Compare commits

..

1 Commits

Author SHA1 Message Date
Adriano d9a40952c4 feat: angle_step auto adattivo a dimensione template
Halcon-style: angle_step_deg=0 attiva derivazione automatica
step = atan(2/max_side) deg, clampato [0.5, 10]. Template grande
ottiene step fine, piccolo step grosso. auto_tune emette il valore
calcolato direttamente.

_refine_angle ora usa _effective_angle_step() per coerenza con
training quando la modalita auto e attiva.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-04 15:27:35 +02:00
2 changed files with 30 additions and 13 deletions
+5 -2
View File
@@ -220,8 +220,11 @@ def auto_tune(template_bgr: np.ndarray, mask: np.ndarray | None = None) -> dict:
else:
min_score = 0.45
# angle step: 5° default; se simmetria, mantengo step ma range ridotto
angle_step = 5.0
# angle step adattivo (Halcon-style): atan(2/max_side) deg, clampato.
# Template grande → step fine (rotazione minima visibile su perimetro).
# Template piccolo → step grosso (over-sampling = sprecato).
max_side = max(h, w)
angle_step = float(np.clip(np.degrees(np.arctan2(2.0, max_side)), 1.0, 8.0))
result = {
"backend": "line",
+25 -11
View File
@@ -197,12 +197,31 @@ class LineShapeMatcher:
n = int(np.floor((s1 - s0) / self.scale_step)) + 1
return [float(s0 + i * self.scale_step) for i in range(n)]
def _auto_angle_step(self) -> float:
"""Step angolare derivato da dimensione template (Halcon-style).
Formula: step ≈ atan(2 / max_side) gradi. Garantisce che la
rotazione minima produca uno spostamento di ≥2 px sul perimetro
del template (sotto sample il matching coarse perde candidati).
Clampato in [0.5°, 10°].
"""
max_side = max(self.template_size) if self.template_size != (0, 0) else 64
step = math.degrees(math.atan2(2.0, float(max_side)))
return float(np.clip(step, 0.5, 10.0))
def _effective_angle_step(self) -> float:
"""Risolve angle_step_deg gestendo modalità auto (<=0)."""
if self.angle_step_deg <= 0:
return self._auto_angle_step()
return self.angle_step_deg
def _angle_list(self) -> list[float]:
a0, a1 = self.angle_range_deg
if self.angle_step_deg <= 0 or a0 >= a1:
step = self._effective_angle_step()
if step <= 0 or a0 >= a1:
return [float(a0)]
n = int(np.floor((a1 - a0) / self.angle_step_deg))
return [float(a0 + i * self.angle_step_deg) for i in range(n)]
n = int(np.floor((a1 - a0) / step))
return [float(a0 + i * step) for i in range(n)]
# --- Training ------------------------------------------------------
@@ -415,7 +434,7 @@ class LineShapeMatcher:
if original_score is not None and original_score >= 0.99:
return (angle_deg, original_score, cx, cy)
if search_radius is None:
search_radius = self.angle_step_deg / 2.0
search_radius = self._effective_angle_step() / 2.0
h, w = template_gray.shape
sw = max(16, int(round(w * scale)))
@@ -572,7 +591,6 @@ class LineShapeMatcher:
subpixel: bool = True,
verify_ncc: bool = True,
verify_threshold: float = 0.4,
ncc_skip_above: float = 0.85,
coarse_angle_factor: int = 2,
scale_penalty: float = 0.0,
) -> list[Match]:
@@ -803,14 +821,10 @@ class LineShapeMatcher:
ang_f, score_f, cx_f, cy_f = self._refine_angle(
spread0, bit_active_full, self.template_gray, cx_f, cy_f,
var.angle_deg, var.scale, mask_full,
search_radius=self.angle_step_deg / 2.0,
search_radius=self._effective_angle_step() / 2.0,
original_score=score,
)
# NCC verify lazy (Halcon-style): skip se shape-score gia molto
# alto (probabilita falso positivo trascurabile). NCC e l'op
# piu costosa per match (warp + corr), quindi vale la pena
# saltarlo quando il gradiente shape e gia conclusivo.
if verify_ncc and float(score_f) < ncc_skip_above:
if verify_ncc:
ncc = self._verify_ncc(gray0, cx_f, cy_f, ang_f, var.scale)
if ncc < verify_threshold:
continue