def prune(self):
"""
This function calculates a threshold using the canny algorithm, then using that threshold it finds the necessary
contours of the OpenCV image. Then it iterates through the contours and finds all the contours that have an area
larger than 200, which gave the best results. This also populates a dictionary with all the points corresponding
to each layer. The contours are drawn on a white canvas
:return: Canvas with drawn contours and dictionary points
"""
thresh = cv2.Canny(self.image, 1, 300)
contours, _ = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE)
canvas = np.add(np.zeros(self.image.shape, np.uint8), 255)
layers = []
layer_id = 1
for c in contours:
if cv2.contourArea(c) > 150:
layer = Layer(layer_id)
points_array = []
approx = cv2.approxPolyDP(c, 0.0001*cv2.arcLength(c, True), True)
cv2.drawContours(canvas, [approx], -1, (0, 0, 0), 5)
for point in c:
points_array.append(Point(point[0][0], point[0][1]))
layer.layer_name="test"
layer.points = points_array
layers.append(layer)
# append layer to layers aray
layer_id += 1
return canvas, layers
