diff --git a/src/pc/gui/main_window.py b/src/pc/gui/main_window.py
index bd4fd57..c5696eb 100644
--- a/src/pc/gui/main_window.py
+++ b/src/pc/gui/main_window.py
@@ -454,6 +454,14 @@
             all_methods,
         )
 
+        self._spin_residual_thresh = self._create_spin(
+            ip.residual_thresh, 0.0, 50.0, 1.0,
+        )
+        self._add_image_row(
+            "残差除去:", self._spin_residual_thresh,
+            all_methods,
+        )
+
         # --- 案C/D: RANSAC ---
         self._spin_ransac_thresh = self._create_spin(
             ip.ransac_thresh, 1.0, 30.0, 1.0,
@@ -684,6 +692,9 @@
             self._spin_neighbor_thresh.setValue(
                 ip.neighbor_thresh,
             )
+            self._spin_residual_thresh.setValue(
+                ip.residual_thresh,
+            )
             self._spin_ransac_thresh.setValue(
                 ip.ransac_thresh,
             )
@@ -751,6 +762,9 @@
         ip.neighbor_thresh = (
             self._spin_neighbor_thresh.value()
         )
+        ip.residual_thresh = (
+            self._spin_residual_thresh.value()
+        )
         # 案C: RANSAC
         ip.ransac_thresh = (
             self._spin_ransac_thresh.value()
diff --git a/src/pc/vision/line_detector.py b/src/pc/vision/line_detector.py
index 2433580..85d77d4 100644
--- a/src/pc/vision/line_detector.py
+++ b/src/pc/vision/line_detector.py
@@ -56,6 +56,7 @@
         width_tolerance: 幅フィルタの上限倍率
         median_ksize: 中心点列の移動メディアンフィルタサイズ（0 で無効）
         neighbor_thresh: 近傍外れ値除去の閾値（px，0 で無効）
+        residual_thresh: 残差反復除去の閾値（px，0 で無効）
         valley_gauss_ksize: 谷検出の行ごとガウシアンカーネルサイズ
         valley_min_depth: 谷として認識する最小深度
         valley_max_deviation: 追跡予測からの最大許容偏差（px）
@@ -97,6 +98,7 @@
     # ロバストフィッティング（全手法共通）
     median_ksize: int = 7
     neighbor_thresh: float = 10.0
+    residual_thresh: float = 8.0
 
     # 透視補正付き幅フィルタ（0 で無効）
     width_near: int = 0
@@ -265,6 +267,7 @@
         binary, params.min_line_width,
         median_ksize=params.median_ksize,
         neighbor_thresh=params.neighbor_thresh,
+        residual_thresh=params.residual_thresh,
     )
 
 
@@ -318,6 +321,7 @@
         binary, params.min_line_width,
         median_ksize=params.median_ksize,
         neighbor_thresh=params.neighbor_thresh,
+        residual_thresh=params.residual_thresh,
     )
 
 
@@ -370,6 +374,7 @@
         ransac_iter=params.ransac_iter,
         median_ksize=params.median_ksize,
         neighbor_thresh=params.neighbor_thresh,
+        residual_thresh=params.residual_thresh,
     )
 
 
@@ -496,6 +501,7 @@
     ransac_iter: int = 0,
     median_ksize: int = 0,
     neighbor_thresh: float = 0.0,
+    residual_thresh: float = 0.0,
 ) -> LineDetectResult:
     """行ごとの中心点に多項式をフィッティングする
 
@@ -511,6 +517,7 @@
         ransac_iter: RANSAC 反復回数
         median_ksize: 移動メディアンのカーネルサイズ
         neighbor_thresh: 近傍外れ値除去の閾値 px
+        residual_thresh: 残差反復除去の閾値 px
 
     Returns:
         線検出の結果
@@ -540,6 +547,7 @@
         cy, cx,
         median_ksize=median_ksize,
         neighbor_thresh=neighbor_thresh,
+        residual_thresh=residual_thresh,
         ransac_thresh=ransac_thresh,
         ransac_iter=ransac_iter,
     )
@@ -603,6 +611,7 @@
     cx: np.ndarray,
     median_ksize: int,
     neighbor_thresh: float,
+    residual_thresh: float = 0.0,
     weights: np.ndarray | None = None,
     ransac_thresh: float = 0.0,
     ransac_iter: int = 0,
@@ -611,14 +620,16 @@
 
     全検出手法で共通に使えるロバストなフィッティング．
     (1) 移動メディアンフィルタでスパイクを平滑化
-    (2) 近傍中央値からの偏差で外れ値を除去
+    (2) 近傍中央値からの偏差で外れ値を除去（複数パス）
     (3) 重み付き最小二乗（または RANSAC）でフィッティング
+    (4) 残差ベースの反復除去で外れ値を最終除去
 
     Args:
         cy: 中心点の y 座標配列
         cx: 中心点の x 座標配列
         median_ksize: 移動メディアンのカーネルサイズ（0 で無効）
         neighbor_thresh: 近傍外れ値除去の閾値 px（0 で無効）
+        residual_thresh: 残差除去の閾値 px（0 で無効）
         weights: 各点の信頼度（None で均等）
         ransac_thresh: RANSAC 閾値（0 以下で無効）
         ransac_iter: RANSAC 反復回数
@@ -641,15 +652,26 @@
             hi = min(len(cx_clean), i + half + 1)
             cx_clean[i] = float(np.median(cx[lo:hi]))
 
-    # (2) 近傍外れ値除去
+    # (2) 近傍外れ値除去（複数パス）
     if neighbor_thresh > 0:
         half_n = 3
-        for i in range(len(cx_clean)):
-            lo = max(0, i - half_n)
-            hi = min(len(cx_clean), i + half_n + 1)
-            local_med = float(np.median(cx_clean[lo:hi]))
-            if abs(cx_clean[i] - local_med) > neighbor_thresh:
-                mask[i] = False
+        for _ in range(3):
+            new_mask = np.ones(len(cx_clean), dtype=bool)
+            for i in range(len(cx_clean)):
+                if not mask[i]:
+                    continue
+                lo = max(0, i - half_n)
+                hi = min(len(cx_clean), i + half_n + 1)
+                neighbors = cx_clean[lo:hi][mask[lo:hi]]
+                if len(neighbors) == 0:
+                    new_mask[i] = False
+                    continue
+                local_med = float(np.median(neighbors))
+                if abs(cx_clean[i] - local_med) > neighbor_thresh:
+                    new_mask[i] = False
+            if np.array_equal(mask, mask & new_mask):
+                break
+            mask = mask & new_mask
 
         cy = cy[mask]
         cx_clean = cx_clean[mask]
@@ -661,15 +683,39 @@
 
     # (3) フィッティング
     if ransac_thresh > 0 and ransac_iter > 0:
-        return _ransac_polyfit(
+        coeffs = _ransac_polyfit(
             cy, cx_clean, 2, ransac_iter, ransac_thresh,
         )
+    elif weights is not None:
+        coeffs = np.polyfit(cy, cx_clean, 2, w=weights)
+    else:
+        coeffs = np.polyfit(cy, cx_clean, 2)
 
-    # 重み付き最小二乗
-    if weights is not None:
-        return np.polyfit(cy, cx_clean, 2, w=weights)
+    if coeffs is None:
+        return None
 
-    return np.polyfit(cy, cx_clean, 2)
+    # (4) 残差ベースの反復除去
+    if residual_thresh > 0:
+        for _ in range(5):
+            poly = np.poly1d(coeffs)
+            residuals = np.abs(cx_clean - poly(cy))
+            inlier = residuals < residual_thresh
+            if np.all(inlier):
+                break
+            if np.sum(inlier) < MIN_FIT_ROWS:
+                break
+            cy = cy[inlier]
+            cx_clean = cx_clean[inlier]
+            if weights is not None:
+                weights = weights[inlier]
+            if weights is not None:
+                coeffs = np.polyfit(
+                    cy, cx_clean, 2, w=weights,
+                )
+            else:
+                coeffs = np.polyfit(cy, cx_clean, 2)
+
+    return coeffs
 
 
 def _no_detection(
@@ -1008,6 +1054,7 @@
         cy, cx,
         median_ksize=params.median_ksize,
         neighbor_thresh=params.neighbor_thresh,
+        residual_thresh=params.residual_thresh,
         weights=w_arr,
         ransac_thresh=params.ransac_thresh,
         ransac_iter=params.ransac_iter,