[Refactor] Camera keys (#805)

* replace all cam_intrinsic

* revert cam_intrinsic in test

* remove rect and fix lint

mmdet3d/apis/inference.py

@@ -423,7 +423,7 @@ def show_proj_det_result_meshlab(data,
             img_metas=data['img_metas'][0][0],
             show=show)
     elif box_mode == Box3DMode.CAM:
-        if 'cam_intrinsic' not in data['img_metas'][0][0]:
+        if 'cam2img' not in data['img_metas'][0][0]:
             raise NotImplementedError(
                 'camera intrinsic matrix is not provided')
 
@@ -434,7 +434,7 @@ def show_proj_det_result_meshlab(data,
             img,
             None,
             show_bboxes,
-            data['img_metas'][0][0]['cam_intrinsic'],
+            data['img_metas'][0][0]['cam2img'],
             out_dir,
             file_name,
             box_mode='camera',

mmdet3d/core/visualizer/image_vis.py

@@ -162,7 +162,7 @@ def draw_depth_bbox3d_on_img(bboxes3d,
 
 def draw_camera_bbox3d_on_img(bboxes3d,
                               raw_img,
-                              cam_intrinsic,
+                              cam2img,
                               img_metas,
                               color=(0, 255, 0),
                               thickness=1):
@@ -172,7 +172,7 @@ def draw_camera_bbox3d_on_img(bboxes3d,
         bboxes3d (:obj:`CameraInstance3DBoxes`, shape=[M, 7]):
             3d bbox in camera coordinate system to visualize.
         raw_img (numpy.array): The numpy array of image.
-        cam_intrinsic (dict): Camera intrinsic matrix,
+        cam2img (dict): Camera intrinsic matrix,
             denoted as `K` in depth bbox coordinate system.
         img_metas (dict): Useless here.
         color (tuple[int]): The color to draw bboxes. Default: (0, 255, 0).
@@ -181,16 +181,16 @@ def draw_camera_bbox3d_on_img(bboxes3d,
     from mmdet3d.core.bbox import points_cam2img
 
     img = raw_img.copy()
-    cam_intrinsic = copy.deepcopy(cam_intrinsic)
+    cam2img = copy.deepcopy(cam2img)
     corners_3d = bboxes3d.corners
     num_bbox = corners_3d.shape[0]
     points_3d = corners_3d.reshape(-1, 3)
-    if not isinstance(cam_intrinsic, torch.Tensor):
-        cam_intrinsic = torch.from_numpy(np.array(cam_intrinsic))
-    cam_intrinsic = cam_intrinsic.reshape(3, 3).float().cpu()
+    if not isinstance(cam2img, torch.Tensor):
+        cam2img = torch.from_numpy(np.array(cam2img))
+    cam2img = cam2img.reshape(3, 3).float().cpu()
 
     # project to 2d to get image coords (uv)
-    uv_origin = points_cam2img(points_3d, cam_intrinsic)
+    uv_origin = points_cam2img(points_3d, cam2img)
     uv_origin = (uv_origin - 1).round()
     imgfov_pts_2d = uv_origin[..., :2].reshape(num_bbox, 8, 2).numpy()
 

mmdet3d/datasets/nuscenes_mono_dataset.py

@@ -628,7 +628,7 @@ class NuScenesMonoDataset(CocoDataset):
                 img,
                 gt_bboxes,
                 pred_bboxes,
-                img_metas['cam_intrinsic'],
+                img_metas['cam2img'],
                 out_dir,
                 file_name,
                 box_mode='camera',

mmdet3d/datasets/pipelines/formating.py

@@ -101,6 +101,7 @@ class Collect3D(object):
         - 'pad_shape': image shape after padding
         - 'lidar2img': transform from lidar to image
         - 'depth2img': transform from depth to image
+        - 'cam2img': transform from camera to image
         - 'pcd_horizontal_flip': a boolean indicating if point cloud is \
             flipped horizontally
         - 'pcd_vertical_flip': a boolean indicating if point cloud is \
@@ -108,13 +109,9 @@ class Collect3D(object):
         - 'box_mode_3d': 3D box mode
         - 'box_type_3d': 3D box type
         - 'img_norm_cfg': a dict of normalization information:
-
             - mean: per channel mean subtraction
             - std: per channel std divisor
             - to_rgb: bool indicating if bgr was converted to rgb
-        - 'rect': rectification matrix
-        - 'Trv2c': transformation from velodyne to camera coordinate
-        - 'P2': transformation betweeen cameras
         - 'pcd_trans': point cloud transformations
         - 'sample_idx': sample index
         - 'pcd_scale_factor': point cloud scale factor
@@ -125,22 +122,22 @@ class Collect3D(object):
         keys (Sequence[str]): Keys of results to be collected in ``data``.
         meta_keys (Sequence[str], optional): Meta keys to be converted to
             ``mmcv.DataContainer`` and collected in ``data[img_metas]``.
-            Default: ('filename', 'ori_shape', 'img_shape', 'lidar2img', \
-            'pad_shape', 'scale_factor', 'flip', 'pcd_horizontal_flip', \
-            'pcd_vertical_flip', 'box_mode_3d', 'box_type_3d', \
-            'img_norm_cfg', 'rect', 'Trv2c', 'P2', 'pcd_trans', \
+            Default: ('filename', 'ori_shape', 'img_shape', 'lidar2img',
+            'depth2img', 'cam2img', 'pad_shape', 'scale_factor', 'flip',
+            'pcd_horizontal_flip', 'pcd_vertical_flip', 'box_mode_3d',
+            'box_type_3d', 'img_norm_cfg', 'pcd_trans',
             'sample_idx', 'pcd_scale_factor', 'pcd_rotation', 'pts_filename')
     """
 
     def __init__(self,
                  keys,
                  meta_keys=('filename', 'ori_shape', 'img_shape', 'lidar2img',
-                            'depth2img', 'pad_shape', 'scale_factor', 'flip',
-                            'cam_intrinsic', 'pcd_horizontal_flip',
+                            'depth2img', 'cam2img', 'pad_shape',
+                            'scale_factor', 'flip', 'pcd_horizontal_flip',
                             'pcd_vertical_flip', 'box_mode_3d', 'box_type_3d',
-                            'img_norm_cfg', 'rect', 'Trv2c', 'P2', 'pcd_trans',
-                            'sample_idx', 'pcd_scale_factor', 'pcd_rotation',
-                            'pts_filename', 'transformation_3d_flow')):
+                            'img_norm_cfg', 'pcd_trans', 'sample_idx',
+                            'pcd_scale_factor', 'pcd_rotation', 'pts_filename',
+                            'transformation_3d_flow')):
         self.keys = keys
         self.meta_keys = meta_keys
 

mmdet3d/datasets/pipelines/loading.py

@@ -90,7 +90,7 @@ class LoadImageFromFileMono3D(LoadImageFromFile):
             dict: The dict contains loaded image and meta information.
         """
         super().__call__(results)
-        results['cam_intrinsic'] = results['img_info']['cam_intrinsic']
+        results['cam2img'] = results['img_info']['cam_intrinsic']
         return results
 
 

mmdet3d/datasets/pipelines/transforms_3d.py

@@ -130,11 +130,10 @@ class RandomFlip3D(RandomFlip):
                 w - input_dict['centers2d'][..., 0]
             # need to modify the horizontal position of camera center
             # along u-axis in the image (flip like centers2d)
-            # ['cam_intrinsic'][0][2] = c_u
+            # ['cam2img'][0][2] = c_u
             # see more details and examples at
             # https://github.com/open-mmlab/mmdetection3d/pull/744
-            input_dict['cam_intrinsic'][0][2] = \
-                w - input_dict['cam_intrinsic'][0][2]
+            input_dict['cam2img'][0][2] = w - input_dict['cam2img'][0][2]
 
     def __call__(self, input_dict):
         """Call function to flip points, values in the ``bbox3d_fields`` and \

mmdet3d/models/dense_heads/fcos_mono3d_head.py

@@ -590,7 +590,7 @@ class FCOSMono3DHead(AnchorFreeMono3DHead):
         Returns:
             tuples[Tensor]: Predicted 3D boxes, scores, labels and attributes.
         """
-        view = np.array(input_meta['cam_intrinsic'])
+        view = np.array(input_meta['cam2img'])
         scale_factor = input_meta['scale_factor']
         cfg = self.test_cfg if cfg is None else cfg
         assert len(cls_scores) == len(bbox_preds) == len(mlvl_points)

mmdet3d/models/detectors/single_stage_mono3d.py

@@ -189,11 +189,10 @@ class SingleStageMono3DDetector(SingleStageDetector):
             if isinstance(data['img_metas'][0], DC):
                 img_filename = data['img_metas'][0]._data[0][batch_id][
                     'filename']
-                cam_intrinsic = data['img_metas'][0]._data[0][batch_id][
-                    'cam_intrinsic']
+                cam2img = data['img_metas'][0]._data[0][batch_id]['cam2img']
             elif mmcv.is_list_of(data['img_metas'][0], dict):
                 img_filename = data['img_metas'][0][batch_id]['filename']
-                cam_intrinsic = data['img_metas'][0][batch_id]['cam_intrinsic']
+                cam2img = data['img_metas'][0][batch_id]['cam2img']
             else:
                 ValueError(
                     f"Unsupported data type {type(data['img_metas'][0])} "
@@ -211,7 +210,7 @@ class SingleStageMono3DDetector(SingleStageDetector):
                 img,
                 None,
                 pred_bboxes,
-                cam_intrinsic,
+                cam2img,
                 out_dir,
                 file_name,
                 'camera',

tests/test_data/test_pipelines/test_loadings/test_loading.py

@@ -304,7 +304,7 @@ def test_load_image_from_file_mono_3d():
         img_info=dict(filename=filename, cam_intrinsic=cam_intrinsic.copy()))
     results = load_image_from_file_mono_3d(input_dict)
     assert results['img'].shape == (900, 1600, 3)
-    assert np.all(results['cam_intrinsic'] == cam_intrinsic)
+    assert np.all(results['cam2img'] == cam_intrinsic)
 
     repr_str = repr(load_image_from_file_mono_3d)
     expected_repr_str = 'LoadImageFromFileMono3D(to_float32=False, ' \

tests/test_models/test_detectors.py

@@ -344,7 +344,7 @@ def test_fcos3d():
     attr_labels = [torch.randint(0, 9, [3], device='cuda')]
     img_metas = [
         dict(
-            cam_intrinsic=[[1260.8474446004698, 0.0, 807.968244525554],
+            cam2img=[[1260.8474446004698, 0.0, 807.968244525554],
                      [0.0, 1260.8474446004698, 495.3344268742088],
                      [0.0, 0.0, 1.0]],
             scale_factor=np.array([1., 1., 1., 1.], dtype=np.float32),

tests/test_models/test_heads/test_heads.py

@@ -1088,7 +1088,7 @@ def test_fcos_mono3d_head():
     attr_labels = [torch.randint(0, 9, [3], device='cuda') for i in range(2)]
     img_metas = [
         dict(
-            cam_intrinsic=[[1260.8474446004698, 0.0, 807.968244525554],
+            cam2img=[[1260.8474446004698, 0.0, 807.968244525554],
                      [0.0, 1260.8474446004698, 495.3344268742088],
                      [0.0, 0.0, 1.0]],
             scale_factor=np.array([1., 1., 1., 1.], dtype=np.float32),

tests/test_runtime/test_apis.py

@@ -189,7 +189,7 @@ def test_show_result_meshlab():
         torch.tensor(
             [[6.4495, -3.9097, -1.7409, 1.5063, 3.1819, 1.4716, 1.8782]]))
     img = np.random.randn(1, 3, 384, 1280)
-    cam_intrinsic = np.array([[100.0, 0.0, 50.0], [0.0, 100.0, 50.0],
+    cam2img = np.array([[100.0, 0.0, 50.0], [0.0, 100.0, 50.0],
                         [0.0, 0.0, 1.0]])
     img_meta = dict(
         filename=filename,
@@ -199,7 +199,7 @@ def test_show_result_meshlab():
         box_type_3d=CameraInstance3DBoxes,
         pcd_trans=np.array([0., 0., 0.]),
         pcd_scale_factor=1.0,
-        cam_intrinsic=cam_intrinsic)
+        cam2img=cam2img)
     data = dict(
         points=[[torch.tensor(points)]], img_metas=[[img_meta]], img=[img])
     result = [

tools/misc/browse_dataset.py

@@ -163,7 +163,7 @@ def show_proj_bbox_img(idx,
             img,
             gt_bboxes,
             None,
-            img_metas['cam_intrinsic'],
+            img_metas['cam2img'],
             out_dir,
             filename,
             box_mode='camera',