSampleCtr = 0;
    m_iCurrentSample = 0;
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setMappingMatrix(QSharedPointer<Eigen::MatrixXf> mat)
{
    QMutexLocker locker(&m_mutex);
    m_mappingMat = mat;
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setNumberAverages(int numAvr)
{
    QMutexLocker locker(&m_mutex);
    m_iNumAverages = qMax(1, numAvr);
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setColormapType(const QString &name)
{
    QMutexLocker locker(&m_mutex);
    m_sColormapType = name;
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setThresholds(double min, double max)
{
    QMutexLocker locker(&m_mutex);
    if (min == 0.0 && max == 0.0) {
        m_bUseAutoNorm = true;
        m_dThreshMin = 0.0;
        m_dThreshMax = 0.0;
    } else {
        m_bUseAutoNorm = false;
        m_dThreshMin = min;
        m_dThreshMax = max;
    }
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setLoopState(bool enabled)
{
    QMutexLocker locker(&m_mutex);
    m_bIsLooping = enabled;
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setSFreq(double sFreq)
{
    QMutexLocker locker(&m_mutex);
    m_dSFreq = qMax(1.0, sFreq);
}
 
//=============================================================================================================
 
void RtSensorDataWorker::setStreamSmoothedData(bool bStreamSmoothedData)
{
    QMutexLocker locker(&m_mutex);
    m_bStreamSmoothedData = bStreamSmoothedData;
}
 
//=============================================================================================================
 
void RtSensorDataWorker::streamData()
{
    QMutexLocker locker(&m_mutex);
 
    // Try to get data from queue
    Eigen::VectorXf vecCurrentData;
 
    if (!m_lDataQ.isEmpty()) {
        vecCurrentData = m_lDataQ.takeFirst();
    } else if (m_bIsLooping && !m_lDataLoopQ.isEmpty()) {
        // Loop: replay from stored data
        vecCurrentData = m_lDataLoopQ[m_iCurrentSample % m_lDataLoopQ.size()];
        m_iCurrentSample++;
    } else {
        // No data available
        return;
    }
 
    // Averaging
    if (m_iNumAverages > 1) {
        if (m_vecAverage.size() != vecCurrentData.size()) {
            m_vecAverage = vecCurrentData;
            m_iSampleCtr = 1;
        } else {
            m_vecAverage += vecCurrentData;
            m_iSampleCtr++;
        }
 
        if (m_iSampleCtr < m_iNumAverages) {
            return; // Accumulate more samples
        }
 
        vecCurrentData = m_vecAverage / static_cast<float>(m_iSampleCtr);
        m_vecAverage = Eigen::VectorXf();
        m_iSampleCtr = 0;
    }
 
    // Check streaming mode
    if (!m_bStreamSmoothedData) {
        // Raw mode: emit measurement vector without mapping
        Eigen::VectorXf rawData = vecCurrentData;
        locker.unlock();
        emit newRtRawSensorData(rawData);
        return;
    }
 
    // Compute per-vertex colors using the dense mapping matrix
    QVector<uint32_t> colors = computeSurfaceColors(vecCurrentData);
 
    // Unlock before emitting (avoid deadlock if slot is direct connection)
    locker.unlock();
 
    if (!colors.isEmpty()) {
        emit newRtSensorColors(m_sSurfaceKey, colors);
    }
}
 
//=============================================================================================================
 
QVector<uint32_t> RtSensorDataWorker::computeSurfaceColors(const Eigen::VectorXf &sensorData) const
{
    if (sensorData.size() == 0 || !m_mappingMat || m_mappingMat->rows() == 0) {
        return QVector<uint32_t>();
    }
 
    // Validate dimensions
    if (m_mappingMat->cols() != sensorData.size()) {
        qWarning() << "RtSensorDataWorker: Mapping matrix cols" << m_mappingMat->cols()
                   << "!= sensor data size" << sensorData.size();
        return QVector<uint32_t>();
    }
 
    // Map sensor data to surface vertices: mapped = M * meas
    Eigen::VectorXf mapped = (*m_mappingMat) * sensorData;
 
    int nVertices = mapped.size();
 
    // Determine normalization range
    float normMin, normMax;
    if (m_bUseAutoNorm) {
        // Symmetric auto-normalization: ±maxAbs
        float maxAbs = 0.0f;
        for (int i = 0; i < nVertices; ++i) {
            maxAbs = std::max(maxAbs, std::abs(mapped(i)));
        }
        if (maxAbs <= 0.0f) maxAbs = 1.0f;
        normMin = -maxAbs;
        normMax = maxAbs;
    } else {
        // Explicit thresholds
        normMin = static_cast<float>(m_dThreshMin);
        normMax = static_cast<float>(m_dThreshMax);
        if (normMax <= normMin) normMax = normMin + 1.0f;
    }
 
    float range = normMax - normMin;
 
    // Convert to per-vertex ABGR colours
    QVector<uint32_t> colors(nVertices);
 
    for (int i = 0; i < nVertices; ++i) {
        // Normalisation: map [normMin, normMax] → [0, 1]
        double norm = static_cast<double>(mapped(i) - normMin) / static_cast<double>(range);
        norm = qBound(0.0, norm, 1.0);
 
        QRgb rgb = DISPLIB::ColorMap::valueToColor(norm, m_sColormapType);
 
        uint32_t r = qRed(rgb);
        uint32_t g = qGreen(rgb);
        uint32_t b = qBlue(rgb);
 
        // Pack as ABGR (same format as BrainSurface uses)
        colors[i] = packABGR(r, g, b);
    }
 
    return colors;
}