Feature/precise positioning

This MR adds the precise positioning functionality of OpenCV found by @hleistra as a function and includes tests for this function. The way the function works is as follows: the function takes an image and optionally the corner points of either the image itself or reference image or both. it then warps the image so that the corner markers align with the reference markers. It finishes by returning the image and new corner markers. This function makes use of the fix_corner_markers function as it needs 4 points to work. A slight change has been made to fix_corner_markers to now insert the corner markers at the right index of the list.

tests/test_pregrader.py

@@ -34,7 +34,7 @@ def test_ideal_crops(box_coords, result, scanned_image_keypoints, scanned_image)
 
 @pytest.mark.parametrize('box_coords, result', [((341, 471), True), ((352, 482), True), ((448, 482), True),
                                                 ((423, 474), True), ((460, 475), False), ((477, 474), True),
-                                                ((87, 548), False)],
+                                                ((87, 544), False)],
                          ids=["1 filled bottom right", "1 filled top left", "5 filled with a bit of 6",
                               "4 fully filled with the label", "6 empty with label",
                               "7 partially  cropped, filled and a part of 6", "B empty with cb at the bottom"])