sunwen/UR
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix detectAttVectorized

Browse Source
This commit is contained in:
kradchen
2023-05-22 14:07:27 +08:00
parent a4a972f99f
commit 5276d1cc42
2 changed files with 43 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
src/transmissionReconstruction/detection/detection.cpp
View File

@@ -84,9 +84,10 @@ namespace Recon {
if(gaussWindow)
{
auto expectedPosWater = (distBlock / expectedSOSWater) * sampleRate + startOffset;
auto windowWidth = calcResult.endSearch-calcResult.startSearch;
#pragma omp parallel for
for (size_t i = 0; i < AscanBlock.getDimSize(1); i++)
{
auto windowWidth = calcResult.endSearch-calcResult.startSearch;
auto t = linspace(-5,5,windowWidth[i]);
auto gauss = exp( -0.1 * (transpose(t) ^ 2));
@@ -105,6 +106,7 @@ namespace Recon {
}
}
else{
#pragma omp parallel for
for (size_t i = 0; i < AscanBlock.getDimSize(1); i++) {
for (size_t j = calcResult.startSearch[i] - 1;
j < calcResult.endSearch[i]; j++) {
@@ -118,7 +120,6 @@ namespace Recon {
return result;
}
//TODO:need test
Aurora::Matrix detectAttVectorized(const Aurora::Matrix &Ascan, const Aurora::Matrix &AscanRef,
const Aurora::Matrix &distRef,
const Aurora::Matrix &sosWaterRef,
@@ -170,24 +171,34 @@ namespace Recon {
auto sampleRate = aScanReconstructionFrequency;
double offsetElectronicSamples = offsetElectronic * sampleRate;
Matrix diffStartSearch;
TimeWindowResult timeResult1;
timeResult1.AscanBlockProcessed = AscanBlock;
TimeWindowResult timeResult2;
timeResult2.AscanBlockProcessed = AscanRefBlock;
if (useTimeWindowing == 1)
{
auto timeResult1 = applyTimeWindowing(AscanBlock, sampleRate, distBlock, sosWaterBlock, expectedSOSWater, offsetElectronicSamples, detectionWindowSOS, minSpeedOfSound, maxSpeedOfSound, gaussWindow);
auto timeResult2= applyTimeWindowing(AscanRefBlock, sampleRate, distBlockRef, sosWaterRefBlock, expectedSOSWater, offsetElectronicSamples, detectionWindowSOS, minSpeedOfSound, maxSpeedOfSound, gaussWindow);
timeResult1 = applyTimeWindowing(AscanBlock, sampleRate, distBlock, sosWaterBlock, expectedSOSWater, offsetElectronicSamples, detectionWindowSOS, minSpeedOfSound, maxSpeedOfSound, gaussWindow);
timeResult2 = applyTimeWindowing(AscanRefBlock, sampleRate, distBlockRef, sosWaterRefBlock, expectedSOSWater, offsetElectronicSamples, detectionWindowSOS, minSpeedOfSound, maxSpeedOfSound, gaussWindow);
diffStartSearch = timeResult1.startSearch - timeResult1.startSearch;
}
auto _AscanBlock = timeResult1.AscanBlockProcessed;
auto _AscanRefBlock = timeResult2.AscanBlockProcessed;
auto m = std::max(size(AscanBlock, 1), size(AscanRefBlock, 1));
auto maxlag = std::max(size(AscanBlock, 1), size(AscanRefBlock, 1)) - 1;
auto m = std::max(size(_AscanBlock, 1), size(_AscanRefBlock, 1));
auto maxlag = std::max(size(_AscanBlock, 1), size(_AscanRefBlock, 1)) - 1;
auto mxl = std::min(maxlag, m-1);
auto ceilLog2 = nextpow2(2*m-1);
auto m2 = pow(2,ceilLog2);
auto x = fft(AscanBlock, m2);
auto y = fft(AscanRefBlock, m2);
auto c1 = ifft(x*conj(y));
auto c = complex(zeros(mxl+mxl+1,c1.getDimSize(1)));
auto x = fft(_AscanBlock, m2);
auto y = fft(_AscanRefBlock, m2);
auto c_1_1 = x*conj(y);
auto c_1_2 = ifft(c_1_1);
auto c1 = real(c_1_2);
auto c = zeros(mxl+mxl+1,c1.getDimSize(1));
#pragma omp parallel for
for (size_t i = 0; i < mxl; i++)
{
c(i,$) = c1(m2-mxl+i, $);
@@ -196,11 +207,12 @@ namespace Recon {
c(mxl+mxl,$) = c1(mxl, $);
auto shiftInSamples =zeros(1,c1.getDimSize(1));
#pragma omp parallel for
for (size_t i = 0; i < c1.getDimSize(1); i++)
{
long rowID=0,colID=0;
max(c(i,$).toMatrix(),All,rowID,colID);
shiftInSamples[i]=-maxlag+colID;
max(c($,i).toMatrix(),All,rowID,colID);
shiftInSamples[i]=-maxlag+colID+rowID;
}
if (useTimeWindowing)
{
@@ -208,5 +220,6 @@ namespace Recon {
}
auto tof = shiftInSamples / sampleRate + distBlock / sosWaterBlock;
auto sosValue = distBlock / tof;
return tof;
}
}

32
test/Detection_Test.cpp
View File

@@ -73,24 +73,19 @@ TEST_F(Detection_Test, calculateAttenuation) {
TEST_F(Detection_Test, applyTimeWindowing) {
MatlabReader m("/home/krad/TestData/timeWindow.mat");
MatlabReader m("/home/krad/TestData/timeWindow2.mat");
auto AscanBlock = m.read("AscanBlock");
auto distBlock = m.read("distBlock");
auto sosBlock = m.read("sosBlock");
auto AscanBlockProcessed = m.read("AscanBlockProcessed");
auto distBlock = m.read("dists");
auto sosBlock = m.read("waterTempBlock");
auto AscanBlockProcessed = m.read("AscanBlock1");
auto startSearch = m.read("startSearch");
auto result = Recon::applyTimeWindowing(AscanBlock, 10000000, distBlock, sosBlock, 1482, 0, 0, 1400, 1650, false);
auto result = Recon::applyTimeWindowing(AscanBlock, 10000000, distBlock, sosBlock, 1.511948131508464e+03, 5.2, 1, 1450, 1550, false);
#pragma omp parallel for
for (size_t i = 0; i < AscanBlockProcessed.getDataSize(); i++)
{
EXPECT_DOUBLE_AE(AscanBlockProcessed[i],result.AscanBlockProcessed[i])<<",index:"<<i;
}
auto AscanBlockProcessed2 = m.read("AscanBlockProcessed2");
result = Recon::applyTimeWindowing(AscanBlock, 10000000, distBlock, sosBlock, 1482, 0, 0, 1400, 1650, true);
for (size_t i = 0; i < AscanBlockProcessed2.getDataSize(); i++)
{
EXPECT_DOUBLE_AE(AscanBlockProcessed2[i],result.AscanBlockProcessed[i])<<",index:"<<i;
}
}
@@ -102,8 +97,8 @@ TEST_F(Detection_Test, detectTofVectorized) {
auto AscanRefBlock = m.read("AscanRefBlock");
auto distBlock = m.read("dists");
auto distBlockRef = m.read("distRefBlock");
auto sosWaterBlock = m.read("sosWaterBlock");
auto sosWaterRefBlock = m.read("sosWaterRefBlock");
auto sosWaterBlock = m.read("waterTempBlock");
auto sosWaterRefBlock = m.read("waterTempRefBlock");
double expectedSOSWater = 1.511948131508464e+03;
auto result = Recon::detectTofVectorized(
@@ -116,5 +111,14 @@ TEST_F(Detection_Test, detectTofVectorized) {
Recon::transParams::minSpeedOfSound,
Recon::transParams::maxSpeedOfSound, Recon::transParams::gaussWindow);
int size = result.getDataSize();
MatlabReader m2("/home/krad/TestData/tofResult.mat");
auto tof = m2.read("tof");
auto sosvalue = m2.read("sosValue");
EXPECT_EQ(tof.getDataSize(), result.getDataSize());
#pragma omp parallel for
for (size_t i = 0; i < tof.getDataSize(); i++)
{
EXPECT_DOUBLE_AE(tof[i],result[i])<<",index:"<<i;
}
}