- Code cleanup on robust_estimation

Lancelot899 · Apr 19, 2016 · c17600c · c17600c
1 parent dfb460f
commit c17600c
Show file tree

Hide file tree

Showing 10 changed files with 18 additions and 15 deletions.
diff --git a/src/openMVG/robust_estimation/guided_matching.hpp b/src/openMVG/robust_estimation/guided_matching.hpp
@@ -368,9 +368,8 @@ void GuidedMatching_Fundamental_Fast(
       itBs != buckets.begin() + bucket_stop; ++itBs)
     {
       const Bucket_vec & bucket = *itBs;
-      for(Bucket_vec::const_iterator itB = bucket.begin(); itB != bucket.end(); ++itB)
+      for( const auto & i : bucket )
       {
-        const IndexT i = *itB;
         // Compute descriptor distance
         const double descDist = lRegions.SquaredDescriptorDistance(i, &rRegions, j);
         // Update the corresponding points & distance (if required)

diff --git a/src/openMVG/robust_estimation/rand_sampling.hpp b/src/openMVG/robust_estimation/rand_sampling.hpp
@@ -35,10 +35,11 @@
 namespace openMVG {
 namespace robust{
 
+// rperrot : is this legal ? why using an anonymous namespace here ? 
 namespace
 {
   std::default_random_engine random_generator;
-}
+} // 
 
 /**
 * Pick a random subset of the integers [0, total), in random order.

diff --git a/src/openMVG/robust_estimation/robust_estimator_ACRansac.hpp b/src/openMVG/robust_estimation/robust_estimator_ACRansac.hpp
@@ -49,7 +49,7 @@
 namespace openMVG {
 namespace robust{
 
-namespace acrancac_nfa_internal {
+namespace acransac_nfa_internal {
 
 /// logarithm (base 10) of binomial coefficient
 template <typename T>
@@ -368,7 +368,7 @@ std::pair<double, double> ACRANSAC(const Kernel &kernel,
 
   // Initialize the NFA computation interface
   // (quantified NFA computation is used if a valid upper bound is provided)
-  acrancac_nfa_internal::NFA_Interface<Kernel> nfa_interface
+  acransac_nfa_internal::NFA_Interface<Kernel> nfa_interface
     (kernel, maxThreshold, (precision!=std::numeric_limits<double>::infinity()));
 
   // Output parameters

diff --git a/src/openMVG/robust_estimation/robust_estimator_ACRansacKernelAdaptator.hpp b/src/openMVG/robust_estimation/robust_estimator_ACRansacKernelAdaptator.hpp
@@ -27,6 +27,8 @@
 //  by the generic ACRANSAC routine.
 //
 
+#include "openMVG/numeric/numeric.h"
+
 namespace openMVG {
 namespace robust{
 

diff --git a/src/openMVG/robust_estimation/robust_estimator_LMeds.hpp b/src/openMVG/robust_estimation/robust_estimator_LMeds.hpp
@@ -28,8 +28,8 @@ namespace robust{
 /// IJCV 1998
 template <typename Kernel>
   double LeastMedianOfSquares(const Kernel &kernel,
-	  typename Kernel::Model * model = NULL,
-    double* outlierThreshold = NULL,
+	  typename Kernel::Model * model = nullptr ,
+    double* outlierThreshold = nullptr ,
     double outlierRatio=0.5,
 	  double minProba=0.99)
 {
@@ -63,7 +63,7 @@ template <typename Kernel>
       }
 
 			// Compute median
-			std::vector<double>::iterator itMedian = residuals.begin() +
+			auto itMedian = residuals.begin() +
 				std::size_t( total_samples*(1.-outlierRatio) );
 			std::nth_element(residuals.begin(), itMedian, residuals.end());
 			double median = *itMedian;

diff --git a/src/openMVG/robust_estimation/robust_estimator_MaxConsensus.hpp b/src/openMVG/robust_estimation/robust_estimator_MaxConsensus.hpp
@@ -33,7 +33,7 @@ namespace robust{
 template<typename Kernel, typename Scorer>
 typename Kernel::Model MaxConsensus(const Kernel &kernel,
   const Scorer &scorer,
-  std::vector<size_t> *best_inliers = NULL, size_t max_iteration = 1024) {
+  std::vector<size_t> *best_inliers = nullptr , size_t max_iteration = 1024) {
 
     const size_t min_samples = Kernel::MINIMUM_SAMPLES;
     const size_t total_samples = kernel.NumSamples();

diff --git a/src/openMVG/robust_estimation/robust_estimator_MaxConsensus_test.cpp b/src/openMVG/robust_estimation/robust_estimator_MaxConsensus_test.cpp
@@ -105,7 +105,7 @@ TEST(MaxConsensusLineFitter, RealisticCase) {
 
   //-- Add some noise (for the asked percentage amount)
   int nbPtToNoise = (int) NbPoints*inlierPourcentAmount/100.0;
-  vector<size_t> vec_samples; // Fit with unique random index
+  std::vector<size_t> vec_samples; // Fit with unique random index
   UniformSample(nbPtToNoise, NbPoints, &vec_samples);
   for(size_t i = 0; i <vec_samples.size(); ++i)
   {

diff --git a/src/openMVG/robust_estimation/robust_estimator_Ransac.hpp b/src/openMVG/robust_estimation/robust_estimator_Ransac.hpp
@@ -13,6 +13,7 @@
 #include <limits>
 #include <numeric>
 #include <vector>
+#include <cassert>
 
 namespace openMVG {
 namespace robust{
@@ -33,8 +34,8 @@ template<typename Kernel, typename Scorer>
 typename Kernel::Model RANSAC(
   const Kernel &kernel,
   const Scorer &scorer,
-  std::vector<size_t> *best_inliers = NULL,
-  double *best_score = NULL,
+  std::vector<size_t> *best_inliers = nullptr ,
+  double *best_score = nullptr ,
   double outliers_probability = 1e-2)
 {
   assert(outliers_probability < 1.0);

diff --git a/src/openMVG/robust_estimation/robust_estimator_Ransac_test.cpp b/src/openMVG/robust_estimation/robust_estimator_Ransac_test.cpp
@@ -87,7 +87,7 @@ TEST(MaxConsensusLineFitter, RealisticCase) {
 
   //-- Add some noise (for the asked percentage amount)
   int nbPtToNoise = (int) NbPoints*inlierPourcentAmount/100.0;
-  vector<size_t> vec_samples; // Fit with unique random index
+  std::vector<size_t> vec_samples; // Fit with unique random index
   UniformSample(nbPtToNoise, NbPoints, &vec_samples);
   for(size_t i = 0; i <vec_samples.size(); ++i)
   {

diff --git a/src/openMVG/robust_estimation/score_evaluator.hpp b/src/openMVG/robust_estimation/score_evaluator.hpp
@@ -28,11 +28,11 @@
 #ifndef OPENMVG_ROBUST_ESTIMATION_SCORE_EVALUATOR_H_
 #define OPENMVG_ROBUST_ESTIMATION_SCORE_EVALUATOR_H_
 
+#include <vector>
+
 namespace openMVG {
 namespace robust{
 
-using namespace std;
-
 /// Templated Functor class to evaluate a given model over a set of samples.
 template<typename Kernel>
 class ScorerEvaluator {