Documentation

Finds edges in an image using the Canny86 algorithm.

Example:

cannyImg
    :: forall shape channels depth
     . (Mat shape channels depth ~ Lambda)
    => Mat shape ('S 1) depth
cannyImg = exceptError $
  canny 30 200 Nothing CannyNormL1 lambda

Determines strong corners on an image.

The function finds the most prominent corners in the image or in the specified image region.

• Function calculates the corner quality measure at every source image pixel using the cornerMinEigenVal or cornerHarris.
• Function performs a non-maximum suppression (the local maximums in 3 x 3 neighborhood are retained).
• The corners with the minimal eigenvalue less than 𝚚𝚞𝚊𝚕𝚒𝚝𝚢𝙻𝚎𝚟𝚎𝚕 * max(x,y) qualityMeasureMap(x,y) are rejected.
• The remaining corners are sorted by the quality measure in the descending order.
• Function throws away each corner for which there is a stronger corner at a distance less than maxDistance.

Example:

goodFeaturesToTrackTraces
    :: forall (width    :: Nat)
              (height   :: Nat)
              (channels :: Nat)
              (depth    :: *)
     . (Mat (ShapeT [height, width]) ('S channels) ('S depth) ~ Frog)
    => Mat (ShapeT [height, width]) ('S channels) ('S depth)
goodFeaturesToTrackTraces = exceptError $ do
  imgG <- cvtColor bgr gray frog
  let features = goodFeaturesToTrack imgG 20 0.01 0.5 Nothing Nothing CornerMinEigenVal
  withMatM (Proxy :: Proxy [height, width])
           (Proxy :: Proxy channels)
           (Proxy :: Proxy depth)
           white $ \imgM -> do
    void $ matCopyToM imgM (V2 0 0) frog Nothing
    forM_ features $ \f -> do
      circle imgM (round <$> f :: V2 Int32) 2 blue 5 LineType_AA 0

Finds circles in a grayscale image using a modification of the Hough transformation.

Example:

houghCircleTraces
    :: forall (width    :: Nat)
              (height   :: Nat)
              (channels :: Nat)
              (depth    :: *)
     . (Mat (ShapeT [height, width]) ('S channels) ('S depth) ~ Circles_1000x625)
    => Mat (ShapeT [height, width]) ('S channels) ('S depth)
houghCircleTraces = exceptError $ do
  imgG <- cvtColor bgr gray circles_1000x625
  let circles = houghCircles 1 10 Nothing Nothing Nothing Nothing imgG
  withMatM (Proxy :: Proxy [height, width])
           (Proxy :: Proxy channels)
           (Proxy :: Proxy depth)
           white $ \imgM -> do
    void $ matCopyToM imgM (V2 0 0) circles_1000x625 Nothing
    forM_ circles $ \c -> do
      circle imgM (round <$> circleCenter c :: V2 Int32) (round (circleRadius c)) blue 1 LineType_AA 0

Example:

houghLinesPTraces
  :: forall (width    :: Nat)
            (height   :: Nat)
            (channels :: Nat)
            (depth    :: *  )
   . (Mat (ShapeT [height, width]) ('S channels) ('S depth) ~ Building_868x600)
  => Mat (ShapeT [height, width]) ('S channels) ('S depth)
houghLinesPTraces = exceptError $ do
    edgeImg <- canny 50 200 Nothing CannyNormL1 building_868x600
    edgeImgBgr <- cvtColor gray bgr edgeImg
    withMatM (Proxy :: Proxy [height, width])
             (Proxy :: Proxy channels)
             (Proxy :: Proxy depth)
             white $ \imgM -> do
      edgeImgM <- thaw edgeImg
      lineSegments <- houghLinesP 1 (pi / 180) 80 (Just 30) (Just 10) edgeImgM
      void $ matCopyToM imgM (V2 0 0) edgeImgBgr Nothing
      forM_ lineSegments $ \lineSegment -> do
        line imgM
             (lineSegmentStart lineSegment)
             (lineSegmentStop  lineSegment)
             red 2 LineType_8 0

A flag, indicating whether to use the more accurate L2 norm or the default L1 norm.