Documentation

Resizes an image

To shrink an image, it will generally look best with InterArea interpolation, whereas to enlarge an image, it will generally look best with InterCubic (slow) or InterLinear (faster but still looks OK).

Example:

resizeInterAreaImg :: Mat ('S ['D, 'D]) ('S 3) ('S Word8)
resizeInterAreaImg = exceptError $
    withMatM (h ::: w + (w `div` 2) ::: Z)
             (Proxy :: Proxy 3)
             (Proxy :: Proxy Word8)
             transparent $ \imgM -> do
      birds_resized <-
        pureExcept $ resize (ResizeRel $ pure 0.5) InterArea birds_768x512
      matCopyToM imgM (V2 0 0) birds_768x512 Nothing
      matCopyToM imgM (V2 w 0) birds_resized Nothing
      lift $ arrowedLine imgM (V2 startX y) (V2 pointX y) red 4 LineType_8 0 0.15
  where
    [h, w] = miShape $ matInfo birds_768x512
    startX = round $ fromIntegral w * (0.95 :: Double)
    pointX = round $ fromIntegral w * (1.05 :: Double)
    y = h `div` 4

OpenCV Sphinx doc

Applies an affine transformation to an image

Example:

rotateBirds :: Mat (ShapeT [2, 3]) ('S 1) ('S Double)
rotateBirds = getRotationMatrix2D (V2 256 170 :: V2 CFloat) 45 0.75

warpAffineImg :: Kodak_512x341
warpAffineImg = exceptError $
    warpAffine birds_512x341 rotateBirds InterArea False False (BorderConstant black)

warpAffineInvImg :: Kodak_512x341
warpAffineInvImg = exceptError $
    warpAffine warpAffineImg rotateBirds InterCubic True False (BorderConstant black)

OpenCV Sphinx doc

Applies a perspective transformation to an image

OpenCV Sphinx doc

Inverts an affine transformation

OpenCV Sphinx doc

Calculates a perspective transformation matrix for 2D perspective transform

OpenCV Sphinx doc

Calculates an affine matrix of 2D rotation

OpenCV Sphinx doc

Applies a generic geometrical transformation to an image.

The function remap transforms the source image using the specified map:

dst(x,y) = src(map(x,y))

Example:

remapImg
  :: forall (width    :: Nat)
            (height   :: Nat)
            (channels :: Nat)
            (depth    :: *  )
   . (Mat ('S ['S height, 'S width]) ('S channels) ('S depth) ~ Kodak_512x341)
  => Mat ('S ['S height, 'S width]) ('S channels) ('S depth)
remapImg = exceptError $ remap birds_512x341 transform InterLinear (BorderConstant black)
  where
    transform = exceptError $
                matFromFunc (Proxy :: Proxy [height, width])
                            (Proxy :: Proxy 2)
                            (Proxy :: Proxy Float)
                            exampleFunc

    exampleFunc [_y,  x] 0 = wobble x w
    exampleFunc [ y, _x] 1 = wobble y h
    exampleFunc _pos _channel = error "impossible"

    wobble :: Int -> Float -> Float
    wobble v s = let v' = fromIntegral v
                     n = v' / s
                 in v' + (s * 0.05 * sin (n * 2 * pi * 5))

    w = fromInteger $ natVal (Proxy :: Proxy width)
    h = fromInteger $ natVal (Proxy :: Proxy height)

OpenCV documentation

The function transforms an image to compensate radial and tangential lens distortion.

Those pixels in the destination image, for which there is no correspondent pixels in the source image, are filled with zeros (black color).

The camera matrix and the distortion parameters can be determined using calibrateCamera . If the resolution of images is different from the resolution used at the calibration stage, f_x, f_y, c_x and c_y need to be scaled accordingly, while the distortion coefficients remain the same.

Example:

undistortImg
  :: forall (width    :: Nat)
            (height   :: Nat)
            (channels :: Nat)
            (depth    :: *  )
   . (Mat ('S ['S height, 'S width]) ('S channels) ('S depth) ~ Kodak_512x341)
  => Mat ('S ['S height, 'S width]) ('S channels) ('S depth)
undistortImg = undistort birds_512x341 intrinsics coefficients
  where intrinsics :: M33 Float
        intrinsics =
          V3 (V3 15840.8      0      2049)
             (V3     0    15830.3    1097)
             (V3     0        0         1)

        coefficients :: Matx51d
        coefficients = unsafePerformIO $
          newMatx51d
            (-2.239145913492247)
             13.674526561736648
              3.650187848850095e-2
            (-2.0042015752853796e-2)
            (-0.44790921357620456)