Task02 Максим Федоров ITMO

src/phg/matching/descriptor_matcher.cpp

@@ -1,14 +1,31 @@
 #include "descriptor_matcher.h"
 
+#include <algorithm>
+#include <vector>
+
 #include <opencv2/flann/miniflann.hpp>
 #include "flann_factory.h"
 
 void phg::DescriptorMatcher::filterMatchesRatioTest(const std::vector<std::vector<cv::DMatch>> &matches,
                                                     std::vector<cv::DMatch> &filtered_matches)
 {
     filtered_matches.clear();
+    filtered_matches.reserve(matches.size());
+
+    const float ratio = 0.75f;
+
+    for (const std::vector<cv::DMatch>& knn_matches : matches) {
+        if (knn_matches.size() < 2) {
+            continue;
+        }
 
-    throw std::runtime_error("not implemented yet");
+        const cv::DMatch& best = knn_matches[0];
+        const cv::DMatch& second_best = knn_matches[1];
+
+        if (best.distance < ratio * second_best.distance) {
+            filtered_matches.push_back(best);
+        }
+    }
 }
 
 
@@ -35,6 +52,7 @@ void phg::DescriptorMatcher::filterMatchesClusters(const std::vector<cv::DMatch>
         points_query.at<cv::Point2f>(i) = keypoints_query[matches[i].queryIdx].pt;
         points_train.at<cv::Point2f>(i) = keypoints_train[matches[i].trainIdx].pt;
     }
+
 //
 //    // размерность всего 2, так что точное KD-дерево
 //    std::shared_ptr<cv::flann::IndexParams> index_params = flannKdTreeIndexParams(TODO);
@@ -73,4 +91,63 @@ void phg::DescriptorMatcher::filterMatchesClusters(const std::vector<cv::DMatch>
 //
 //    метч остается, если левое и правое множества первых total_neighbors соседей в радиусах поиска(radius2_query, radius2_train) имеют как минимум consistent_matches общих элементов
 //    // TODO заполнить filtered_matches
+
+    std::shared_ptr<cv::flann::IndexParams> index_params = flannKdTreeIndexParams(1);
+    std::shared_ptr<cv::flann::SearchParams> search_params = flannKsTreeSearchParams(64);
+
+    std::shared_ptr<cv::flann::Index> index_query = flannKdTreeIndex(points_query, index_params);
+    std::shared_ptr<cv::flann::Index> index_train = flannKdTreeIndex(points_train, index_params);
+
+    cv::Mat indices_query(n_matches, total_neighbours, CV_32SC1);
+    cv::Mat distances2_query(n_matches, total_neighbours, CV_32FC1);
+    cv::Mat indices_train(n_matches, total_neighbours, CV_32SC1);
+    cv::Mat distances2_train(n_matches, total_neighbours, CV_32FC1);
+
+    index_query->knnSearch(points_query, indices_query, distances2_query, (int)total_neighbours, *search_params);
+    index_train->knnSearch(points_train, indices_train, distances2_train, (int)total_neighbours, *search_params);
+
+    float radius2_query, radius2_train;
+    {
+        std::vector<double> max_dists2_query(n_matches);
+        std::vector<double> max_dists2_train(n_matches);
+        for (int i = 0; i < n_matches; ++i) {
+            max_dists2_query[i] = distances2_query.at<float>(i, (int)total_neighbours - 1);
+            max_dists2_train[i] = distances2_train.at<float>(i, (int)total_neighbours - 1);
+        }
+
+        int median_pos = n_matches / 2;
+        std::nth_element(max_dists2_query.begin(), max_dists2_query.begin() + median_pos, max_dists2_query.end());
+        std::nth_element(max_dists2_train.begin(), max_dists2_train.begin() + median_pos, max_dists2_train.end());
+
+        radius2_query = max_dists2_query[median_pos] * radius_limit_scale * radius_limit_scale;
+        radius2_train = max_dists2_train[median_pos] * radius_limit_scale * radius_limit_scale;
+    }
+
+    filtered_matches.reserve(matches.size());
+
+    std::vector<char> neighbour_in_query(n_matches, 0);
+    for (int i = 0; i < n_matches; ++i) {
+        std::fill(neighbour_in_query.begin(), neighbour_in_query.end(), 0);
+
+        for (size_t j = 0; j < total_neighbours; ++j) {
+            int idx = indices_query.at<int>(i, (int)j);
+            float dist2 = distances2_query.at<float>(i, (int)j);
+            if (idx >= 0 && dist2 <= radius2_query) {
+                neighbour_in_query[idx] = 1;
+            }
+        }
+
+        int consistent_count = 0;
+        for (size_t j = 0; j < total_neighbours; ++j) {
+            int idx = indices_train.at<int>(i, (int)j);
+            float dist2 = distances2_train.at<float>(i, (int)j);
+            if (idx >= 0 && dist2 <= radius2_train && neighbour_in_query[idx]) {
+                ++consistent_count;
+            }
+        }
+
+        if (consistent_count >= (int)consistent_matches) {
+            filtered_matches.push_back(matches[i]);
+        }
+    }
 }

src/phg/matching/flann_matcher.cpp

@@ -1,21 +1,54 @@
 #include <iostream>
+#include <cmath>
 #include "flann_matcher.h"
 #include "flann_factory.h"
 
 
 phg::FlannMatcher::FlannMatcher()
 {
     // параметры для приближенного поиска
-//    index_params = flannKdTreeIndexParams(TODO);
-//    search_params = flannKsTreeSearchParams(TODO);
+    index_params = flannKdTreeIndexParams(4);
+    search_params = flannKsTreeSearchParams(32);
 }
 
 void phg::FlannMatcher::train(const cv::Mat &train_desc)
 {
+    index_params = flannKdTreeIndexParams(4);
+    search_params = flannKsTreeSearchParams(32);
+
     flann_index = flannKdTreeIndex(train_desc, index_params);
 }
 
 void phg::FlannMatcher::knnMatch(const cv::Mat &query_desc, std::vector<std::vector<cv::DMatch>> &matches, int k) const
 {
-    throw std::runtime_error("not implemented yet");
+    matches.clear();
+    matches.resize(query_desc.rows);
+
+    if (query_desc.rows == 0) {
+        return;
+    }
+
+    cv::Mat indices;
+    cv::Mat distances;
+    flann_index->knnSearch(query_desc, indices, distances, k);
+
+    for (int qi = 0; qi < query_desc.rows; ++qi) {
+        std::vector<cv::DMatch>& dst = matches[qi];
+        dst.clear();
+        dst.reserve(k);
+
+        for (int j = 0; j < k; ++j) {
+            int train_idx = indices.at<int>(qi, j);
+            if (train_idx < 0) {
+                continue;
+            }
+
+            cv::DMatch match;
+            match.distance = std::sqrt(distances.at<float>(qi, j));
+            match.imgIdx = 0;
+            match.queryIdx = qi;
+            match.trainIdx = train_idx;
+            dst.push_back(match);
+        }
+    }
 }

src/phg/sfm/homography.cpp

@@ -64,6 +64,7 @@ namespace {
     {
         std::vector<std::vector<double>> A;
         std::vector<double> H;
+        A.reserve(8);
 
         double xs0[4] = {l0.x, l1.x, l2.x, l3.x};
         double xs1[4] = {r0.x, r1.x, r2.x, r3.x};
@@ -84,8 +85,9 @@ namespace {
             double w1 = ws1[i];
 
             // 8 elements of matrix + free term as needed by gauss routine
-//            A.push_back({TODO});
-//            A.push_back({TODO});
+            (void)w1;
+            A.push_back({x0, y0, w0, 0.0, 0.0, 0.0, -x0 * x1, -y0 * x1, x1 * w0});
+            A.push_back({0.0, 0.0, 0.0, x0, y0, w0, -x0 * y1, -y0 * y1, y1 * w0});
         }
 
         int res = gauss(A, H);
@@ -168,57 +170,60 @@ namespace {
         // * (простое описание для понимания)
         // * [3] http://ikrisoft.blogspot.com/2015/01/ransac-with-contrario-approach.html
 
-//        const int n_matches = points_lhs.size();
-//
-//        // https://en.wikipedia.org/wiki/Random_sample_consensus#Parameters
-//        const int n_trials = TODO;
-//
-//        const int n_samples = TODO;
-//        uint64_t seed = 1;
-//        const double reprojection_error_threshold_px = 2;
-//
-//        int best_support = 0;
-//        cv::Mat best_H;
-//
-//        std::vector<int> sample;
-//        for (int i_trial = 0; i_trial < n_trials; ++i_trial) {
-//            randomSample(sample, n_matches, n_samples, &seed);
-//
-//            cv::Mat H = estimateHomography4Points(points_lhs[sample[0]], points_lhs[sample[1]], points_lhs[sample[2]], points_lhs[sample[3]],
-//                                                  points_rhs[sample[0]], points_rhs[sample[1]], points_rhs[sample[2]], points_rhs[sample[3]]);
-//
-//            int support = 0;
-//            for (int i_point = 0; i_point < n_matches; ++i_point) {
-//                try {
-//                    cv::Point2d proj = phg::transformPoint(points_lhs[i_point], H);
-//                    if (cv::norm(proj - cv::Point2d(points_rhs[i_point])) < reprojection_error_threshold_px) {
-//                        ++support;
-//                    }
-//                } catch (const std::exception &e)
-//                {
-//                    std::cerr << e.what() << std::endl;
-//                }
-//            }
-//
-//            if (support > best_support) {
-//                best_support = support;
-//                best_H = H;
-//
-//                std::cout << "estimateHomographyRANSAC : support: " << best_support << "/" << n_matches << std::endl;
-//
-//                if (best_support == n_matches) {
-//                    break;
-//                }
-//            }
-//        }
-//
-//        std::cout << "estimateHomographyRANSAC : best support: " << best_support << "/" << n_matches << std::endl;
-//
-//        if (best_support == 0) {
-//            throw std::runtime_error("estimateHomographyRANSAC : failed to estimate homography");
-//        }
-//
-//        return best_H;
+        const int n_matches = points_lhs.size();
+
+        // https://en.wikipedia.org/wiki/Random_sample_consensus#Parameters
+        const int n_trials = 1000;
+
+        const int n_samples = 4;
+        uint64_t seed = 1;
+        const double reprojection_error_threshold_px = 2.0;
+
+        int best_support = 0;
+        cv::Mat best_H;
+
+        std::vector<int> sample;
+        for (int i_trial = 0; i_trial < n_trials; ++i_trial) {
+            randomSample(sample, n_matches, n_samples, &seed);
+
+            cv::Mat H;
+            try {
+                H = estimateHomography4Points(points_lhs[sample[0]], points_lhs[sample[1]], points_lhs[sample[2]], points_lhs[sample[3]],
+                                              points_rhs[sample[0]], points_rhs[sample[1]], points_rhs[sample[2]], points_rhs[sample[3]]);
+            } catch (const std::exception &) {
+                continue;
+            }
+
+            int support = 0;
+            for (int i_point = 0; i_point < n_matches; ++i_point) {
+                try {
+                    cv::Point2d proj = phg::transformPoint(points_lhs[i_point], H);
+                    if (cv::norm(proj - cv::Point2d(points_rhs[i_point])) < reprojection_error_threshold_px) {
+                        ++support;
+                    }
+                } catch (const std::exception &) {
+                }
+            }
+
+            if (support > best_support) {
+                best_support = support;
+                best_H = H;
+
+                std::cout << "estimateHomographyRANSAC : support: " << best_support << "/" << n_matches << std::endl;
+
+                if (best_support == n_matches) {
+                    break;
+                }
+            }
+        }
+
+        std::cout << "estimateHomographyRANSAC : best support: " << best_support << "/" << n_matches << std::endl;
+
+        if (best_support == 0) {
+            throw std::runtime_error("estimateHomographyRANSAC : failed to estimate homography");
+        }
+
+        return best_H;
     }
 
 }
@@ -238,7 +243,14 @@ cv::Mat phg::findHomographyCV(const std::vector<cv::Point2f> &points_lhs, const
 // таким преобразованием внутри занимается функции cv::perspectiveTransform и cv::warpPerspective
 cv::Point2d phg::transformPoint(const cv::Point2d &pt, const cv::Mat &T)
 {
-    throw std::runtime_error("not implemented yet");
+    double x = pt.x;
+    double y = pt.y;
+
+    double tx = T.at<double>(0, 0) * x + T.at<double>(0, 1) * y + T.at<double>(0, 2);
+    double ty = T.at<double>(1, 0) * x + T.at<double>(1, 1) * y + T.at<double>(1, 2);
+    double tw = T.at<double>(2, 0) * x + T.at<double>(2, 1) * y + T.at<double>(2, 2);
+
+    return cv::Point2d(tx / tw, ty / tw);
 }
 
 cv::Point2d phg::transformPointCV(const cv::Point2d &pt, const cv::Mat &T) {

src/phg/sfm/panorama_stitcher.cpp

@@ -3,6 +3,7 @@
 
 #include <libutils/bbox2.h>
 #include <iostream>
+#include <vector>
 
 /*
  * imgs - список картинок
@@ -23,7 +24,44 @@ cv::Mat phg::stitchPanorama(const std::vector<cv::Mat> &imgs,
     {
         // здесь надо посчитать вектор Hs
         // при этом можно обойтись n_images - 1 вызовами функтора homography_builder
-        throw std::runtime_error("not implemented yet");
+        std::vector<cv::Mat> edge_H(n_images);
+        std::vector<char> edge_ready(n_images, 0);
+        std::vector<char> H_ready(n_images, 0);
+
+        for (int i = 0; i < n_images; ++i) {
+            if (parent[i] == -1) {
+                Hs[i] = cv::Mat::eye(3, 3, CV_64FC1);
+                H_ready[i] = 1;
+            }
+        }
+
+        for (int i = 0; i < n_images; ++i) {
+            if (H_ready[i]) {
+                continue;
+            }
+
+            std::vector<int> chain;
+            int cur = i;
+            while (!H_ready[cur]) {
+                chain.push_back(cur);
+
+                int p = parent[cur];
+
+                if (!edge_ready[cur]) {
+                    edge_H[cur] = homography_builder(imgs[cur], imgs[p]);
+                    edge_ready[cur] = 1;
+                }
+
+                cur = p;
+            }
+
+            for (int k = (int)chain.size() - 1; k >= 0; --k) {
+                int v = chain[k];
+                int p = parent[v];
+                Hs[v] = Hs[p] * edge_H[v];
+                H_ready[v] = 1;
+            }
+        }
     }
 
     bbox2<double, cv::Point2d> bbox;