update

python_server/app/routes/depth.py

@@ -24,9 +24,10 @@ def depth(req: DepthRequest):
     """
     try:
         pil = b64_to_pil_image(req.image_b64).convert("RGB")
+        w, h = pil.size
         predictor = get_depth_predictor()
         depth_arr = predictor(pil)  # type: ignore[misc]
-        png_bytes = depth_to_png16_bytes(np.asarray(depth_arr))
+        png_bytes = depth_to_png16_bytes(np.asarray(depth_arr), target_hw=(h, w))
         return Response(content=png_bytes, media_type="image/png")
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e)) from e

python_server/app/services/depth_io.py

@@ -6,8 +6,16 @@ import numpy as np
 from PIL import Image
 
 
-def depth_to_png16_bytes(depth: np.ndarray) -> bytes:
-    depth = np.asarray(depth, dtype=np.float32)
+def depth_to_png16_bytes(depth: np.ndarray, target_hw: tuple[int, int] | None = None) -> bytes:
+    depth = np.asarray(depth, dtype=np.float32).squeeze()
+    if depth.ndim != 2:
+        raise ValueError(f"depth 应为二维数组，实际 shape={depth.shape}")
+    if target_hw is not None:
+        th, tw = int(target_hw[0]), int(target_hw[1])
+        if th > 0 and tw > 0 and (depth.shape[0] != th or depth.shape[1] != tw):
+            im = Image.fromarray(depth, mode="F")
+            im = im.resize((tw, th), Image.Resampling.BILINEAR)
+            depth = np.asarray(im, dtype=np.float32)
     dmin = float(depth.min())
     dmax = float(depth.max())
     if dmax > dmin: