diff --git a/DATA_SuPER/DEM_SuPER.py b/DATA_SuPER/DEM_SuPER.py
index 36a6675..53a643c 100644
--- a/DATA_SuPER/DEM_SuPER.py
+++ b/DATA_SuPER/DEM_SuPER.py
@@ -231,7 +231,7 @@ def process_granule(
         dem_file_list = glob.glob(os.path.join(unzip_dir, f"*{name}.hgt"))
         out_tif_name = f"DEM.{mode_str}.{tile_id}.2000.{name}.tif"
     elif mode_str == "ALOSDEM":
-        dem_file_list = glob.glob(os.path.join(unzip_dir, f"*{name}.tif"))
+        dem_file_list = glob.glob(os.path.join(unzip_dir, "*.dem.tif"))
         out_tif_name = f"DEM.{mode_str}.{tile_id}.2011.{name}.tif"
     output_file = os.path.join(output_dir, out_tif_name)
     if not os.path.isfile(output_file):
@@ -250,12 +250,17 @@ def process_granule(
                     raster.attrs["scale_factor"] = 1
                 dem_raster_list.append(raster)
             if len(dem_raster_list) >= 1:
+                # 先逐一裁剪再镶嵌合成（仅在提供ROI且要求裁剪时执行）
+                if roi is not None and clip:
+                    clipped_list = []
+                    for raster in dem_raster_list:
+                        clipped = clip_image(raster, roi, clip_by_box=True)
+                        clipped_list.append(clipped)
+                # 镶嵌合成
                 if name == "slope" or name == "aspect":
                     dem_mosaiced = mosaic_images(dem_raster_list, nodata=-9999)
                 else:
                     dem_mosaiced = mosaic_images(dem_raster_list, nodata=-32768)
-                if roi is not None and clip:
-                    dem_mosaiced = clip_image(dem_mosaiced, roi, clip_by_box=True)
                 dem_mosaiced.rio.to_raster(
                     output_file, driver="COG", compress="DEFLATE"
                 )
@@ -270,12 +275,13 @@ def process_granule(
 
 def main(
     output_root_dir: str,
-    region: list,
+    region_file: str,
     asset_name: list,
     tile_id: str,
     mode_str: str = "NASADEM",
 ):
-    bbox = tuple(list(region.total_bounds))
+    region = gpd.read_file(region_file)
+    bbox = tuple(list(region.total_bounds)) # 最大外接矩形
     results_urls = []
     mode_str = mode_str.upper()
     output_root_dir = os.path.join(output_root_dir, mode_str)
@@ -297,9 +303,12 @@ def main(
 
     # 默认覆盖上一次检索记录
     results_urls = earthdata_search(asset_name, roi=bbox)
+    if mode_str == "ALOSDEM":
+        # ALOSDEM 数据仅保存含 "FBS" 字符串的 URL
+        results_urls = [url for url in results_urls if "FBS" in url[0]]
     with open(results_urls_file, "w") as f:
         json.dump(results_urls, f)
-    logging.info(f"Found {len(results_urls)} {mode_str} granules.")
+    logging.info(f"Found {len(results_urls)} {mode_str} granules after filtering.")
     if len(results_urls) == 0:
         return
 
@@ -348,13 +357,13 @@ def main(
 
 if __name__ == "__main__":
     earthaccess.login(strategy="netrc", persist=True)
-    # region = gpd.read_file("./data/vectors/wuling_guanqu_polygon.geojson")
+    # region_file = "./data/vectors/wuling_guanqu_polygon.geojson"
     # tile_id = "49REL"
     tile_id = "Wuhan"
-    region = gpd.read_file(f"./data/vectors/{tile_id}.geojson")
+    region_file = f"./data/vectors/{tile_id}.geojson"
     # asset_name = ["NASADEM_HGT", "NASADEM_SC"]
     # mode_str = "NASADEM"
     asset_name = ["ALOS_PSR_RTC_HIGH"]
     mode_str = "ALOSDEM"
     output_root_dir = ".\\data\\DEM"
-    main(output_root_dir, region, asset_name, tile_id, mode_str)
+    main(output_root_dir, region_file, asset_name, tile_id, mode_str)