diff --git a/src/transformers/models/owlv2/image_processing_owlv2.py b/src/transformers/models/owlv2/image_processing_owlv2.py
index a79cc57a6c94..a272c8fdbce7 100644
--- a/src/transformers/models/owlv2/image_processing_owlv2.py
+++ b/src/transformers/models/owlv2/image_processing_owlv2.py
@@ -291,7 +291,7 @@ def pad(
         image = pad(
             image=image,
             padding=((0, size - height), (0, size - width)),
-            constant_values=0.5,
+            constant_values=0.0,
             data_format=data_format,
             input_data_format=input_data_format,
         )
diff --git a/src/transformers/models/owlv2/image_processing_owlv2_fast.py b/src/transformers/models/owlv2/image_processing_owlv2_fast.py
index 25022f4f6c8c..f1a8a79fb81e 100644
--- a/src/transformers/models/owlv2/image_processing_owlv2_fast.py
+++ b/src/transformers/models/owlv2/image_processing_owlv2_fast.py
@@ -228,7 +228,7 @@ def post_process_image_guided_detection(self, outputs, threshold=0.0, nms_thresh
 
         return results
 
-    def _pad_images(self, images: "torch.Tensor", constant_value: float = 0.5) -> "torch.Tensor":
+    def _pad_images(self, images: "torch.Tensor", constant_value: float = 0.0) -> "torch.Tensor":
         """
         Pad an image with zeros to the given size.
         """
@@ -245,7 +245,7 @@ def pad(
         self,
         images: list["torch.Tensor"],
         disable_grouping: Optional[bool],
-        constant_value: float = 0.5,
+        constant_value: float = 0.0,
         **kwargs,
     ) -> list["torch.Tensor"]:
         """
@@ -351,7 +351,7 @@ def _preprocess(
         processed_images = reorder_images(processed_images_grouped, grouped_images_index)
 
         if do_pad:
-            processed_images = self.pad(processed_images, constant_value=0.5, disable_grouping=disable_grouping)
+            processed_images = self.pad(processed_images, constant_value=0.0, disable_grouping=disable_grouping)
 
         grouped_images, grouped_images_index = group_images_by_shape(
             processed_images, disable_grouping=disable_grouping
diff --git a/src/transformers/models/owlv2/modular_owlv2.py b/src/transformers/models/owlv2/modular_owlv2.py
index c58db1efd46e..590fa5b4b31c 100644
--- a/src/transformers/models/owlv2/modular_owlv2.py
+++ b/src/transformers/models/owlv2/modular_owlv2.py
@@ -52,7 +52,7 @@ class Owlv2ImageProcessorFast(OwlViTImageProcessorFast):
     crop_size = None
     do_center_crop = None
 
-    def _pad_images(self, images: "torch.Tensor", constant_value: float = 0.5) -> "torch.Tensor":
+    def _pad_images(self, images: "torch.Tensor", constant_value: float = 0.0) -> "torch.Tensor":
         """
         Pad an image with zeros to the given size.
         """
@@ -69,7 +69,7 @@ def pad(
         self,
         images: list["torch.Tensor"],
         disable_grouping: Optional[bool],
-        constant_value: float = 0.5,
+        constant_value: float = 0.0,
         **kwargs,
     ) -> list["torch.Tensor"]:
         """
@@ -175,7 +175,7 @@ def _preprocess(
         processed_images = reorder_images(processed_images_grouped, grouped_images_index)
 
         if do_pad:
-            processed_images = self.pad(processed_images, constant_value=0.5, disable_grouping=disable_grouping)
+            processed_images = self.pad(processed_images, constant_value=0.0, disable_grouping=disable_grouping)
 
         grouped_images, grouped_images_index = group_images_by_shape(
             processed_images, disable_grouping=disable_grouping