lot's of changes; nonfree & photo modules added; SIFT & SURF -> nonfree module; Inpainting -> photo; refactored features2d (ORB is still failing tests), optimized brute-force matcher and made it non-template.
This commit is contained in:
Vadim Pisarevsky
@@ -94,27 +94,6 @@ HarrisResponses(const Mat& img, vector<KeyPoint>& pts, int blockSize, float harr
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
struct KeypointResponseGreaterThanThreshold
|
||||
{
|
||||
KeypointResponseGreaterThanThreshold(float _value) :
|
||||
value(_value)
|
||||
{
|
||||
}
|
||||
inline bool operator()(const KeyPoint& kpt) const
|
||||
{
|
||||
return kpt.response >= value;
|
||||
}
|
||||
float value;
|
||||
};
|
||||
|
||||
struct KeypointResponseGreater
|
||||
{
|
||||
inline bool operator()(const KeyPoint& kp1, const KeyPoint& kp2) const
|
||||
{
|
||||
return kp1.response > kp2.response;
|
||||
}
|
||||
};
|
||||
|
||||
static float IC_Angle(const Mat& image, const int half_k, Point2f pt,
|
||||
const vector<int> & u_max)
|
||||
{
|
||||
@@ -224,37 +203,37 @@ static void computeOrbDescriptor(const KeyPoint& kpt,
|
||||
{
|
||||
for (int i = 0; i < dsize; ++i, pattern += 16)
|
||||
{
|
||||
int t0, t1, t2, t3, a, b, k, val;
|
||||
int t0, t1, t2, t3, u, v, k, val;
|
||||
t0 = GET_VALUE(0); t1 = GET_VALUE(1);
|
||||
t2 = GET_VALUE(2); t3 = GET_VALUE(3);
|
||||
a = 0, b = 2;
|
||||
if( t1 > t0 ) t0 = t1, a = 1;
|
||||
if( t3 > t2 ) t2 = t3, b = 3;
|
||||
k = t0 > t2 ? a : b;
|
||||
u = 0, v = 2;
|
||||
if( t1 > t0 ) t0 = t1, u = 1;
|
||||
if( t3 > t2 ) t2 = t3, v = 3;
|
||||
k = t0 > t2 ? u : v;
|
||||
val = k;
|
||||
|
||||
t0 = GET_VALUE(4); t1 = GET_VALUE(5);
|
||||
t2 = GET_VALUE(6); t3 = GET_VALUE(7);
|
||||
a = 0, b = 2;
|
||||
if( t1 > t0 ) t0 = t1, a = 1;
|
||||
if( t3 > t2 ) t2 = t3, b = 3;
|
||||
k = t0 > t2 ? a : b;
|
||||
u = 0, v = 2;
|
||||
if( t1 > t0 ) t0 = t1, u = 1;
|
||||
if( t3 > t2 ) t2 = t3, v = 3;
|
||||
k = t0 > t2 ? u : v;
|
||||
val |= k << 2;
|
||||
|
||||
t0 = GET_VALUE(8); t1 = GET_VALUE(9);
|
||||
t2 = GET_VALUE(10); t3 = GET_VALUE(11);
|
||||
a = 0, b = 2;
|
||||
if( t1 > t0 ) t0 = t1, a = 1;
|
||||
if( t3 > t2 ) t2 = t3, b = 3;
|
||||
k = t0 > t2 ? a : b;
|
||||
u = 0, v = 2;
|
||||
if( t1 > t0 ) t0 = t1, u = 1;
|
||||
if( t3 > t2 ) t2 = t3, v = 3;
|
||||
k = t0 > t2 ? u : v;
|
||||
val |= k << 4;
|
||||
|
||||
t0 = GET_VALUE(12); t1 = GET_VALUE(13);
|
||||
t2 = GET_VALUE(14); t3 = GET_VALUE(15);
|
||||
a = 0, b = 2;
|
||||
if( t1 > t0 ) t0 = t1, a = 1;
|
||||
if( t3 > t2 ) t2 = t3, b = 3;
|
||||
k = t0 > t2 ? a : b;
|
||||
u = 0, v = 2;
|
||||
if( t1 > t0 ) t0 = t1, u = 1;
|
||||
if( t3 > t2 ) t2 = t3, v = 3;
|
||||
k = t0 > t2 ? u : v;
|
||||
val |= k << 6;
|
||||
|
||||
desc[i] = (uchar)val;
|
||||
@@ -567,168 +546,195 @@ static void makeRandomPattern(int patchSize, Point* pattern, int npoints)
|
||||
}
|
||||
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
|
||||
///////////////////////////////////////////////////////////////////////////////////////////////////////////
|
||||
|
||||
void ORB::CommonParams::read(const FileNode& fn)
|
||||
static Algorithm* createORB() { return new ORB; }
|
||||
static AlgorithmInfo orb_info("Feature2D.ORB", createORB);
|
||||
|
||||
AlgorithmInfo* ORB::info() const
|
||||
{
|
||||
scale_factor_ = fn["scaleFactor"];
|
||||
n_levels_ = int(fn["nLevels"]);
|
||||
first_level_ = int(fn["firstLevel"]);
|
||||
edge_threshold_ = fn["edgeThreshold"];
|
||||
patch_size_ = fn["patchSize"];
|
||||
WTA_K_ = fn["WTA_K"];
|
||||
if( WTA_K_ == 0 ) WTA_K_ = 2;
|
||||
score_type_ = fn["scoreType"];
|
||||
static volatile bool initialized = false;
|
||||
if( !initialized )
|
||||
{
|
||||
orb_info.addParam(this, "nFeatures", nfeatures);
|
||||
orb_info.addParam(this, "scaleFactor", scaleFactor);
|
||||
orb_info.addParam(this, "nLevels", nlevels);
|
||||
orb_info.addParam(this, "firstLevel", firstLevel);
|
||||
orb_info.addParam(this, "edgeThreshold", edgeThreshold);
|
||||
orb_info.addParam(this, "patchSize", patchSize);
|
||||
orb_info.addParam(this, "WTA_K", WTA_K);
|
||||
orb_info.addParam(this, "scoreType", scoreType);
|
||||
|
||||
initialized = true;
|
||||
}
|
||||
return &orb_info;
|
||||
}
|
||||
|
||||
void ORB::CommonParams::write(FileStorage& fs) const
|
||||
|
||||
static inline float getScale(int level, int firstLevel, double scaleFactor)
|
||||
{
|
||||
fs << "scaleFactor" << scale_factor_;
|
||||
fs << "nLevels" << int(n_levels_);
|
||||
fs << "firstLevel" << int(first_level_);
|
||||
fs << "edgeThreshold" << int(edge_threshold_);
|
||||
fs << "patchSize" << int(patch_size_);
|
||||
fs << "WTA_K" << WTA_K_;
|
||||
fs << "scoreType" << score_type_;
|
||||
}
|
||||
|
||||
void ORB::read(const FileNode& fn)
|
||||
{
|
||||
CommonParams params;
|
||||
params.read(fn);
|
||||
int n_features = int(fn["nFeatures"]);
|
||||
*this = ORB(n_features, params);
|
||||
}
|
||||
|
||||
void ORB::write(FileStorage& fs) const
|
||||
{
|
||||
params_.write(fs);
|
||||
fs << "nFeatures" << int(n_features_);
|
||||
}
|
||||
|
||||
static inline float get_scale(const ORB::CommonParams& params, int level)
|
||||
{
|
||||
return std::pow(params.scale_factor_, float(level) - float(params.first_level_));
|
||||
return (float)std::pow(scaleFactor, (double)(level - firstLevel));
|
||||
}
|
||||
|
||||
/** Constructor
|
||||
* @param detector_params parameters to use
|
||||
*/
|
||||
ORB::ORB(size_t n_features, const CommonParams & detector_params) :
|
||||
params_(detector_params), n_features_(n_features)
|
||||
{
|
||||
// fill the extractors and descriptors for the corresponding scales
|
||||
int n_levels = (int)params_.n_levels_;
|
||||
float factor = (float)(1.0 / params_.scale_factor_);
|
||||
float n_desired_features_per_scale = n_features_*(1 - factor)/(1 - (float)pow((double)factor, (double)n_levels));
|
||||
|
||||
n_features_per_level_.resize(n_levels);
|
||||
int sum_n_features = 0;
|
||||
for( int level = 0; level < n_levels-1; level++ )
|
||||
{
|
||||
n_features_per_level_[level] = cvRound(n_desired_features_per_scale);
|
||||
sum_n_features += n_features_per_level_[level];
|
||||
n_desired_features_per_scale *= factor;
|
||||
}
|
||||
n_features_per_level_[n_levels-1] = n_features - sum_n_features;
|
||||
|
||||
// Make sure we forget about what is too close to the boundary
|
||||
//params_.edge_threshold_ = std::max(params_.edge_threshold_, params_.patch_size_/2 + kKernelWidth / 2 + 2);
|
||||
|
||||
// pre-compute the end of a row in a circular patch
|
||||
int half_patch_size = params_.patch_size_ / 2;
|
||||
u_max_.resize(half_patch_size + 1);
|
||||
for (int v = 0; v <= half_patch_size * sqrt(2.f) / 2 + 1; ++v)
|
||||
u_max_[v] = cvRound(sqrt(float(half_patch_size * half_patch_size - v * v)));
|
||||
|
||||
// Make sure we are symmetric
|
||||
for (int v = half_patch_size, v_0 = 0; v >= half_patch_size * sqrt(2.f) / 2; --v)
|
||||
{
|
||||
while (u_max_[v_0] == u_max_[v_0 + 1])
|
||||
++v_0;
|
||||
u_max_[v] = v_0;
|
||||
++v_0;
|
||||
}
|
||||
|
||||
const int npoints = 512;
|
||||
Point pattern_buf[npoints];
|
||||
const Point* pattern0 = (const Point*)bit_pattern_31_;
|
||||
if( params_.patch_size_ != 31 )
|
||||
{
|
||||
pattern0 = pattern_buf;
|
||||
makeRandomPattern(params_.patch_size_, pattern_buf, npoints);
|
||||
}
|
||||
|
||||
CV_Assert( params_.WTA_K_ == 2 || params_.WTA_K_ == 3 || params_.WTA_K_ == 4 );
|
||||
|
||||
if( params_.WTA_K_ == 2 )
|
||||
std::copy(pattern0, pattern0 + npoints, back_inserter(pattern));
|
||||
else
|
||||
{
|
||||
int ntuples = descriptorSize()*4;
|
||||
initializeOrbPattern(pattern0, pattern, ntuples, params_.WTA_K_, npoints);
|
||||
}
|
||||
}
|
||||
ORB::ORB(int _nfeatures, float _scaleFactor, int _nlevels, int _edgeThreshold,
|
||||
int _firstLevel, int WTA_K, int _scoreType, int _patchSize) :
|
||||
nfeatures(_nfeatures), scaleFactor(_scaleFactor), nlevels(_nlevels),
|
||||
edgeThreshold(_edgeThreshold), firstLevel(_firstLevel), WTA_K(WTA_K),
|
||||
scoreType(_scoreType), patchSize(_patchSize)
|
||||
{}
|
||||
|
||||
|
||||
/** destructor to empty the patterns */
|
||||
ORB::~ORB()
|
||||
{
|
||||
}
|
||||
|
||||
/** returns the descriptor size in bytes */
|
||||
int ORB::descriptorSize() const
|
||||
{
|
||||
return kBytes;
|
||||
}
|
||||
|
||||
int ORB::descriptorType() const
|
||||
{
|
||||
return CV_8U;
|
||||
}
|
||||
|
||||
/** Compute the ORB features and descriptors on an image
|
||||
* @param img the image to compute the features and descriptors on
|
||||
* @param mask the mask to apply
|
||||
* @param keypoints the resulting keypoints
|
||||
*/
|
||||
void ORB::operator()(const Mat &image, const Mat &mask, vector<KeyPoint> & keypoints)
|
||||
void ORB::operator()(InputArray image, InputArray mask, vector<KeyPoint>& keypoints) const
|
||||
{
|
||||
Mat empty_descriptors;
|
||||
this->operator ()(image, mask, keypoints, empty_descriptors, true, false);
|
||||
(*this)(image, mask, keypoints, noArray(), false);
|
||||
}
|
||||
|
||||
/** Compute the ORB features and descriptors on an image
|
||||
* @param img the image to compute the features and descriptors on
|
||||
* @param mask the mask to apply
|
||||
|
||||
/** Compute the ORB keypoint orientations
|
||||
* @param image the image to compute the features and descriptors on
|
||||
* @param integral_image the integral image of the iamge (can be empty, but the computation will be slower)
|
||||
* @param scale the scale at which we compute the orientation
|
||||
* @param keypoints the resulting keypoints
|
||||
* @param descriptors the resulting descriptors
|
||||
* @param useProvidedKeypoints if true, the keypoints are used as an input
|
||||
*/
|
||||
void ORB::operator()(const Mat &image, const Mat &mask, vector<KeyPoint> & keypoints,
|
||||
Mat & descriptors, bool useProvidedKeypoints)
|
||||
static void computeOrientation(const Mat& image, vector<KeyPoint>& keypoints,
|
||||
int halfPatchSize, const vector<int>& umax)
|
||||
{
|
||||
this->operator ()(image, mask, keypoints, descriptors, !useProvidedKeypoints, true);
|
||||
}
|
||||
|
||||
//takes keypoints and culls them by the response
|
||||
static void cull(vector<KeyPoint>& keypoints, size_t n_points)
|
||||
{
|
||||
//this is only necessary if the keypoints size is greater than the number of desired points.
|
||||
if (keypoints.size() > n_points)
|
||||
// Process each keypoint
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints.begin(),
|
||||
keypointEnd = keypoints.end(); keypoint != keypointEnd; ++keypoint)
|
||||
{
|
||||
if (n_points==0) {
|
||||
keypoints.clear();
|
||||
return;
|
||||
}
|
||||
//first use nth element to partition the keypoints into the best and worst.
|
||||
std::nth_element(keypoints.begin(), keypoints.begin() + n_points, keypoints.end(), KeypointResponseGreater());
|
||||
//this is the boundary response, and in the case of FAST may be ambigous
|
||||
float ambiguous_response = keypoints[n_points - 1].response;
|
||||
//use std::partition to grab all of the keypoints with the boundary response.
|
||||
vector<KeyPoint>::const_iterator new_end =
|
||||
std::partition(keypoints.begin() + n_points, keypoints.end(),
|
||||
KeypointResponseGreaterThanThreshold(ambiguous_response));
|
||||
//resize the keypoints, given this new end point. nth_element and partition reordered the points inplace
|
||||
keypoints.resize(new_end - keypoints.begin());
|
||||
keypoint->angle = IC_Angle(image, halfPatchSize, keypoint->pt, umax);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/** Compute the ORB keypoints on an image
|
||||
* @param image_pyramid the image pyramid to compute the features and descriptors on
|
||||
* @param mask_pyramid the masks to apply at every level
|
||||
* @param keypoints the resulting keypoints, clustered per level
|
||||
*/
|
||||
static void computeKeyPoints(const vector<Mat>& imagePyramid,
|
||||
const vector<Mat>& maskPyramid,
|
||||
vector<vector<KeyPoint> >& allKeypoints,
|
||||
int nfeatures, int firstLevel, double scaleFactor,
|
||||
int edgeThreshold, int patchSize, int scoreType )
|
||||
{
|
||||
int nlevels = (int)imagePyramid.size();
|
||||
vector<int> nfeaturesPerLevel(nlevels);
|
||||
|
||||
// fill the extractors and descriptors for the corresponding scales
|
||||
float factor = (float)(1.0 / scaleFactor);
|
||||
float ndesiredFeaturesPerScale = nfeatures*(1 - factor)/(1 - (float)pow((double)factor, (double)nlevels));
|
||||
|
||||
int sumFeatures = 0;
|
||||
for( int level = 0; level < nlevels-1; level++ )
|
||||
{
|
||||
nfeaturesPerLevel[level] = cvRound(ndesiredFeaturesPerScale);
|
||||
sumFeatures += nfeaturesPerLevel[level];
|
||||
ndesiredFeaturesPerScale *= factor;
|
||||
}
|
||||
nfeaturesPerLevel[nlevels-1] = std::max(nfeatures - sumFeatures, 0);
|
||||
|
||||
// Make sure we forget about what is too close to the boundary
|
||||
//edge_threshold_ = std::max(edge_threshold_, patch_size_/2 + kKernelWidth / 2 + 2);
|
||||
|
||||
// pre-compute the end of a row in a circular patch
|
||||
int halfPatchSize = patchSize / 2;
|
||||
vector<int> umax(halfPatchSize + 1);
|
||||
|
||||
int v, v0, vmax = cvFloor(halfPatchSize * sqrt(2.f) / 2 + 1);
|
||||
int vmin = cvCeil(halfPatchSize * sqrt(2.f) / 2);
|
||||
for (v = 0; v <= vmax; ++v)
|
||||
umax[v] = cvRound(sqrt(halfPatchSize * halfPatchSize - v * v));
|
||||
|
||||
// Make sure we are symmetric
|
||||
for (v = halfPatchSize, v0 = 0; v >= vmin; --v)
|
||||
{
|
||||
while (umax[v0] == umax[v0 + 1])
|
||||
++v0;
|
||||
umax[v] = v0;
|
||||
++v0;
|
||||
}
|
||||
|
||||
allKeypoints.resize(nlevels);
|
||||
|
||||
for (int level = 0; level < nlevels; ++level)
|
||||
{
|
||||
int nfeatures = nfeaturesPerLevel[level];
|
||||
allKeypoints[level].reserve(nfeatures*2);
|
||||
|
||||
vector<KeyPoint> & keypoints = allKeypoints[level];
|
||||
|
||||
// Detect FAST features, 20 is a good threshold
|
||||
FastFeatureDetector fd(20, true);
|
||||
fd.detect(imagePyramid[level], keypoints, maskPyramid[level]);
|
||||
|
||||
// Remove keypoints very close to the border
|
||||
KeyPointsFilter::runByImageBorder(keypoints, imagePyramid[level].size(), edgeThreshold);
|
||||
|
||||
if( scoreType == ORB::HARRIS_SCORE )
|
||||
{
|
||||
// Keep more points than necessary as FAST does not give amazing corners
|
||||
KeyPointsFilter::retainBest(keypoints, 2 * nfeatures);
|
||||
|
||||
// Compute the Harris cornerness (better scoring than FAST)
|
||||
HarrisResponses(imagePyramid[level], keypoints, 7, HARRIS_K);
|
||||
}
|
||||
|
||||
//cull to the final desired level, using the new Harris scores or the original FAST scores.
|
||||
KeyPointsFilter::retainBest(keypoints, nfeatures);
|
||||
|
||||
float sf = getScale(level, firstLevel, scaleFactor);
|
||||
|
||||
// Set the level of the coordinates
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints.begin(),
|
||||
keypointEnd = keypoints.end(); keypoint != keypointEnd; ++keypoint)
|
||||
{
|
||||
keypoint->octave = level;
|
||||
keypoint->size = patchSize*sf;
|
||||
}
|
||||
|
||||
computeOrientation(imagePyramid[level], keypoints, halfPatchSize, umax);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/** Compute the ORB decriptors
|
||||
* @param image the image to compute the features and descriptors on
|
||||
* @param integral_image the integral image of the image (can be empty, but the computation will be slower)
|
||||
* @param level the scale at which we compute the orientation
|
||||
* @param keypoints the keypoints to use
|
||||
* @param descriptors the resulting descriptors
|
||||
*/
|
||||
static void computeDescriptors(const Mat& image, vector<KeyPoint>& keypoints, Mat& descriptors,
|
||||
const vector<Point>& pattern, int dsize, int WTA_K)
|
||||
{
|
||||
//convert to grayscale if more than one color
|
||||
CV_Assert(image.type() == CV_8UC1);
|
||||
//create the descriptor mat, keypoints.size() rows, BYTES cols
|
||||
descriptors = Mat::zeros((int)keypoints.size(), dsize, CV_8UC1);
|
||||
|
||||
for (size_t i = 0; i < keypoints.size(); i++)
|
||||
computeOrbDescriptor(keypoints[i], image, &pattern[0], descriptors.ptr((int)i), dsize, WTA_K);
|
||||
}
|
||||
|
||||
|
||||
/** Compute the ORB features and descriptors on an image
|
||||
* @param img the image to compute the features and descriptors on
|
||||
@@ -738,24 +744,25 @@ static void cull(vector<KeyPoint>& keypoints, size_t n_points)
|
||||
* @param do_keypoints if true, the keypoints are computed, otherwise used as an input
|
||||
* @param do_descriptors if true, also computes the descriptors
|
||||
*/
|
||||
void ORB::operator()(const Mat &image_in, const Mat &mask, vector<KeyPoint> & keypoints_in_out,
|
||||
Mat& descriptors, bool do_keypoints, bool do_descriptors)
|
||||
void ORB::operator()( InputArray _image, InputArray _mask, vector<KeyPoint>& _keypoints,
|
||||
OutputArray _descriptors, bool useProvidedKeypoints) const
|
||||
{
|
||||
if (((!do_keypoints) && (!do_descriptors)) || (image_in.empty()))
|
||||
bool do_keypoints = !useProvidedKeypoints;
|
||||
bool do_descriptors = _descriptors.needed();
|
||||
|
||||
if( (!do_keypoints && !do_descriptors) || _image.empty() )
|
||||
return;
|
||||
|
||||
//ROI handling
|
||||
const int HARRIS_BLOCK_SIZE = 9;
|
||||
int half_patch_size = params_.patch_size_ / 2;
|
||||
int border = std::max(params_.edge_threshold_, std::max(half_patch_size, HARRIS_BLOCK_SIZE/2))+1;
|
||||
int halfPatchSize = patchSize / 2;
|
||||
int border = std::max(edgeThreshold, std::max(halfPatchSize, HARRIS_BLOCK_SIZE/2))+1;
|
||||
|
||||
Mat image;
|
||||
if (image_in.type() != CV_8UC1)
|
||||
cvtColor(image_in, image, CV_BGR2GRAY);
|
||||
else
|
||||
image = image_in;
|
||||
Mat image = _image.getMat(), mask = _mask.getMat();
|
||||
if( image.type() != CV_8UC1 )
|
||||
cvtColor(_image, image, CV_BGR2GRAY);
|
||||
|
||||
int n_levels = (int)params_.n_levels_;
|
||||
int nlevels = this->nlevels;
|
||||
|
||||
if( !do_keypoints )
|
||||
{
|
||||
@@ -768,46 +775,46 @@ void ORB::operator()(const Mat &image_in, const Mat &mask, vector<KeyPoint> & ke
|
||||
//
|
||||
// In short, ultimately the descriptor should
|
||||
// ignore octave parameter and deal only with the keypoint size.
|
||||
n_levels = 0;
|
||||
for( size_t i = 0; i < keypoints_in_out.size(); i++ )
|
||||
n_levels = std::max(n_levels, std::max(keypoints_in_out[i].octave, 0));
|
||||
n_levels++;
|
||||
nlevels = 0;
|
||||
for( size_t i = 0; i < _keypoints.size(); i++ )
|
||||
nlevels = std::max(nlevels, std::max(_keypoints[i].octave, 0));
|
||||
nlevels++;
|
||||
}
|
||||
|
||||
// Pre-compute the scale pyramids
|
||||
vector<Mat> image_pyramid(n_levels), mask_pyramid(n_levels);
|
||||
for (int level = 0; level < n_levels; ++level)
|
||||
vector<Mat> imagePyramid(nlevels), maskPyramid(nlevels);
|
||||
for (int level = 0; level < nlevels; ++level)
|
||||
{
|
||||
float scale = 1/get_scale(params_, level);
|
||||
float scale = 1/getScale(level, firstLevel, scale);
|
||||
Size sz(cvRound(image.cols*scale), cvRound(image.rows*scale));
|
||||
Size wholeSize(sz.width + border*2, sz.height + border*2);
|
||||
Mat temp(wholeSize, image.type()), masktemp;
|
||||
image_pyramid[level] = temp(Rect(border, border, sz.width, sz.height));
|
||||
imagePyramid[level] = temp(Rect(border, border, sz.width, sz.height));
|
||||
|
||||
if( !mask.empty() )
|
||||
{
|
||||
masktemp = Mat(wholeSize, mask.type());
|
||||
mask_pyramid[level] = masktemp(Rect(border, border, sz.width, sz.height));
|
||||
maskPyramid[level] = masktemp(Rect(border, border, sz.width, sz.height));
|
||||
}
|
||||
|
||||
// Compute the resized image
|
||||
if (level != (int)params_.first_level_)
|
||||
if( level != firstLevel )
|
||||
{
|
||||
if( level < (int)params_.first_level_ )
|
||||
if( level < firstLevel )
|
||||
{
|
||||
resize(image, image_pyramid[level], sz, scale, scale, INTER_LINEAR);
|
||||
resize(image, imagePyramid[level], sz, scale, scale, INTER_LINEAR);
|
||||
if (!mask.empty())
|
||||
resize(mask, mask_pyramid[level], sz, scale, scale, INTER_LINEAR);
|
||||
copyMakeBorder(image_pyramid[level], temp, border, border, border, border,
|
||||
resize(mask, maskPyramid[level], sz, scale, scale, INTER_LINEAR);
|
||||
copyMakeBorder(imagePyramid[level], temp, border, border, border, border,
|
||||
BORDER_REFLECT_101+BORDER_ISOLATED);
|
||||
}
|
||||
else
|
||||
{
|
||||
float sf = params_.scale_factor_;
|
||||
resize(image_pyramid[level-1], image_pyramid[level], sz, 1./sf, 1./sf, INTER_LINEAR);
|
||||
float sf = scaleFactor;
|
||||
resize(imagePyramid[level-1], imagePyramid[level], sz, 1./sf, 1./sf, INTER_LINEAR);
|
||||
if (!mask.empty())
|
||||
resize(mask_pyramid[level-1], mask_pyramid[level], sz, 1./sf, 1./sf, INTER_LINEAR);
|
||||
copyMakeBorder(image_pyramid[level], temp, border, border, border, border,
|
||||
resize(maskPyramid[level-1], maskPyramid[level], sz, 1./sf, 1./sf, INTER_LINEAR);
|
||||
copyMakeBorder(imagePyramid[level], temp, border, border, border, border,
|
||||
BORDER_REFLECT_101+BORDER_ISOLATED);
|
||||
}
|
||||
}
|
||||
@@ -815,22 +822,24 @@ void ORB::operator()(const Mat &image_in, const Mat &mask, vector<KeyPoint> & ke
|
||||
{
|
||||
copyMakeBorder(image, temp, border, border, border, border,
|
||||
BORDER_REFLECT_101);
|
||||
image.copyTo(image_pyramid[level]);
|
||||
image.copyTo(imagePyramid[level]);
|
||||
if( !mask.empty() )
|
||||
mask.copyTo(mask_pyramid[level]);
|
||||
mask.copyTo(maskPyramid[level]);
|
||||
}
|
||||
|
||||
if( !mask.empty() )
|
||||
copyMakeBorder(mask_pyramid[level], masktemp, border, border, border, border,
|
||||
copyMakeBorder(maskPyramid[level], masktemp, border, border, border, border,
|
||||
BORDER_CONSTANT+BORDER_ISOLATED);
|
||||
}
|
||||
|
||||
// Pre-compute the keypoints (we keep the best over all scales, so this has to be done beforehand
|
||||
vector < vector<KeyPoint> > all_keypoints;
|
||||
if (do_keypoints)
|
||||
vector < vector<KeyPoint> > allKeypoints;
|
||||
if( do_keypoints )
|
||||
{
|
||||
// Get keypoints, those will be far enough from the border that no check will be required for the descriptor
|
||||
computeKeyPoints(image_pyramid, mask_pyramid, all_keypoints);
|
||||
computeKeyPoints(imagePyramid, maskPyramid, allKeypoints,
|
||||
nfeatures, firstLevel, scaleFactor,
|
||||
edgeThreshold, patchSize, scoreType);
|
||||
|
||||
// make sure we have the right number of keypoints keypoints
|
||||
/*vector<KeyPoint> temp;
|
||||
@@ -842,7 +851,7 @@ void ORB::operator()(const Mat &image_in, const Mat &mask, vector<KeyPoint> & ke
|
||||
keypoints.clear();
|
||||
}
|
||||
|
||||
cull(temp, n_features_);
|
||||
KeyPoint::retainBest(temp, n_features_);
|
||||
|
||||
for (vector<KeyPoint>::iterator keypoint = temp.begin(),
|
||||
keypoint_end = temp.end(); keypoint != keypoint_end; ++keypoint)
|
||||
@@ -851,48 +860,72 @@ void ORB::operator()(const Mat &image_in, const Mat &mask, vector<KeyPoint> & ke
|
||||
else
|
||||
{
|
||||
// Remove keypoints very close to the border
|
||||
KeyPointsFilter::runByImageBorder(keypoints_in_out, image.size(), params_.edge_threshold_);
|
||||
KeyPointsFilter::runByImageBorder(_keypoints, image.size(), edgeThreshold);
|
||||
|
||||
// Cluster the input keypoints depending on the level they were computed at
|
||||
all_keypoints.resize(n_levels);
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints_in_out.begin(),
|
||||
keypoint_end = keypoints_in_out.end(); keypoint != keypoint_end; ++keypoint)
|
||||
all_keypoints[keypoint->octave].push_back(*keypoint);
|
||||
allKeypoints.resize(nlevels);
|
||||
for (vector<KeyPoint>::iterator keypoint = _keypoints.begin(),
|
||||
keypointEnd = _keypoints.end(); keypoint != keypointEnd; ++keypoint)
|
||||
allKeypoints[keypoint->octave].push_back(*keypoint);
|
||||
|
||||
// Make sure we rescale the coordinates
|
||||
for (int level = 0; level < n_levels; ++level)
|
||||
for (int level = 0; level < nlevels; ++level)
|
||||
{
|
||||
if (level == (int)params_.first_level_)
|
||||
if (level == firstLevel)
|
||||
continue;
|
||||
|
||||
vector<KeyPoint> & keypoints = all_keypoints[level];
|
||||
float scale = 1/get_scale(params_, level);
|
||||
vector<KeyPoint> & keypoints = allKeypoints[level];
|
||||
float scale = 1/getScale(level, firstLevel, scaleFactor);
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints.begin(),
|
||||
keypoint_end = keypoints.end(); keypoint != keypoint_end; ++keypoint)
|
||||
keypointEnd = keypoints.end(); keypoint != keypointEnd; ++keypoint)
|
||||
keypoint->pt *= scale;
|
||||
}
|
||||
}
|
||||
|
||||
if (do_descriptors)
|
||||
Mat descriptors;
|
||||
vector<Point> pattern;
|
||||
|
||||
if( do_descriptors )
|
||||
{
|
||||
int nkeypoints = 0;
|
||||
for (int level = 0; level < n_levels; ++level)
|
||||
nkeypoints += (int)all_keypoints[level].size();
|
||||
for (int level = 0; level < nlevels; ++level)
|
||||
nkeypoints += (int)allKeypoints[level].size();
|
||||
if( nkeypoints == 0 )
|
||||
descriptors.release();
|
||||
_descriptors.release();
|
||||
else
|
||||
descriptors.create(nkeypoints, descriptorSize(), CV_8U);
|
||||
{
|
||||
_descriptors.create(nkeypoints, descriptorSize(), CV_8U);
|
||||
descriptors = _descriptors.getMat();
|
||||
}
|
||||
|
||||
const int npoints = 512;
|
||||
Point patternbuf[npoints];
|
||||
const Point* pattern0 = (const Point*)bit_pattern_31_;
|
||||
|
||||
if( patchSize != 31 )
|
||||
{
|
||||
pattern0 = patternbuf;
|
||||
makeRandomPattern(patchSize, patternbuf, npoints);
|
||||
}
|
||||
|
||||
CV_Assert( WTA_K == 2 || WTA_K == 3 || WTA_K == 4 );
|
||||
|
||||
if( WTA_K == 2 )
|
||||
std::copy(pattern0, pattern0 + npoints, std::back_inserter(pattern));
|
||||
else
|
||||
{
|
||||
int ntuples = descriptorSize()*4;
|
||||
initializeOrbPattern(pattern0, pattern, ntuples, WTA_K, npoints);
|
||||
}
|
||||
}
|
||||
|
||||
keypoints_in_out.clear();
|
||||
_keypoints.clear();
|
||||
int offset = 0;
|
||||
for (int level = 0; level < n_levels; ++level)
|
||||
for (int level = 0; level < nlevels; ++level)
|
||||
{
|
||||
// Get the features and compute their orientation
|
||||
vector<KeyPoint>& keypoints = all_keypoints[level];
|
||||
vector<KeyPoint>& keypoints = allKeypoints[level];
|
||||
int nkeypoints = (int)keypoints.size();
|
||||
if (nkeypoints==0)
|
||||
continue;
|
||||
|
||||
// Compute the descriptors
|
||||
if (do_descriptors)
|
||||
@@ -900,122 +933,33 @@ void ORB::operator()(const Mat &image_in, const Mat &mask, vector<KeyPoint> & ke
|
||||
Mat desc = descriptors.rowRange(offset, offset + nkeypoints);
|
||||
offset += nkeypoints;
|
||||
// preprocess the resized image
|
||||
Mat& working_mat = image_pyramid[level];
|
||||
Mat& workingMat = imagePyramid[level];
|
||||
//boxFilter(working_mat, working_mat, working_mat.depth(), Size(5,5), Point(-1,-1), true, BORDER_REFLECT_101);
|
||||
GaussianBlur(working_mat, working_mat, Size(7, 7), 2, 2, BORDER_REFLECT_101);
|
||||
computeDescriptors(working_mat, Mat(), level, keypoints, desc);
|
||||
GaussianBlur(workingMat, workingMat, Size(7, 7), 2, 2, BORDER_REFLECT_101);
|
||||
computeDescriptors(workingMat, keypoints, desc, pattern, descriptorSize(), WTA_K);
|
||||
}
|
||||
|
||||
// Copy to the output data
|
||||
if (level != (int)params_.first_level_)
|
||||
if (level != firstLevel)
|
||||
{
|
||||
float scale = get_scale(params_, level);
|
||||
float scale = getScale(level, firstLevel, scaleFactor);
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints.begin(),
|
||||
keypoint_end = keypoints.end(); keypoint != keypoint_end; ++keypoint)
|
||||
keypointEnd = keypoints.end(); keypoint != keypointEnd; ++keypoint)
|
||||
keypoint->pt *= scale;
|
||||
}
|
||||
// And add the keypoints to the output
|
||||
keypoints_in_out.insert(keypoints_in_out.end(), keypoints.begin(), keypoints.end());
|
||||
_keypoints.insert(_keypoints.end(), keypoints.begin(), keypoints.end());
|
||||
}
|
||||
}
|
||||
|
||||
/** Compute the ORB keypoints on an image
|
||||
* @param image_pyramid the image pyramid to compute the features and descriptors on
|
||||
* @param mask_pyramid the masks to apply at every level
|
||||
* @param keypoints the resulting keypoints, clustered per level
|
||||
*/
|
||||
void ORB::computeKeyPoints(const vector<Mat>& image_pyramid,
|
||||
const vector<Mat>& mask_pyramid,
|
||||
vector<vector<KeyPoint> >& all_keypoints_out) const
|
||||
{
|
||||
all_keypoints_out.resize(params_.n_levels_);
|
||||
|
||||
for (int level = 0; level < (int)params_.n_levels_; ++level)
|
||||
{
|
||||
int n_features = n_features_per_level_[level];
|
||||
all_keypoints_out[level].reserve(n_features*2);
|
||||
|
||||
vector<KeyPoint> & keypoints = all_keypoints_out[level];
|
||||
|
||||
// Detect FAST features, 20 is a good threshold
|
||||
FastFeatureDetector fd(20, true);
|
||||
fd.detect(image_pyramid[level], keypoints, mask_pyramid[level]);
|
||||
|
||||
// Remove keypoints very close to the border
|
||||
KeyPointsFilter::runByImageBorder(keypoints, image_pyramid[level].size(), params_.edge_threshold_);
|
||||
|
||||
if( params_.score_type_ == CommonParams::HARRIS_SCORE )
|
||||
{
|
||||
// Keep more points than necessary as FAST does not give amazing corners
|
||||
cull(keypoints, 2 * n_features);
|
||||
|
||||
// Compute the Harris cornerness (better scoring than FAST)
|
||||
HarrisResponses(image_pyramid[level], keypoints, 7, HARRIS_K);
|
||||
}
|
||||
|
||||
//cull to the final desired level, using the new Harris scores or the original FAST scores.
|
||||
cull(keypoints, n_features);
|
||||
|
||||
float sf = get_scale(params_, level);
|
||||
|
||||
// Set the level of the coordinates
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints.begin(),
|
||||
keypoint_end = keypoints.end(); keypoint != keypoint_end; ++keypoint)
|
||||
{
|
||||
keypoint->octave = level;
|
||||
keypoint->size = params_.patch_size_*sf;
|
||||
}
|
||||
|
||||
computeOrientation(image_pyramid[level], Mat(), level, keypoints);
|
||||
}
|
||||
}
|
||||
|
||||
/** Compute the ORB keypoint orientations
|
||||
* @param image the image to compute the features and descriptors on
|
||||
* @param integral_image the integral image of the iamge (can be empty, but the computation will be slower)
|
||||
* @param scale the scale at which we compute the orientation
|
||||
* @param keypoints the resulting keypoints
|
||||
*/
|
||||
void ORB::computeOrientation(const Mat& image, const Mat&, unsigned int /*scale*/,
|
||||
vector<KeyPoint>& keypoints) const
|
||||
void ORB::detectImpl( const Mat& image, vector<KeyPoint>& keypoints, const Mat& mask) const
|
||||
{
|
||||
int half_patch_size = params_.patch_size_/2;
|
||||
|
||||
// Process each keypoint
|
||||
for (vector<KeyPoint>::iterator keypoint = keypoints.begin(),
|
||||
keypoint_end = keypoints.end(); keypoint != keypoint_end; ++keypoint)
|
||||
{
|
||||
keypoint->angle = IC_Angle(image, half_patch_size, keypoint->pt, u_max_);
|
||||
}
|
||||
}
|
||||
(*this)(image, mask, keypoints, noArray(), false);
|
||||
}
|
||||
|
||||
/** Compute the integral image and upadte the cached values
|
||||
* @param image the image to compute the features and descriptors on
|
||||
* @param level the scale at which we compute the orientation
|
||||
* @param descriptors the resulting descriptors
|
||||
*/
|
||||
void ORB::computeIntegralImage(const Mat&, unsigned int, Mat&)
|
||||
void ORB::computeImpl( const Mat& image, vector<KeyPoint>& keypoints, Mat& descriptors) const
|
||||
{
|
||||
}
|
||||
|
||||
/** Compute the ORB decriptors
|
||||
* @param image the image to compute the features and descriptors on
|
||||
* @param integral_image the integral image of the image (can be empty, but the computation will be slower)
|
||||
* @param level the scale at which we compute the orientation
|
||||
* @param keypoints the keypoints to use
|
||||
* @param descriptors the resulting descriptors
|
||||
*/
|
||||
void ORB::computeDescriptors(const Mat& image, const Mat& /*integral_image*/, unsigned int,
|
||||
vector<KeyPoint>& keypoints, Mat& descriptors) const
|
||||
{
|
||||
//convert to grayscale if more than one color
|
||||
CV_Assert(image.type() == CV_8UC1);
|
||||
//create the descriptor mat, keypoints.size() rows, BYTES cols
|
||||
int dsize = descriptorSize();
|
||||
descriptors = Mat::zeros((int)keypoints.size(), dsize, CV_8UC1);
|
||||
|
||||
for (size_t i = 0; i < keypoints.size(); i++)
|
||||
computeOrbDescriptor(keypoints[i], image, &pattern[0], descriptors.ptr((int)i), dsize, params_.WTA_K_);
|
||||
(*this)(image, Mat(), keypoints, descriptors, true);
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user