Programming for Engineers in Python
           Recitation 12 
         Image Processing 
               
        Plan: Image Processing with numpy 
           Binary segmentation 
           Image gradient 
           Image brightening 
           Morphological operators  
             Erosion 
             Dilation 
             Smoothing 
           Denoising 
           Ternary segmentation 
   2 
                                                                                      A 2D table of values 
                                                                                      (pixels), each in 
             Grayscale Image                                                          0..255: 
                                                                                      •  0 = Black 
                                                                                      •  255 = White 
         105   114  116   116  160   121   97   90   124  119   188  144   112  116   78    32   19   61    40 
     3 
                Image processing – basic functions
        Reading an image from disk: 
            from scipy import misc 
            im = misc.imread('C:/Koala.jpg') 
        Creating an “empty” image matrix: 
            im = numpy.zeros( (height,width), dtype=numpy.uint8 ) 
        Important: each pixel is in the range 0-255 (type np.uint8) 
            o Numerical operations cause overflow, e.g.: 
                numpy.uint8(200) + numpy.uint8(100) = 44 = 300 mod 256 
            o Therefore, before doing numerical operations on image pixels, convert the 
                whole image or a specific pixel to int 32-bit using numpy.int_ : 
                a = numpy.uint8(200) ; b = numpy.uint8(100) 
                numpy.int_(a) + numpy.int_(b) = 300 
      4