560 lines
19 KiB
C++
560 lines
19 KiB
C++
|
#include "features2d.h"
|
||
|
|
|
||
|
|
AKAZE AKAZE_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::AKAZE>(cv::AKAZE::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void AKAZE_Close(AKAZE a) {
|
||
|
|
delete a;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints AKAZE_Detect(AKAZE a, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints AKAZE_DetectAndCompute(AKAZE a, Mat src, Mat mask, Mat desc) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detectAndCompute(*src, *mask, detected, *desc);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
AgastFeatureDetector AgastFeatureDetector_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::AgastFeatureDetector>(cv::AgastFeatureDetector::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void AgastFeatureDetector_Close(AgastFeatureDetector a) {
|
||
|
|
delete a;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints AgastFeatureDetector_Detect(AgastFeatureDetector a, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
BRISK BRISK_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::BRISK>(cv::BRISK::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void BRISK_Close(BRISK b) {
|
||
|
|
delete b;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints BRISK_Detect(BRISK b, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*b)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints BRISK_DetectAndCompute(BRISK b, Mat src, Mat mask, Mat desc) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*b)->detectAndCompute(*src, *mask, detected, *desc);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
GFTTDetector GFTTDetector_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::GFTTDetector>(cv::GFTTDetector::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void GFTTDetector_Close(GFTTDetector a) {
|
||
|
|
delete a;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints GFTTDetector_Detect(GFTTDetector a, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
KAZE KAZE_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::KAZE>(cv::KAZE::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void KAZE_Close(KAZE a) {
|
||
|
|
delete a;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints KAZE_Detect(KAZE a, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints KAZE_DetectAndCompute(KAZE a, Mat src, Mat mask, Mat desc) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detectAndCompute(*src, *mask, detected, *desc);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
MSER MSER_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::MSER>(cv::MSER::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void MSER_Close(MSER a) {
|
||
|
|
delete a;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints MSER_Detect(MSER a, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*a)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
FastFeatureDetector FastFeatureDetector_Create() {
|
||
|
|
return new cv::Ptr<cv::FastFeatureDetector>(cv::FastFeatureDetector::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void FastFeatureDetector_Close(FastFeatureDetector f) {
|
||
|
|
delete f;
|
||
|
|
}
|
||
|
|
|
||
|
|
FastFeatureDetector FastFeatureDetector_CreateWithParams(int threshold, bool nonmaxSuppression, int type) {
|
||
|
|
return new cv::Ptr<cv::FastFeatureDetector>(cv::FastFeatureDetector::create(threshold,nonmaxSuppression,static_cast<cv::FastFeatureDetector::DetectorType>(type)));
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints FastFeatureDetector_Detect(FastFeatureDetector f, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*f)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
ORB ORB_Create() {
|
||
|
|
return new cv::Ptr<cv::ORB>(cv::ORB::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
ORB ORB_CreateWithParams(int nfeatures, float scaleFactor, int nlevels, int edgeThreshold, int firstLevel, int WTA_K, int scoreType, int patchSize, int fastThreshold) {
|
||
|
|
return new cv::Ptr<cv::ORB>(cv::ORB::create(nfeatures, scaleFactor, nlevels, edgeThreshold, firstLevel, WTA_K, static_cast<cv::ORB::ScoreType>(scoreType), patchSize, fastThreshold));
|
||
|
|
}
|
||
|
|
|
||
|
|
void ORB_Close(ORB o) {
|
||
|
|
delete o;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints ORB_Detect(ORB o, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*o)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints ORB_DetectAndCompute(ORB o, Mat src, Mat mask, Mat desc) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*o)->detectAndCompute(*src, *mask, detected, *desc);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
cv::SimpleBlobDetector::Params ConvertCParamsToCPPParams(SimpleBlobDetectorParams params) {
|
||
|
|
cv::SimpleBlobDetector::Params converted;
|
||
|
|
|
||
|
|
converted.blobColor = params.blobColor;
|
||
|
|
converted.filterByArea = params.filterByArea;
|
||
|
|
converted.filterByCircularity = params.filterByCircularity;
|
||
|
|
converted.filterByColor = params.filterByColor;
|
||
|
|
converted.filterByConvexity = params.filterByConvexity;
|
||
|
|
converted.filterByInertia = params.filterByInertia;
|
||
|
|
converted.maxArea = params.maxArea;
|
||
|
|
converted.maxCircularity = params.maxCircularity;
|
||
|
|
converted.maxConvexity = params.maxConvexity;
|
||
|
|
converted.maxInertiaRatio = params.maxInertiaRatio;
|
||
|
|
converted.maxThreshold = params.maxThreshold;
|
||
|
|
converted.minArea = params.minArea;
|
||
|
|
converted.minCircularity = params.minCircularity;
|
||
|
|
converted.minConvexity = params.minConvexity;
|
||
|
|
converted.minDistBetweenBlobs = params.minDistBetweenBlobs;
|
||
|
|
converted.minInertiaRatio = params.minInertiaRatio;
|
||
|
|
converted.minRepeatability = params.minRepeatability;
|
||
|
|
converted.minThreshold = params.minThreshold;
|
||
|
|
converted.thresholdStep = params.thresholdStep;
|
||
|
|
|
||
|
|
return converted;
|
||
|
|
}
|
||
|
|
|
||
|
|
SimpleBlobDetectorParams ConvertCPPParamsToCParams(cv::SimpleBlobDetector::Params params) {
|
||
|
|
SimpleBlobDetectorParams converted;
|
||
|
|
|
||
|
|
converted.blobColor = params.blobColor;
|
||
|
|
converted.filterByArea = params.filterByArea;
|
||
|
|
converted.filterByCircularity = params.filterByCircularity;
|
||
|
|
converted.filterByColor = params.filterByColor;
|
||
|
|
converted.filterByConvexity = params.filterByConvexity;
|
||
|
|
converted.filterByInertia = params.filterByInertia;
|
||
|
|
converted.maxArea = params.maxArea;
|
||
|
|
converted.maxCircularity = params.maxCircularity;
|
||
|
|
converted.maxConvexity = params.maxConvexity;
|
||
|
|
converted.maxInertiaRatio = params.maxInertiaRatio;
|
||
|
|
converted.maxThreshold = params.maxThreshold;
|
||
|
|
converted.minArea = params.minArea;
|
||
|
|
converted.minCircularity = params.minCircularity;
|
||
|
|
converted.minConvexity = params.minConvexity;
|
||
|
|
converted.minDistBetweenBlobs = params.minDistBetweenBlobs;
|
||
|
|
converted.minInertiaRatio = params.minInertiaRatio;
|
||
|
|
converted.minRepeatability = params.minRepeatability;
|
||
|
|
converted.minThreshold = params.minThreshold;
|
||
|
|
converted.thresholdStep = params.thresholdStep;
|
||
|
|
|
||
|
|
return converted;
|
||
|
|
}
|
||
|
|
|
||
|
|
SimpleBlobDetector SimpleBlobDetector_Create_WithParams(SimpleBlobDetectorParams params){
|
||
|
|
cv::SimpleBlobDetector::Params actualParams;
|
||
|
|
return new cv::Ptr<cv::SimpleBlobDetector>(cv::SimpleBlobDetector::create(ConvertCParamsToCPPParams(params)));
|
||
|
|
}
|
||
|
|
|
||
|
|
SimpleBlobDetector SimpleBlobDetector_Create() {
|
||
|
|
return new cv::Ptr<cv::SimpleBlobDetector>(cv::SimpleBlobDetector::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
SimpleBlobDetectorParams SimpleBlobDetectorParams_Create() {
|
||
|
|
return ConvertCPPParamsToCParams(cv::SimpleBlobDetector::Params());
|
||
|
|
}
|
||
|
|
|
||
|
|
void SimpleBlobDetector_Close(SimpleBlobDetector b) {
|
||
|
|
delete b;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints SimpleBlobDetector_Detect(SimpleBlobDetector b, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*b)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
BFMatcher BFMatcher_Create() {
|
||
|
|
return new cv::Ptr<cv::BFMatcher>(cv::BFMatcher::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
BFMatcher BFMatcher_CreateWithParams(int normType, bool crossCheck) {
|
||
|
|
return new cv::Ptr<cv::BFMatcher>(cv::BFMatcher::create(normType, crossCheck));
|
||
|
|
}
|
||
|
|
|
||
|
|
void BFMatcher_Close(BFMatcher b) {
|
||
|
|
delete b;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct MultiDMatches BFMatcher_KnnMatch(BFMatcher b, Mat query, Mat train, int k) {
|
||
|
|
std::vector< std::vector<cv::DMatch> > matches;
|
||
|
|
(*b)->knnMatch(*query, *train, matches, k);
|
||
|
|
|
||
|
|
DMatches *dms = new DMatches[matches.size()];
|
||
|
|
for (size_t i = 0; i < matches.size(); ++i) {
|
||
|
|
DMatch *dmatches = new DMatch[matches[i].size()];
|
||
|
|
for (size_t j = 0; j < matches[i].size(); ++j) {
|
||
|
|
DMatch dmatch = {matches[i][j].queryIdx, matches[i][j].trainIdx, matches[i][j].imgIdx,
|
||
|
|
matches[i][j].distance};
|
||
|
|
dmatches[j] = dmatch;
|
||
|
|
}
|
||
|
|
dms[i] = {dmatches, (int) matches[i].size()};
|
||
|
|
}
|
||
|
|
MultiDMatches ret = {dms, (int) matches.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct MultiDMatches BFMatcher_KnnMatchWithParams(BFMatcher b, Mat query, Mat train, int k, Mat mask, bool compactResult) {
|
||
|
|
std::vector< std::vector<cv::DMatch> > matches;
|
||
|
|
(*b)->knnMatch(*query, *train, matches, k, *mask, compactResult);
|
||
|
|
|
||
|
|
DMatches *dms = new DMatches[matches.size()];
|
||
|
|
for (size_t i = 0; i < matches.size(); ++i) {
|
||
|
|
DMatch *dmatches = new DMatch[matches[i].size()];
|
||
|
|
for (size_t j = 0; j < matches[i].size(); ++j) {
|
||
|
|
DMatch dmatch = {matches[i][j].queryIdx, matches[i][j].trainIdx, matches[i][j].imgIdx,
|
||
|
|
matches[i][j].distance};
|
||
|
|
dmatches[j] = dmatch;
|
||
|
|
}
|
||
|
|
dms[i] = {dmatches, (int) matches[i].size()};
|
||
|
|
}
|
||
|
|
MultiDMatches ret = {dms, (int) matches.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
FlannBasedMatcher FlannBasedMatcher_Create() {
|
||
|
|
return new cv::Ptr<cv::FlannBasedMatcher>(cv::FlannBasedMatcher::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void FlannBasedMatcher_Close(FlannBasedMatcher f) {
|
||
|
|
delete f;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct MultiDMatches FlannBasedMatcher_KnnMatch(FlannBasedMatcher f, Mat query, Mat train, int k) {
|
||
|
|
std::vector< std::vector<cv::DMatch> > matches;
|
||
|
|
(*f)->knnMatch(*query, *train, matches, k);
|
||
|
|
|
||
|
|
DMatches *dms = new DMatches[matches.size()];
|
||
|
|
for (size_t i = 0; i < matches.size(); ++i) {
|
||
|
|
DMatch *dmatches = new DMatch[matches[i].size()];
|
||
|
|
for (size_t j = 0; j < matches[i].size(); ++j) {
|
||
|
|
DMatch dmatch = {matches[i][j].queryIdx, matches[i][j].trainIdx, matches[i][j].imgIdx,
|
||
|
|
matches[i][j].distance};
|
||
|
|
dmatches[j] = dmatch;
|
||
|
|
}
|
||
|
|
dms[i] = {dmatches, (int) matches[i].size()};
|
||
|
|
}
|
||
|
|
MultiDMatches ret = {dms, (int) matches.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct MultiDMatches FlannBasedMatcher_KnnMatchWithParams(FlannBasedMatcher f, Mat query, Mat train, int k, Mat mask, bool compactResult) {
|
||
|
|
std::vector< std::vector<cv::DMatch> > matches;
|
||
|
|
(*f)->knnMatch(*query, *train, matches, k, *mask, compactResult);
|
||
|
|
|
||
|
|
DMatches *dms = new DMatches[matches.size()];
|
||
|
|
for (size_t i = 0; i < matches.size(); ++i) {
|
||
|
|
DMatch *dmatches = new DMatch[matches[i].size()];
|
||
|
|
for (size_t j = 0; j < matches[i].size(); ++j) {
|
||
|
|
DMatch dmatch = {matches[i][j].queryIdx, matches[i][j].trainIdx, matches[i][j].imgIdx,
|
||
|
|
matches[i][j].distance};
|
||
|
|
dmatches[j] = dmatch;
|
||
|
|
}
|
||
|
|
dms[i] = {dmatches, (int) matches[i].size()};
|
||
|
|
}
|
||
|
|
MultiDMatches ret = {dms, (int) matches.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
void DrawKeyPoints(Mat src, struct KeyPoints kp, Mat dst, Scalar s, int flags) {
|
||
|
|
std::vector<cv::KeyPoint> keypts;
|
||
|
|
cv::KeyPoint keypt;
|
||
|
|
|
||
|
|
for (int i = 0; i < kp.length; ++i) {
|
||
|
|
keypt = cv::KeyPoint(kp.keypoints[i].x, kp.keypoints[i].y,
|
||
|
|
kp.keypoints[i].size, kp.keypoints[i].angle, kp.keypoints[i].response,
|
||
|
|
kp.keypoints[i].octave, kp.keypoints[i].classID);
|
||
|
|
keypts.push_back(keypt);
|
||
|
|
}
|
||
|
|
|
||
|
|
cv::Scalar color = cv::Scalar(s.val1, s.val2, s.val3, s.val4);
|
||
|
|
|
||
|
|
cv::drawKeypoints(*src, keypts, *dst, color, static_cast<cv::DrawMatchesFlags>(flags));
|
||
|
|
}
|
||
|
|
|
||
|
|
SIFT SIFT_Create() {
|
||
|
|
// TODO: params
|
||
|
|
return new cv::Ptr<cv::SIFT>(cv::SIFT::create());
|
||
|
|
}
|
||
|
|
|
||
|
|
void SIFT_Close(SIFT d) {
|
||
|
|
delete d;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints SIFT_Detect(SIFT d, Mat src) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*d)->detect(*src, detected);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
struct KeyPoints SIFT_DetectAndCompute(SIFT d, Mat src, Mat mask, Mat desc) {
|
||
|
|
std::vector<cv::KeyPoint> detected;
|
||
|
|
(*d)->detectAndCompute(*src, *mask, detected, *desc);
|
||
|
|
|
||
|
|
KeyPoint* kps = new KeyPoint[detected.size()];
|
||
|
|
|
||
|
|
for (size_t i = 0; i < detected.size(); ++i) {
|
||
|
|
KeyPoint k = {detected[i].pt.x, detected[i].pt.y, detected[i].size, detected[i].angle,
|
||
|
|
detected[i].response, detected[i].octave, detected[i].class_id
|
||
|
|
};
|
||
|
|
kps[i] = k;
|
||
|
|
}
|
||
|
|
|
||
|
|
KeyPoints ret = {kps, (int)detected.size()};
|
||
|
|
return ret;
|
||
|
|
}
|
||
|
|
|
||
|
|
void DrawMatches(Mat img1, struct KeyPoints kp1, Mat img2, struct KeyPoints kp2, struct DMatches matches1to2, Mat outImg, const Scalar matchesColor, const Scalar pointColor, struct ByteArray matchesMask, int flags) {
|
||
|
|
std::vector<cv::KeyPoint> kp1vec, kp2vec;
|
||
|
|
cv::KeyPoint keypt;
|
||
|
|
|
||
|
|
for (int i = 0; i < kp1.length; ++i) {
|
||
|
|
keypt = cv::KeyPoint(kp1.keypoints[i].x, kp1.keypoints[i].y,
|
||
|
|
kp1.keypoints[i].size, kp1.keypoints[i].angle, kp1.keypoints[i].response,
|
||
|
|
kp1.keypoints[i].octave, kp1.keypoints[i].classID);
|
||
|
|
kp1vec.push_back(keypt);
|
||
|
|
}
|
||
|
|
|
||
|
|
for (int i = 0; i < kp2.length; ++i) {
|
||
|
|
keypt = cv::KeyPoint(kp2.keypoints[i].x, kp2.keypoints[i].y,
|
||
|
|
kp2.keypoints[i].size, kp2.keypoints[i].angle, kp2.keypoints[i].response,
|
||
|
|
kp2.keypoints[i].octave, kp2.keypoints[i].classID);
|
||
|
|
kp2vec.push_back(keypt);
|
||
|
|
}
|
||
|
|
|
||
|
|
cv::Scalar cvmatchescolor = cv::Scalar(matchesColor.val1, matchesColor.val2, matchesColor.val3, matchesColor.val4);
|
||
|
|
cv::Scalar cvpointcolor = cv::Scalar(pointColor.val1, pointColor.val2, pointColor.val3, pointColor.val4);
|
||
|
|
|
||
|
|
std::vector<cv::DMatch> dmatchvec;
|
||
|
|
cv::DMatch dm;
|
||
|
|
|
||
|
|
for (int i = 0; i < matches1to2.length; i++) {
|
||
|
|
dm = cv::DMatch(matches1to2.dmatches[i].queryIdx, matches1to2.dmatches[i].trainIdx,
|
||
|
|
matches1to2.dmatches[i].imgIdx, matches1to2.dmatches[i].distance);
|
||
|
|
dmatchvec.push_back(dm);
|
||
|
|
}
|
||
|
|
|
||
|
|
std::vector<char> maskvec;
|
||
|
|
|
||
|
|
for (int i = 0; i < matchesMask.length; i++) {
|
||
|
|
maskvec.push_back(matchesMask.data[i]);
|
||
|
|
}
|
||
|
|
|
||
|
|
cv::drawMatches(*img1, kp1vec, *img2, kp2vec, dmatchvec, *outImg, cvmatchescolor, cvpointcolor, maskvec, static_cast<cv::DrawMatchesFlags>(flags));
|
||
|
|
}
|