1
0
mirror of https://github.com/opencv/opencv.git synced 2026-07-29 23:33:05 +04:00

Merge pull request #29267 from vrabaud:aruco

Fix speed regression in Aruco identify #29267

### Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [x] There is a reference to the original bug report and related work
- [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
      Patch to opencv_extra has the same branch name.
- [ ] The feature is well documented and sample code can be built with the project CMake
This commit is contained in:
Vincent Rabaud
2026-06-11 19:54:39 +02:00
committed by GitHub
parent 4dee742731
commit fa55ed2837
@@ -76,6 +76,35 @@ void Dictionary::writeDictionary(FileStorage& fs, const String &name)
bool Dictionary::identify(const Mat &onlyCellPixelRatio, CV_OUT int &idx, CV_OUT int &rotation, double maxCorrectionRate, float validBitIdThreshold) const {
CV_Assert(onlyCellPixelRatio.rows == markerSize && onlyCellPixelRatio.cols == markerSize);
// Fill bit masks of cells that are not black (not0) and not white (not1).
const int s = (markerSize * markerSize + 8 - 1) / 8;
AutoBuffer<uint8_t> temp(4 * s);
uint8_t* not0 = temp.data(), * not1 = not0 + s;
int not0Byte = 0, not1Byte = 0, currentByte = 0, currentBit = 0;
for(int j = 0; j < markerSize; j++) {
const float* cellPixelRatioRow = onlyCellPixelRatio.ptr<float>(j);
for(int i = 0; i < markerSize; i++) {
not0Byte <<= 1; not1Byte <<= 1;
if(cellPixelRatioRow[i] > validBitIdThreshold) not0Byte |= 1;
if(cellPixelRatioRow[i] < 1 - validBitIdThreshold) not1Byte |= 1;
++currentBit;
if(currentBit == 8) {
not0[currentByte] = not0Byte;
not1[currentByte] = not1Byte;
not0Byte = not1Byte = 0;
++currentByte;
currentBit = 0;
}
}
}
if (currentBit != 0) {
not0[currentByte] = not0Byte;
not1[currentByte] = not1Byte;
}
uint8_t* notXor = not1 + s, * temp0 = notXor + s;
// Computing: notXor = not0 ^ not1
hal::xor8u(not0, s, not1, s, notXor, s, s, 1, nullptr);
int maxCorrectionRecalculed = int(double(maxCorrectionBits) * maxCorrectionRate);
idx = -1; // by default, not found
@@ -84,25 +113,21 @@ bool Dictionary::identify(const Mat &onlyCellPixelRatio, CV_OUT int &idx, CV_OUT
for(int m = 0; m < bytesList.rows; m++) {
int currentMinDistance = markerSize * markerSize + 1;
int currentRotation = -1;
for(int r = 0; r < 4; r++) {
Mat bitsRot = getBitsFromByteList(bytesList.rowRange(m, m + 1), markerSize, r);
bitsRot.convertTo(bitsRot, CV_32F);
// Loop over all bits dictBitsList [m, markerSize * markerSize, 4]; onlyCellPixelRatio [markerSize, markerSize]
int currentHamming = 0;
for(int i = 0; i < markerSize; i++) {
for(int j = 0; j < markerSize; j++) {
// If detected bit is too far from the ground truth, consider it false.
if(fabs(onlyCellPixelRatio.at<float>(i, j) - static_cast<float>(bitsRot.at<float>(i, j))) > validBitIdThreshold){
currentHamming++;
}
}
}
const uchar* bytesRot = bytesList.ptr(m);
for(int r = 0; r < 4; r++, bytesRot += s) {
// Error if: (marker is 0 and input is not 0) or (marker is 1 and input is not 1)
// i.e. if: (!bytesRot && not0) || (bytesRot && not1)
// This is actually: not0 ^ ((not0 ^ not1) & bytesRot)
// Computing: temp0 = (not0 ^ not1) & bytesRot
hal::and8u(notXor, s, bytesRot, s, temp0, s, s, 1, nullptr);
// Computing the final result (xor is performed internally).
int currentHamming = cv::hal::normHamming(not0, temp0, s);
if(currentHamming < currentMinDistance) {
currentMinDistance = currentHamming;
currentRotation = r;
// Break for perfect distance.
if (currentMinDistance == 0) break;
}
}