Fix canny

modules/core/src/image/vpCannyEdgeDetection.cpp

@@ -235,7 +235,7 @@ vpCannyEdgeDetection::performFilteringAndGradientComputation(const vpImage<unsig
 /**
  * \brief Get the interpolation weights and offsets.
  *
- * \param[in] absoluteTheta : The absolute value of the angle of the edge, expressed in degrees.
+ * \param[in] positiveTheta : The positive value of the angle of the edge, expressed in degrees.
  * \param[out] alpha : The weight of the first point used for the interpolation.
  * \param[out] beta : The weight of the second point used for the interpolation.
  * \param[out] dRowGradAlpha : The offset along the row attached to the alpha weight.
@@ -244,45 +244,45 @@ vpCannyEdgeDetection::performFilteringAndGradientComputation(const vpImage<unsig
  * \param[out] dColGradBeta : The offset along the column attached to the beta weight.
  */
 void
-getInterpolationWeightsAndOffsets(const float &absoluteTheta,
+getInterpolationWeightsAndOffsets(const float &positiveTheta,
                  float &alpha, float &beta,
                  int &dRowGradAlpha, int &dRowGradBeta,
                  int &dColGradAlpha, int &dColGradBeta
 )
 {
   float thetaMin = 0.f;
-  if (absoluteTheta < 45.f) {
+  if (positiveTheta < 45.f) {
     // Angles between 0 and 45 deg rely on the horizontal and diagonal points
     dColGradAlpha = 1;
     dColGradBeta = 1;
     dRowGradAlpha = 0;
     dRowGradBeta = -1;
   }
-  else if (absoluteTheta >= 45.f && absoluteTheta < 90.f) {
+  else if (positiveTheta >= 45.f && positiveTheta < 90.f) {
     // Angles between 45 and 90 deg rely on the diagonal and vertical points
     thetaMin = 45.f;
     dColGradAlpha = 1;
     dColGradBeta = 0;
     dRowGradAlpha = -1;
     dRowGradBeta = -1;
   }
-  else if (absoluteTheta >= 90.f && absoluteTheta < 135.f) {
+  else if (positiveTheta >= 90.f && positiveTheta < 135.f) {
     // Angles between 90 and 135 deg rely on the vertical and diagonal points
     thetaMin = 90.f;
     dColGradAlpha = 0;
     dColGradBeta = -1;
     dRowGradAlpha = -1;
     dRowGradBeta = -1;
   }
-  else if (absoluteTheta >= 135.f && absoluteTheta < 180.f) {
+  else if (positiveTheta >= 135.f && positiveTheta < 180.f) {
     // Angles between 135 and 180 deg rely on the vertical and diagonal points
     thetaMin = 135.f;
     dColGradAlpha = -1;
     dColGradBeta = -1;
     dRowGradAlpha = -1;
     dRowGradBeta = 0;
   }
-  beta = (absoluteTheta - thetaMin) / 45.f;
+  beta = (positiveTheta - thetaMin) / 45.f;
   alpha = 1.f - beta;
 }
 
@@ -314,28 +314,35 @@ getManhattanGradient(const vpImage<float> &dIx, const vpImage<float> &dIy, const
 }
 
 /**
- * @brief Get the absolute value of the gradient orientation.
+ * @brief Get the gradient orientation, expressed in degrees, between 0 and +180 degrees.
+ * If the gradient orientation is negative, we add 180 degrees (i.e. M_PI radiants) in
+ * order to keep the same orientation but in the positive direction.
  *
  * @param dIx : Gradient along the horizontal axis.
  * @param dIy : Gradient along the vertical axis.
  * @param row : Index along the vertical axis.
  * @param col : Index along the horizontal axis.
- * @return float The absolute value of the gradient orientation, expressed in degrees.
+ * @return float The positive value of the gradient orientation, expressed in degrees.
  */
 float
-getAbsoluteTheta(const vpImage<float> &dIx, const vpImage<float> &dIy, const int &row, const int &col)
+getPositiveTheta(const vpImage<float> &dIx, const vpImage<float> &dIy, const int &row, const int &col)
 {
-  float absoluteTheta = 0.f;
+  float positiveTheta = 0.f;
   float dx = dIx[row][col];
   float dy = dIy[row][col];
 
   if (std::abs(dx) < std::numeric_limits<float>::epsilon()) {
-    absoluteTheta = 90.f;
+    positiveTheta = 90.f;
   }
   else {
-    absoluteTheta = static_cast<float>(vpMath::deg(std::abs(std::atan2(dy, dx))));
+    // -dy because the y-axis of the image is oriented towards the bottom of the screen
+    // while we later work with a y-axis oriented towards the top when getting the theta quadrant.
+    positiveTheta = static_cast<float>(vpMath::deg(std::atan2(-dy , dx)));
+    if(positiveTheta < 0.f) {
+      positiveTheta += 180.f; // + M_PI in order to be between 0 and M_PI
+    }
   }
-  return absoluteTheta;
+  return positiveTheta;
 }
 
 void
@@ -358,9 +365,9 @@ vpCannyEdgeDetection::performEdgeThinning(const float &lowerThreshold)
       // depending on the gradient orientation
       int dRowAlphaPlus = 0, dRowBetaPlus = 0;
       int dColAphaPlus = 0, dColBetaPlus = 0;
-      float absTheta = getAbsoluteTheta(m_dIx, m_dIy, row, col);
+      float positiveTheta = getPositiveTheta(m_dIx, m_dIy, row, col);
       float alpha = 0.f, beta = 0.f;
-      getInterpolationWeightsAndOffsets(absTheta, alpha, beta, dRowAlphaPlus, dRowBetaPlus, dColAphaPlus, dColBetaPlus);
+      getInterpolationWeightsAndOffsets(positiveTheta, alpha, beta, dRowAlphaPlus, dRowBetaPlus, dColAphaPlus, dColBetaPlus);
       int dRowAlphaMinus = -dRowAlphaPlus, dRowBetaMinus = -dRowBetaPlus;
       int dColAphaMinus = -dColAphaPlus, dColBetaMinus = -dColBetaPlus;
       float gradAlphaPlus = getManhattanGradient(m_dIx, m_dIy, row + dRowAlphaPlus, col + dColAphaPlus);

modules/core/src/image/vpImageFilter.cpp

@@ -684,7 +684,7 @@ float vpImageFilter::computeCannyThreshold(const cv::Mat &cv_I, const cv::Mat *p
   const float range[] = { 0.f, 256.f }; // The upper boundary is exclusive
   const float *ranges[] = { range };
   int channels[] = { 0 };
-  bool dims = 1; // The number of dimensions of the histogram
+  int dims = 1; // The number of dimensions of the histogram
   int histSize[] = { bins };
   bool uniform = true;
   bool accumulate = false; // Clear the histogram at the beginning of calcHist if false, does not clear it otherwise
@@ -741,14 +741,14 @@ void vpImageFilter::computePartialDerivatives(const cv::Mat &cv_I,
       if (normalize) {
         scale = 1. / 8.;
         if (apertureGradient > 3) {
-          scale *= std::pow(1./16., ((apertureGradient -1.)/2.) - 1.);
+          scale *= std::pow(1./2., (apertureGradient * 2. - 3.)); // 1 / 2^(2 x ksize - dx - dy -2) with ksize =apertureGradient and dx xor dy = 1
         }
       }
       if (computeDx) {
-        cv::Sobel(img_blur, cv_dIx, CV_16S, 1, 0, apertureGradient, 1, 0, scale);
+        cv::Sobel(img_blur, cv_dIx, CV_16S, 1, 0, apertureGradient, scale, 0., cv::BORDER_REPLICATE);
       }
       if (computeDy) {
-        cv::Sobel(img_blur, cv_dIy, CV_16S, 0, 1, apertureGradient, 1, 0, scale);
+        cv::Sobel(img_blur, cv_dIy, CV_16S, 0, 1, apertureGradient, scale, 0., cv::BORDER_REPLICATE);
       }
     }
     else if (filteringType == vpImageFilter::CANNY_GBLUR_SCHARR_FILTERING) {