Added

- Normalised Block Filter
- Gaussian Filter
- Median Filter

PalmettoPython/operations/blur.py

@@ -0,0 +1,103 @@
+"""
+Project: Palmetto
+
+Description: Adds the following operations
+
+Blur            -    computes a blog averaging filter
+GaussianBlur    -    applys a Gaussian filter to the image
+MedianFilter    -    applys a median filter to the image
+
+Created on 8 Oct 2014
+
+Updated __updated__
+
+@author:  Jason Forte
+@contact: <[email protected]>
+"""
+import cv2
+import numpy
+
+from structure.Operation import Operation
+
+class Blur(Operation):
+    '''
+    Blur(input_image, options={'ksize': (15,15)}) -> numpy.ndarray
+    
+    options:
+        ksize    -    dimensions of the block for filetering
+    
+    
+    '''
+
+    NAME = 'Normalised Block Filter'
+    def __init__(self, input_image=numpy.ndarray, options=dict):
+        Operation.__init__(self, input_image=input_image, options=options)
+        if not self.options.has_key('ksize'):
+            self.options['ksize'] = (15, 15)
+
+    def execute(self):
+        return cv2.blur(self.input_image, self.options['ksize'])
+
+
+class MedianFilter(Operation):
+    '''
+    MedianFilter(input_image, options={'ksize': (15,15)}) -> numpy.ndarray
+    
+    options:
+        ksize    -    dimensions of the block for filtering
+    
+    
+    '''
+
+    NAME = 'Median Filter'
+    def __init__(self, input_image=numpy.ndarray, options=dict):
+        Operation.__init__(self, input_image=input_image, options=options)
+        if not self.options.has_key('ksize'):
+            self.options['ksize'] = (15, 15)
+
+    def execute(self):
+        return cv2.medianBlur(self.input_image, self.options['ksize'])
+
+
+
+class GaussBlur(Operation):
+    '''
+    GaussianBlur(input_image, options={'window':(x,y), 's_x':3, 's_y': 3})
+    
+    Applys a gaussian filter with window and sigma values for x and y as s_x and s_y
+    respectively
+    '''
+    NAME = "Gaussian Filter"
+    def __init__(self, input_image=numpy.ndarray, options=dict):
+        Operation.__init__(self, input_image=input_image, options=options)
+
+        if not self.options.has_key('window'):
+            self.options['window'] = (15, 15)
+        if not self.options.has_key('s_x'):
+            self.options['s_x'] = 3
+        if not self.options.has_key('s_y'):
+            self.options['s_y'] = 3
+
+    def execute(self):
+        return cv2.GaussianBlur(self.input_image, self.options['window'], self.options['s_x'], self.options['s_y'])
+
+
+if __name__ == '__main__':
+
+    import structure.Base
+
+    sample_image = structure.Base.sample_dir + 'sample.png'
+
+    img = cv2.imread(sample_image, 0)
+
+    op = GaussBlur(img, options={})
+
+    # op = Blur(img, options={})
+
+    # op = MedianFilter(img, options={'ksize':9})
+
+    result = op.execute()
+    cv2.imshow('Blur Original', img)
+    cv2.imshow('Blur Result', result)
+    cv2.waitKey(0)
+    cv2.destroyAllWindows()