ovl.image_filters.image_filter module

ovl.image_filters.image_filter.image_filter(image_filter_function)

A decorator used to pass parameters in two stages, run in a python console “help(ovl.contour_filter)” or ovl.contour_filter.__doc__ or look at ovl.contour_filter documentation for more information image_filter acts like contour_filter except it receives an image (np.ndarray) applies changes on it and then returns it altered. functions decorated by image_filter can be passed to Vision(image_filters_)

Changes that can be applied vary from various blurs like Gaussian blur or Bilateral filter to image rotations translations, contrast and brightness alterations and color spaces changes (HSV, RGB Greyscale conversions)

Filters that change binary images (images that have been threshold by a threshold object such as ovl.Color ovl.MultiColor or ovl.Canny) like morphological functions (ex. erosion, dilation) count as image filters as well and should be decorated using image_filter.

image filter example:

@image_filter
def rotate_by_angle(image, angle):
    (height, width) = image.shape[:2]
    center_xy = (width / 2, height / 2)
    rotation_matrix = cv2.getRotationMatrix2D(center_xy, -angle, 1.0)
    cos = np.abs(rotation_matrix[0, 0])
    sin = np.abs(rotation_matrix[0, 1])
    rotated_image_dimensions = (int((height * sin) + (width * cos)), int((height * cos) + (width * sin)))
    rotation_matrix[0, 2] += (rotated_image_dimensions[0] / 2) - center_xy[0]
    rotation_matrix[1, 2] += (rotated_image_dimensions[1] / 2) - center_xy[1]
    return cv2.warpAffine(image, rotation_matrix, rotated_image_dimensions)

The filter can then be passed to a Vision object to be applied to images in the pipeline.

pipeline = Vision(image_filters=[rotate_by_angle(angle=180)]