From f996353b374e0e59fa45f7d1747db0c897da9b7d Mon Sep 17 00:00:00 2001
From: duxin <duxin@example.com>
Date: Wed, 8 Jul 2026 16:29:33 +0800
Subject: [PATCH] =?UTF-8?q?fix:=20=E5=85=8B=E9=87=8C=E9=87=91=E7=9F=A9?=
 =?UTF-8?q?=E9=98=B5=E9=B2=81=E6=A3=92=E6=80=A7=20=E2=80=94=20=E5=8E=BB?=
 =?UTF-8?q?=E9=87=8D=E9=87=8D=E5=8F=A0=E7=82=B9=20+=20nugget=20=E9=98=B2?=
 =?UTF-8?q?=E5=A5=87=E5=BC=82=E5=8C=96?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

问题: 4326 点仍崩溃回退 IDW，原因两个:
  1. 采样点中存在空间完全重叠点 → 协方差矩阵行列式为 0
  2. 坐标范围出现负数 (X=-793275) → 经/纬度传反，空间距离扭曲

修复:
  1. read_csv_data: drop_duplicates(subset=['proj_x','proj_y'])
     坐标 <0.01m 的重叠点只保留首个，防止矩阵奇异化
  2. OrdinaryKriging (2处): nugget=1e-6
     微小固有方差强制打破矩阵奇异性，病态矩阵仍可求逆
---
 src/postprocessing/map.py | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/src/postprocessing/map.py b/src/postprocessing/map.py
index dc9558c..1f8fa65 100644
--- a/src/postprocessing/map.py
+++ b/src/postprocessing/map.py
@@ -813,6 +813,7 @@ class ContentMapper:
         ok = OrdinaryKriging(
             local_pts[:, 0], local_pts[:, 1], local_vals,
             variogram_model='spherical',
+            nugget=1e-6,  # 微小 nugget 打破矩阵奇异性，防止协方差矩阵求逆崩溃
             verbose=False,
             enable_plotting=False,
         )
@@ -921,7 +922,15 @@ class ContentMapper:
             print(f"已加载不确定性数据列: {uncertainty_col}")
             print(f"不确定性值范围: {gdf['uncertainty'].min():.4f} - {gdf['uncertainty'].max():.4f}")
 
-        print(f"成功读取 {len(gdf)} 个数据点")
+        # 防御性去重：空间重叠点 (X,Y < 0.01m) 只保留首个，防止克里金矩阵奇异化
+        initial_count = len(gdf)
+        gdf = gdf.drop_duplicates(subset=['proj_x', 'proj_y'], keep='first')
+        final_count = len(gdf)
+        if final_count < initial_count:
+            print(f"  [数学防御] 成功检测并剔除了 {initial_count - final_count} 个空间完全重叠点，"
+                  f"防止克里金矩阵奇异化！")
+
+        print(f"成功读取 {final_count} 个数据点")
         return gdf
 
     def read_boundary_shapefile(self, shp_file):