shinestacker 1.1.0__py3-none-any.whl → 1.2.1__py3-none-any.whl

shinestacker/_version.py

@@ -1 +1 @@
-__version__ = '1.1.0'
+__version__ = '1.2.1'

shinestacker/algorithms/__init__.py

@@ -8,6 +8,8 @@ from .balance import BalanceFrames
 from .stack import FocusStackBunch, FocusStack
 from .depth_map import DepthMapStack
 from .pyramid import PyramidStack
+from .pyramid_tiles import PyramidTilesStack
+from .pyramid_auto import PyramidAutoStack
 from .multilayer import MultiLayer
 from .noise_detection import NoiseDetection, MaskNoise
 from .vignetting import Vignetting
@@ -16,5 +18,6 @@ logger.addHandler(logging.NullHandler())
 
 __all__ = [
 	'StackJob', 'CombinedActions', 'AlignFrames', 'BalanceFrames', 'FocusStackBunch', 'FocusStack',
-	'DepthMapStack', 'PyramidStack', 'MultiLayer', 'NoiseDetection', 'MaskNoise', 'Vignetting'
+	'DepthMapStack', 'PyramidStack', 'PyramidTilesStack', 'PyramidAutoStack', 'MultiLayer',
+	'NoiseDetection', 'MaskNoise', 'Vignetting'
 ]

shinestacker/algorithms/align.py

@@ -1,10 +1,11 @@
-# pylint: disable=C0114, C0115, C0116, E1101, R0914, R0913, R0917, R0912, R0915, R0902
+# pylint: disable=C0114, C0115, C0116, E1101, R0914, R0913, R0917, R0912, R0915, R0902, E1121, W0102
 import logging
+import math
 import numpy as np
 import matplotlib.pyplot as plt
 import cv2
 from .. config.constants import constants
-from .. core.exceptions import AlignmentError, InvalidOptionError
+from .. core.exceptions import InvalidOptionError
 from .. core.colors import color_str
 from .utils import img_8bit, img_bw_8bit, save_plot, img_subsample
 from .stack_framework import SubAction
@@ -29,6 +30,7 @@ _DEFAULT_ALIGNMENT_CONFIG = {
     'refine_iters': constants.DEFAULT_REFINE_ITERS,
     'align_confidence': constants.DEFAULT_ALIGN_CONFIDENCE,
     'max_iters': constants.DEFAULT_ALIGN_MAX_ITERS,
+    'abort_abnormal': constants.DEFAULT_ALIGN_ABORT_ABNORMAL,
     'border_mode': constants.DEFAULT_BORDER_MODE,
     'border_value': constants.DEFAULT_BORDER_VALUE,
     'border_blur': constants.DEFAULT_BORDER_BLUR,
@@ -44,8 +46,91 @@ _cv2_border_mode_map = {
     constants.BORDER_REPLICATE_BLUR: cv2.BORDER_REPLICATE
 }
 
+_AFFINE_THRESHOLDS = {
+    'max_rotation': 10.0,  # degrees
+    'min_scale': 0.9,
+    'max_scale': 1.1,
+    'max_shear': 5.0,  # degrees
+    'max_translation_ratio': 0.1,  # 10% of image dimension
+}
+
+_HOMOGRAPHY_THRESHOLDS = {
+    'max_skew': 10.0,  # degrees
+    'max_scale_change': 1.5,  # max area change ratio
+    'max_aspect_ratio': 2.0,  # max aspect ratio change
+}
+
+
+def decompose_affine_matrix(m):
+    a, b, tx = m[0, 0], m[0, 1], m[0, 2]
+    c, d, ty = m[1, 0], m[1, 1], m[1, 2]
+    scale_x = math.sqrt(a**2 + b**2)
+    scale_y = math.sqrt(c**2 + d**2)
+    rotation = math.degrees(math.atan2(b, a))
+    shear = math.degrees(math.atan2(-c, d)) - rotation
+    shear = (shear + 180) % 360 - 180
+    return (scale_x, scale_y), rotation, shear, (tx, ty)
+
+
+def check_affine_matrix(m, img_shape, affine_thresholds=_AFFINE_THRESHOLDS):
+    if affine_thresholds is None:
+        return True, "No thresholds provided"
+    (scale_x, scale_y), rotation, shear, (tx, ty) = decompose_affine_matrix(m)
+    h, w = img_shape[:2]
+    reasons = []
+    if abs(rotation) > affine_thresholds['max_rotation']:
+        reasons.append(f"rotation too large ({rotation:.1f}°)")
+    if scale_x < affine_thresholds['min_scale'] or scale_x > affine_thresholds['max_scale']:
+        reasons.append(f"x-scale out of range ({scale_x:.2f})")
+    if scale_y < affine_thresholds['min_scale'] or scale_y > affine_thresholds['max_scale']:
+        reasons.append(f"y-scale out of range ({scale_y:.2f})")
+    if abs(shear) > affine_thresholds['max_shear']:
+        reasons.append(f"shear too large ({shear:.1f}°)")
+    max_tx = w * affine_thresholds['max_translation_ratio']
+    max_ty = h * affine_thresholds['max_translation_ratio']
+    if abs(tx) > max_tx:
+        reasons.append(f"x-translation too large (|{tx:.1f}| > {max_tx:.1f})")
+    if abs(ty) > max_ty:
+        reasons.append(f"y-translation too large (|{ty:.1f}| > {max_ty:.1f})")
+    if reasons:
+        return False, "; ".join(reasons)
+    return True, "Transformation within acceptable limits"
 
-def get_good_matches(des_0, des_1, matching_config=None):
+
+def check_homography_distortion(m, img_shape, homography_thresholds=_HOMOGRAPHY_THRESHOLDS):
+    if homography_thresholds is None:
+        return True, "No thresholds provided"
+    h, w = img_shape[:2]
+    corners = np.array([[0, 0], [w, 0], [w, h], [0, h]], dtype=np.float32)
+    transformed = cv2.perspectiveTransform(corners.reshape(1, -1, 2), m).reshape(-1, 2)
+    reasons = []
+    area_orig = w * h
+    area_new = cv2.contourArea(transformed)
+    area_ratio = area_new / area_orig
+    if area_ratio > homography_thresholds['max_scale_change'] or \
+       area_ratio < 1.0 / homography_thresholds['max_scale_change']:
+        reasons.append(f"area change too large ({area_ratio:.2f})")
+    rect = cv2.minAreaRect(transformed.astype(np.float32))
+    (w_rect, h_rect) = rect[1]
+    aspect_ratio = max(w_rect, h_rect) / min(w_rect, h_rect)
+    if aspect_ratio > homography_thresholds['max_aspect_ratio']:
+        reasons.append(f"aspect ratio change too large ({aspect_ratio:.2f})")
+    angles = []
+    for i in range(4):
+        vec1 = transformed[(i + 1) % 4] - transformed[i]
+        vec2 = transformed[(i - 1) % 4] - transformed[i]
+        angle = np.degrees(np.arccos(np.dot(vec1, vec2) /
+                           (np.linalg.norm(vec1) * np.linalg.norm(vec2))))
+        angles.append(angle)
+    max_angle_dev = max(abs(angle - 90) for angle in angles)
+    if max_angle_dev > homography_thresholds['max_skew']:
+        reasons.append(f"angle distortion too large ({max_angle_dev:.1f}°)")
+    if reasons:
+        return False, "; ".join(reasons)
+    return True, "Transformation within acceptable limits"
+
+
+def get_good_matches(des_0, des_ref, matching_config=None):
     matching_config = {**_DEFAULT_MATCHING_CONFIG, **(matching_config or {})}
     match_method = matching_config['match_method']
     good_matches = []
@@ -54,12 +139,12 @@ def get_good_matches(des_0, des_1, matching_config=None):
             {'algorithm': matching_config['flann_idx_kdtree'],
              'trees': matching_config['flann_trees']},
             {'checks': matching_config['flann_checks']})
-        matches = flann.knnMatch(des_0, des_1, k=2)
+        matches = flann.knnMatch(des_0, des_ref, k=2)
         good_matches = [m for m, n in matches
                         if m.distance < matching_config['threshold'] * n.distance]
     elif match_method == constants.MATCHING_NORM_HAMMING:
         bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)
-        good_matches = sorted(bf.match(des_0, des_1), key=lambda x: x.distance)
+        good_matches = sorted(bf.match(des_0, des_ref), key=lambda x: x.distance)
     else:
         raise InvalidOptionError(
             'match_method', match_method,
@@ -87,14 +172,14 @@ def validate_align_config(detector, descriptor, match_method):
                          " require matching method Hamming distance")
 
 
-def detect_and_compute(img_0, img_1, feature_config=None, matching_config=None):
+def detect_and_compute(img_0, img_ref, feature_config=None, matching_config=None):
     feature_config = {**_DEFAULT_FEATURE_CONFIG, **(feature_config or {})}
     matching_config = {**_DEFAULT_MATCHING_CONFIG, **(matching_config or {})}
     feature_config_detector = feature_config['detector']
     feature_config_descriptor = feature_config['descriptor']
     match_method = matching_config['match_method']
     validate_align_config(feature_config_detector, feature_config_descriptor, match_method)
-    img_bw_0, img_bw_1 = img_bw_8bit(img_0), img_bw_8bit(img_1)
+    img_bw_0, img_bw_ref = img_bw_8bit(img_0), img_bw_8bit(img_ref)
     detector_map = {
         constants.DETECTOR_SIFT: cv2.SIFT_create,
         constants.DETECTOR_ORB: cv2.ORB_create,
@@ -114,12 +199,12 @@ def detect_and_compute(img_0, img_1, feature_config=None, matching_config=None):
                                    constants.DETECTOR_AKAZE,
                                    constants.DETECTOR_BRISK):
         kp_0, des_0 = detector.detectAndCompute(img_bw_0, None)
-        kp_1, des_1 = detector.detectAndCompute(img_bw_1, None)
+        kp_ref, des_ref = detector.detectAndCompute(img_bw_ref, None)
     else:
         descriptor = descriptor_map[feature_config_descriptor]()
         kp_0, des_0 = descriptor.compute(img_bw_0, detector.detect(img_bw_0, None))
-        kp_1, des_1 = descriptor.compute(img_bw_1, detector.detect(img_bw_1, None))
-    return kp_0, kp_1, get_good_matches(des_0, des_1, matching_config)
+        kp_ref, des_ref = descriptor.compute(img_bw_ref, detector.detect(img_bw_ref, None))
+    return kp_0, kp_ref, get_good_matches(des_0, des_ref, matching_config)
 
 
 def find_transform(src_pts, dst_pts, transform=constants.DEFAULT_TRANSFORM,
@@ -151,8 +236,10 @@ def find_transform(src_pts, dst_pts, transform=constants.DEFAULT_TRANSFORM,
     return result
 
 
-def align_images(img_1, img_0, feature_config=None, matching_config=None, alignment_config=None,
-                 plot_path=None, callbacks=None):
+def align_images(img_ref, img_0, feature_config=None, matching_config=None, alignment_config=None,
+                 plot_path=None, callbacks=None,
+                 affine_thresholds=_AFFINE_THRESHOLDS,
+                 homography_thresholds=_HOMOGRAPHY_THRESHOLDS):
     feature_config = {**_DEFAULT_FEATURE_CONFIG, **(feature_config or {})}
     matching_config = {**_DEFAULT_MATCHING_CONFIG, **(matching_config or {})}
     alignment_config = {**_DEFAULT_ALIGNMENT_CONFIG, **(alignment_config or {})}
@@ -163,19 +250,23 @@ def align_images(img_1, img_0, feature_config=None, matching_config=None, alignm
     min_matches = 4 if alignment_config['transform'] == constants.ALIGN_HOMOGRAPHY else 3
     if callbacks and 'message' in callbacks:
         callbacks['message']()
-    h_ref, w_ref = img_1.shape[:2]
+    h_ref, w_ref = img_ref.shape[:2]
     h0, w0 = img_0.shape[:2]
     subsample = alignment_config['subsample']
+    if subsample == 0:
+        img_res = (float(h0) / constants.ONE_KILO) * (float(w0) / constants.ONE_KILO)
+        target_res = constants.DEFAULT_ALIGN_RES_TARGET_MPX
+        subsample = int(1 + math.floor(img_res / target_res))
     fast_subsampling = alignment_config['fast_subsampling']
     min_good_matches = alignment_config['min_good_matches']
     while True:
         if subsample > 1:
             img_0_sub = img_subsample(img_0, subsample, fast_subsampling)
-            img_1_sub = img_subsample(img_1, subsample, fast_subsampling)
+            img_ref_sub = img_subsample(img_ref, subsample, fast_subsampling)
         else:
-            img_0_sub, img_1_sub = img_0, img_1
-        kp_0, kp_1, good_matches = detect_and_compute(img_0_sub, img_1_sub,
-                                                      feature_config, matching_config)
+            img_0_sub, img_ref_sub = img_0, img_ref
+        kp_0, kp_ref, good_matches = detect_and_compute(img_0_sub, img_ref_sub,
+                                                        feature_config, matching_config)
         n_good_matches = len(good_matches)
         if n_good_matches > min_good_matches or subsample == 1:
             break
@@ -191,7 +282,7 @@ def align_images(img_1, img_0, feature_config=None, matching_config=None, alignm
     if n_good_matches >= min_matches:
         transform = alignment_config['transform']
         src_pts = np.float32([kp_0[m.queryIdx].pt for m in good_matches]).reshape(-1, 1, 2)
-        dst_pts = np.float32([kp_1[m.trainIdx].pt for m in good_matches]).reshape(-1, 1, 2)
+        dst_pts = np.float32([kp_ref[m.trainIdx].pt for m in good_matches]).reshape(-1, 1, 2)
         m, msk = find_transform(src_pts, dst_pts, transform, alignment_config['align_method'],
                                 *(alignment_config[k]
                                   for k in ['rans_threshold', 'max_iters',
@@ -199,11 +290,11 @@ def align_images(img_1, img_0, feature_config=None, matching_config=None, alignm
         if plot_path is not None:
             matches_mask = msk.ravel().tolist()
             img_match = cv2.cvtColor(cv2.drawMatches(
-                img_8bit(img_0_sub), kp_0, img_8bit(img_1_sub),
-                kp_1, good_matches, None, matchColor=(0, 255, 0),
+                img_8bit(img_0_sub), kp_0, img_8bit(img_ref_sub),
+                kp_ref, good_matches, None, matchColor=(0, 255, 0),
                 singlePointColor=None, matchesMask=matches_mask,
                 flags=2), cv2.COLOR_BGR2RGB)
-            plt.figure(figsize=(10, 5))
+            plt.figure(figsize=constants.PLT_FIG_SIZE)
             plt.imshow(img_match, 'gray')
             save_plot(plot_path)
             if callbacks and 'save_plot' in callbacks:
@@ -225,6 +316,20 @@ def align_images(img_1, img_0, feature_config=None, matching_config=None, alignm
                 m[:, 2] = translation_fullres
             else:
                 raise InvalidOptionError("transform", transform)
+
+        transform_type = alignment_config['transform']
+        is_valid = True
+        reason = ""
+        if transform_type == constants.ALIGN_RIGID:
+            is_valid, reason = check_affine_matrix(
+                m, img_0.shape, affine_thresholds)
+        elif transform_type == constants.ALIGN_HOMOGRAPHY:
+            is_valid, reason = check_homography_distortion(
+                m, img_0.shape, homography_thresholds)
+        if not is_valid:
+            if callbacks and 'warning' in callbacks:
+                callbacks['warning'](f"invalid transformation: {reason}")
+            return n_good_matches, None, None
         if callbacks and 'align_message' in callbacks:
             callbacks['align_message']()
         img_mask = np.ones_like(img_0, dtype=np.uint8)
@@ -284,7 +389,7 @@ class AlignFrames(SubAction):
     def sub_msg(self, msg, color=constants.LOG_COLOR_LEVEL_3):
         self.process.sub_message_r(color_str(msg, color))
 
-    def align_images(self, idx, img_1, img_0):
+    def align_images(self, idx, img_ref, img_0):
         idx_str = f"{idx:04d}"
         callbacks = {
             'message': lambda: self.sub_msg(': find matches'),
@@ -303,37 +408,47 @@ class AlignFrames(SubAction):
                 f"{self.process.name}-matches-{idx_str}.pdf"
         else:
             plot_path = None
+        if self.alignment_config['abort_abnormal']:
+            affine_thresholds = _AFFINE_THRESHOLDS
+            homography_thresholds = _HOMOGRAPHY_THRESHOLDS
+        else:
+            affine_thresholds = None
+            homography_thresholds = None
         n_good_matches, _m, img = align_images(
-            img_1, img_0,
+            img_ref, img_0,
             feature_config=self.feature_config,
             matching_config=self.matching_config,
             alignment_config=self.alignment_config,
             plot_path=plot_path,
-            callbacks=callbacks
+            callbacks=callbacks,
+            affine_thresholds=affine_thresholds,
+            homography_thresholds=homography_thresholds
         )
         self.n_matches[idx] = n_good_matches
         if n_good_matches < self.min_matches:
-            self.process.sub_message(f": image not aligned, too few matches found: "
-                                     f"{n_good_matches}", level=logging.CRITICAL)
-            raise AlignmentError(idx, f"Image not aligned, too few matches found: "
-                                 f"{n_good_matches} < {self.min_matches}")
+            self.process.sub_message(color_str(f": image not aligned, too few matches found: "
+                                     f"{n_good_matches}", constants.LOG_COLOR_WARNING),
+                                     level=logging.WARNING)
+            return None
         return img
 
     def begin(self, process):
         self.process = process
-        self.n_matches = np.zeros(process.counts)
+        self.n_matches = np.zeros(process.total_action_counts)
 
     def end(self):
         if self.plot_summary:
-            plt.figure(figsize=(10, 5))
+            plt.figure(figsize=constants.PLT_FIG_SIZE)
             x = np.arange(1, len(self.n_matches) + 1, dtype=int)
             no_ref = x != self.process.ref_idx + 1
             x = x[no_ref]
             y = self.n_matches[no_ref]
-            y_max = y[1] \
-                if self.process.ref_idx == 0 \
-                else y[-1] if self.process.ref_idx == len(y) - 1 \
-                else (y[self.process.ref_idx - 1] + y[self.process.ref_idx]) / 2
+            if self.process.ref_idx == 0:
+                y_max = y[1]
+            elif self.process.ref_idx >= len(y):
+                y_max = y[-1]
+            else:
+                y_max = (y[self.process.ref_idx - 1] + y[self.process.ref_idx]) / 2
 
             plt.plot([self.process.ref_idx + 1, self.process.ref_idx + 1],
                      [0, y_max], color='cornflowerblue', linestyle='--', label='reference frame')