Detect Objects Using Deep Learning raster error

[BryceGardnerIA]

Can anyone make sense of this error. I am trying to detect objects on a historic map.

A raster error has occurred. The messages that follow will provide more detail.

Python raster function is unable to vectorize the data. [Failed to generate table]

Python raster function is unable to vectorize the data.

Traceback (most recent call last):

  File "C:\Users\bgardner\AppData\Local\Temp\ArcGISProTemp28396\CentralSouth.dlpk\ArcGISObjectDetector.py", line 218, in vectorize

    polygon_list, scores, classes = self.tta_scale_detect_objects(

                                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "C:\Users\bgardner\AppData\Local\Temp\ArcGISProTemp28396\CentralSouth.dlpk\ArcGISObjectDetector.py", line 293, in tta_scale_detect_objects

    polygon_list, scores, classes = self.tta_detect_objects(

                                    ^^^^^^^^^^^^^^^^^^^^^^^^

  File "C:\Users\bgardner\AppData\Local\Temp\ArcGISProTemp28396\CentralSouth.dlpk\ArcGISObjectDetector.py", line 370, in tta_detect_objects

    pixelBlocksCopy["raster_pixels"] = (out.data * 256).numpy()

                                        ^^^^^^^^

  File "C:\Users\bgardner\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone-2\Lib\site-packages\fastai\vision\image.py", line 214, in data

    return self.px

           ^^^^^^^

  File "C:\Users\bgardner\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone-2\Lib\site-packages\fastai\vision\image.py", line 145, in px

    self.refresh()

  File "C:\Users\bgardner\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone-2\Lib\site-packages\fastai\vision\image.py", line 132, in refresh

    self._px = _grid_sample(self._px, self.flow, **self.sample_kwargs)

                                      ^^^^^^^^^

  File "C:\Users\bgardner\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone-2\Lib\site-packages\fastai\vision\image.py", line 156, in flow

    self._flow = _affine_grid(self.shape)

                 ^^^^^^^^^^^^^^^^^^^^^^^^

  File "C:\Users\bgardner\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone-2\Lib\site-packages\fastai\vision\image.py", line 547, in _affine_grid

    N, C, H, W = size

    ^^^^^^^^^^

ValueError: not enough values to unpack (expected 4, got 3)

Python Raster Function's .vectorize() method returned nothing.

Failed to execute (DetectObjectsUsingDeepLearning).

Failed at Wednesday, March 26, 2025 2:41:38 PM (Elapsed Time: 45.01 seconds)

[DanPatterson]

via code or the builtin tool?

Detect Objects Using Deep Learning (Raster Analysis)—ArcGIS Pro | Documentation

You may have to track back the error to its source during the creation process to get at the one thrown by which-ever method you used.

[BryceGardnerIA]

Hi Dan! Thank you. This is using the built-in tool. I don't know how to track it back since it's built-in. 

[DanPatterson]

By track back, I meant the procedures under the Usage section.  This ensures the inputs are correct for the tool

[ShivaniPathak]

Hi @BryceGardnerIA, could you please confirm which model architecture (i.e. MaskRCNN, SSD, RetinaNet, etc) are you using? 

[adsilb]

@BryceGardnerIA were you able to work around this at all? I'm running into the same issue on all model runs using Pascal VOC metadata. I'm able to get RCNN Masks to work, but that is basically limiting the types of model I can create. Thanks!