[Fix] Unify camera poses

configs/imvotenet/imvotenet_stage2_16x8_sunrgbd-3d-10class.py

@@ -193,7 +193,7 @@
         type='Collect3D',
         keys=[
             'img', 'gt_bboxes', 'gt_labels', 'points', 'gt_bboxes_3d',
-            'gt_labels_3d', 'calib'
+            'gt_labels_3d'
         ])
 ]
 
@@ -230,7 +230,7 @@
                 type='DefaultFormatBundle3D',
                 class_names=class_names,
                 with_label=False),
-            dict(type='Collect3D', keys=['img', 'points', 'calib'])
+            dict(type='Collect3D', keys=['img', 'points'])
         ]),
 ]
 # construct a pipeline for data and gt loading in show function
@@ -247,7 +247,7 @@
         type='DefaultFormatBundle3D',
         class_names=class_names,
         with_label=False),
-    dict(type='Collect3D', keys=['img', 'points', 'calib'])
+    dict(type='Collect3D', keys=['img', 'points'])
 ]
 
 data = dict(

mmdet3d/apis/inference.py

@@ -155,23 +155,25 @@ def inference_multi_modality_detector(model, pcd, image, ann_file):
         bbox_fields=[],
         mask_fields=[],
         seg_fields=[])
-
-    # depth map points to image conversion
-    if box_mode_3d == Box3DMode.DEPTH:
-        data.update(dict(calib=info['calib']))
-
     data = test_pipeline(data)
 
+    # TODO: this code is dataset-specific. Move lidar2img and
+    #       depth2img to .pkl annotations in the future.
     # LiDAR to image conversion
     if box_mode_3d == Box3DMode.LIDAR:
         rect = info['calib']['R0_rect'].astype(np.float32)
         Trv2c = info['calib']['Tr_velo_to_cam'].astype(np.float32)
         P2 = info['calib']['P2'].astype(np.float32)
         lidar2img = P2 @ rect @ Trv2c
         data['img_metas'][0].data['lidar2img'] = lidar2img
+    # Depth to image conversion
     elif box_mode_3d == Box3DMode.DEPTH:
-        data['calib'][0]['Rt'] = data['calib'][0]['Rt'].astype(np.float32)
-        data['calib'][0]['K'] = data['calib'][0]['K'].astype(np.float32)
+        rt_mat = info['calib']['Rt']
+        # follow Coord3DMode.convert_point
+        rt_mat = np.array([[1, 0, 0], [0, 0, -1], [0, 1, 0]
+                           ]) @ rt_mat.transpose(1, 0)
+        depth2img = info['calib']['K'] @ rt_mat
+        data['img_metas'][0].data['depth2img'] = depth2img
 
     data = collate([data], samples_per_gpu=1)
     if next(model.parameters()).is_cuda:
@@ -182,9 +184,6 @@ def inference_multi_modality_detector(model, pcd, image, ann_file):
         data['img_metas'] = data['img_metas'][0].data
         data['points'] = data['points'][0].data
         data['img'] = data['img'][0].data
-        if box_mode_3d == Box3DMode.DEPTH:
-            data['calib'][0]['Rt'] = data['calib'][0]['Rt'][0].data
-            data['calib'][0]['K'] = data['calib'][0]['K'][0].data
 
     # forward the model
     with torch.no_grad():
@@ -411,17 +410,13 @@ def show_proj_det_result_meshlab(data,
             box_mode='lidar',
             show=show)
     elif box_mode == Box3DMode.DEPTH:
-        if 'calib' not in data.keys():
-            raise NotImplementedError(
-                'camera calibration information is not provided')
-
         show_bboxes = DepthInstance3DBoxes(pred_bboxes, origin=(0.5, 0.5, 0))
 
         show_multi_modality_result(
             img,
             None,
             show_bboxes,
-            data['calib'][0],
+            None,
             out_dir,
             file_name,
             box_mode='depth',

mmdet3d/core/bbox/box_np_ops.py

@@ -50,7 +50,8 @@ def corners_nd(dims, origin=0.5):
 
     Args:
         dims (np.ndarray, shape=[N, ndim]): Array of length per dim
-        origin (list or array or float): origin point relate to smallest point.
+        origin (list or array or float, optional): origin point relate to
+            smallest point. Defaults to 0.5
 
     Returns:
         np.ndarray, shape=[N, 2 ** ndim, ndim]: Returned corners.
@@ -102,7 +103,10 @@ def center_to_corner_box2d(centers, dims, angles=None, origin=0.5):
     Args:
         centers (np.ndarray): Locations in kitti label file with shape (N, 2).
         dims (np.ndarray): Dimensions in kitti label file with shape (N, 2).
-        angles (np.ndarray): Rotation_y in kitti label file with shape (N).
+        angles (np.ndarray, optional): Rotation_y in kitti label file with
+            shape (N). Defaults to None.
+        origin (list or array or float, optional): origin point relate to
+            smallest point. Defaults to 0.5.
 
     Returns:
         np.ndarray: Corners with the shape of (N, 4, 2).
@@ -173,7 +177,7 @@ def rotation_3d_in_axis(points, angles, axis=0):
     Args:
         points (np.ndarray, shape=[N, point_size, 3]]):
         angles (np.ndarray, shape=[N]]):
-        axis (int): Axis to rotate at.
+        axis (int, optional): Axis to rotate at. Defaults to 0.
 
     Returns:
         np.ndarray: Rotated points.
@@ -208,10 +212,13 @@ def center_to_corner_box3d(centers,
     Args:
         centers (np.ndarray): Locations in kitti label file with shape (N, 3).
         dims (np.ndarray): Dimensions in kitti label file with shape (N, 3).
-        angles (np.ndarray): Rotation_y in kitti label file with shape (N).
-        origin (list or array or float): Origin point relate to smallest point.
-            use (0.5, 1.0, 0.5) in camera and (0.5, 0.5, 0) in lidar.
-        axis (int): Rotation axis. 1 for camera and 2 for lidar.
+        angles (np.ndarray, optional): Rotation_y in kitti label file with
+            shape (N). Defaults to None.
+        origin (list or array or float, optional): Origin point relate to
+            smallest point. Use (0.5, 1.0, 0.5) in camera and (0.5, 0.5, 0)
+            in lidar. Defaults to (0.5, 1.0, 0.5).
+        axis (int, optional): Rotation axis. 1 for camera and 2 for lidar.
+            Defaults to 1.
 
     Returns:
         np.ndarray: Corners with the shape of (N, 8, 3).
@@ -308,8 +315,8 @@ def rotation_points_single_angle(points, angle, axis=0):
 
     Args:
         points (np.ndarray, shape=[N, 3]]):
-        angles (np.ndarray, shape=[1]]):
-        axis (int): Axis to rotate at.
+        angle (np.ndarray, shape=[1]]):
+        axis (int, optional): Axis to rotate at. Defaults to 0.
 
     Returns:
         np.ndarray: Rotated points.
@@ -341,7 +348,8 @@ def points_cam2img(points_3d, proj_mat, with_depth=False):
     Args:
         points_3d (np.ndarray): Points in shape (N, 3)
         proj_mat (np.ndarray): Transformation matrix between coordinates.
-        with_depth (bool): Whether to keep depth in the output.
+        with_depth (bool, optional): Whether to keep depth in the output.
+            Defaults to False.
 
     Returns:
         np.ndarray: Points in image coordinates with shape [N, 2].
@@ -420,8 +428,10 @@ def points_in_rbbox(points, rbbox, z_axis=2, origin=(0.5, 0.5, 0)):
     Args:
         points (np.ndarray, shape=[N, 3+dim]): Points to query.
         rbbox (np.ndarray, shape=[M, 7]): Boxes3d with rotation.
-        z_axis (int): Indicate which axis is height.
-        origin (tuple[int]): Indicate the position of box center.
+        z_axis (int, optional): Indicate which axis is height.
+            Defaults to 2.
+        origin (tuple[int], optional): Indicate the position of
+            box center. Defaults to (0.5, 0.5, 0).
 
     Returns:
         np.ndarray, shape=[N, M]: Indices of points in each box.
@@ -479,11 +489,13 @@ def create_anchors_3d_range(feature_size,
         anchor_range (torch.Tensor | list[float]): Range of anchors with
             shape [6]. The order is consistent with that of anchors, i.e.,
             (x_min, y_min, z_min, x_max, y_max, z_max).
-        sizes (list[list] | np.ndarray | torch.Tensor): Anchor size with
-            shape [N, 3], in order of x, y, z.
-        rotations (list[float] | np.ndarray | torch.Tensor): Rotations of
-            anchors in a single feature grid.
-        dtype (type): Data type. Default to np.float32.
+        sizes (list[list] | np.ndarray | torch.Tensor, optional):
+            Anchor size with shape [N, 3], in order of x, y, z.
+            Defaults to ((1.6, 3.9, 1.56), ).
+        rotations (list[float] | np.ndarray | torch.Tensor, optional):
+            Rotations of anchors in a single feature grid.
+            Defaults to (0, np.pi / 2).
+        dtype (type, optional): Data type. Default to np.float32.
 
     Returns:
         np.ndarray: Range based anchors with shape of \
@@ -520,7 +532,8 @@ def center_to_minmax_2d(centers, dims, origin=0.5):
     Args:
         centers (np.ndarray): Center points.
         dims (np.ndarray): Dimensions.
-        origin (list or array or float): origin point relate to smallest point.
+        origin (list or array or float, optional): Origin point relate
+            to smallest point. Defaults to 0.5.
 
     Returns:
         np.ndarray: Minmax points.
@@ -559,6 +572,8 @@ def iou_jit(boxes, query_boxes, mode='iou', eps=0.0):
     Args:
         boxes (np.ndarray): Input bounding boxes with shape of (N, 4).
         query_boxes (np.ndarray): Query boxes with shape of (K, 4).
+        mode (str, optional): IoU mode. Defaults to 'iou'.
+        eps (float, optional): Value added to denominator. Defaults to 0.
 
     Returns:
         np.ndarray: Overlap between boxes and query_boxes
@@ -648,8 +663,10 @@ def get_frustum(bbox_image, C, near_clip=0.001, far_clip=100):
     Args:
         bbox_image (list[int]): box in image coordinates.
         C (np.ndarray): Intrinsics.
-        near_clip (float): Nearest distance of frustum.
-        far_clip (float): Farthest distance of frustum.
+        near_clip (float, optional): Nearest distance of frustum.
+            Defaults to 0.001.
+        far_clip (float, optional): Farthest distance of frustum.
+            Defaults to 100.
 
     Returns:
         np.ndarray, shape=[8, 3]: coordinates of frustum corners.
@@ -742,12 +759,12 @@ def points_in_convex_polygon_3d_jit(points,
 
     Args:
         points (np.ndarray): Input points with shape of (num_points, 3).
-        polygon_surfaces (np.ndarray): Polygon surfaces with shape of \
-            (num_polygon, max_num_surfaces, max_num_points_of_surface, 3). \
-            All surfaces' normal vector must direct to internal. \
+        polygon_surfaces (np.ndarray): Polygon surfaces with shape of
+            (num_polygon, max_num_surfaces, max_num_points_of_surface, 3).
+            All surfaces' normal vector must direct to internal.
             Max_num_points_of_surface must at least 3.
-        num_surfaces (np.ndarray): Number of surfaces a polygon contains \
-            shape of (num_polygon).
+        num_surfaces (np.ndarray, optional): Number of surfaces a polygon
+            contains shape of (num_polygon). Defaults to None.
 
     Returns:
         np.ndarray: Result matrix with the shape of [num_points, num_polygon].
@@ -772,7 +789,8 @@ def points_in_convex_polygon_jit(points, polygon, clockwise=True):
         points (np.ndarray): Input points with the shape of [num_points, 2].
         polygon (np.ndarray): Input polygon with the shape of
             [num_polygon, num_points_of_polygon, 2].
-        clockwise (bool): Indicate polygon is clockwise.
+        clockwise (bool, optional): Indicate polygon is clockwise. Defaults
+            to True.
 
     Returns:
         np.ndarray: Result matrix with the shape of [num_points, num_polygon].
@@ -821,10 +839,11 @@ def boxes3d_to_corners3d_lidar(boxes3d, bottom_center=True):
       2 -------- 1
 
     Args:
-        boxes3d (np.ndarray): Boxes with shape of (N, 7) \
-            [x, y, z, w, l, h, ry] in LiDAR coords, see the definition of ry \
+        boxes3d (np.ndarray): Boxes with shape of (N, 7)
+            [x, y, z, w, l, h, ry] in LiDAR coords, see the definition of ry
             in KITTI dataset.
-        bottom_center (bool): Whether z is on the bottom center of object.
+        bottom_center (bool, optional): Whether z is on the bottom center
+            of object. Defaults to True.
 
     Returns:
         np.ndarray: Box corners with the shape of [N, 8, 3].

mmdet3d/core/bbox/structures/coord_3d_mode.py

@@ -227,21 +227,11 @@ def convert_point(point, src, dst, rt_mat=None):
             if rt_mat is None:
                 rt_mat = arr.new_tensor([[0, 0, 1], [-1, 0, 0], [0, -1, 0]])
         elif src == Coord3DMode.DEPTH and dst == Coord3DMode.CAM:
-            # LIDAR-CAM conversion is different from DEPTH-CAM conversion
-            # because SUNRGB-D camera calibration files are different from
-            # that of KITTI, and currently we keep this hack
             if rt_mat is None:
                 rt_mat = arr.new_tensor([[1, 0, 0], [0, 0, -1], [0, 1, 0]])
-            else:
-                rt_mat = rt_mat.new_tensor(
-                    [[1, 0, 0], [0, 0, -1], [0, 1, 0]]) @ \
-                    rt_mat.transpose(1, 0)
         elif src == Coord3DMode.CAM and dst == Coord3DMode.DEPTH:
             if rt_mat is None:
                 rt_mat = arr.new_tensor([[1, 0, 0], [0, 0, 1], [0, -1, 0]])
-            else:
-                rt_mat = rt_mat @ rt_mat.new_tensor([[1, 0, 0], [0, 0, 1],
-                                                     [0, -1, 0]])
         elif src == Coord3DMode.LIDAR and dst == Coord3DMode.DEPTH:
             if rt_mat is None:
                 rt_mat = arr.new_tensor([[0, -1, 0], [1, 0, 0], [0, 0, 1]])

mmdet3d/core/bbox/structures/utils.py

@@ -111,12 +111,14 @@ def get_box_type(box_type):
     return box_type_3d, box_mode_3d
 
 
-def points_cam2img(points_3d, proj_mat):
+def points_cam2img(points_3d, proj_mat, with_depth=False):
     """Project points from camera coordicates to image coordinates.
 
     Args:
-        points_3d (torch.Tensor): Points in shape (N, 3)
+        points_3d (torch.Tensor): Points in shape (N, 3).
         proj_mat (torch.Tensor): Transformation matrix between coordinates.
+        with_depth (bool, optional): Whether to keep depth in the output.
+            Defaults to False.
 
     Returns:
         torch.Tensor: Points in image coordinates with shape [N, 2].
@@ -141,6 +143,9 @@ def points_cam2img(points_3d, proj_mat):
         [points_3d, points_3d.new_ones(*points_shape)], dim=-1)
     point_2d = torch.matmul(points_4, proj_mat.t())
     point_2d_res = point_2d[..., :2] / point_2d[..., 2:3]
+
+    if with_depth:
+        return torch.cat([point_2d_res, point_2d[..., 2:3]], dim=-1)
     return point_2d_res
 
 

mmdet3d/core/visualizer/image_vis.py

@@ -120,6 +120,7 @@ def draw_lidar_bbox3d_on_img(bboxes3d,
     return plot_rect3d_on_img(img, num_bbox, imgfov_pts_2d, color, thickness)
 
 
+# TODO: remove third parameter in all functions here in favour of img_metas
 def draw_depth_bbox3d_on_img(bboxes3d,
                              raw_img,
                              calibs,
@@ -137,35 +138,22 @@ def draw_depth_bbox3d_on_img(bboxes3d,
         color (tuple[int]): The color to draw bboxes. Default: (0, 255, 0).
         thickness (int, optional): The thickness of bboxes. Default: 1.
     """
-    from mmdet3d.core import Coord3DMode
     from mmdet3d.core.bbox import points_cam2img
     from mmdet3d.models import apply_3d_transformation
 
     img = raw_img.copy()
-    calibs = copy.deepcopy(calibs)
     img_metas = copy.deepcopy(img_metas)
     corners_3d = bboxes3d.corners
     num_bbox = corners_3d.shape[0]
     points_3d = corners_3d.reshape(-1, 3)
-    assert ('Rt' in calibs.keys() and 'K' in calibs.keys()), \
-        'Rt and K matrix should be provided as camera caliberation information'
-    if not isinstance(calibs['Rt'], torch.Tensor):
-        calibs['Rt'] = torch.from_numpy(np.array(calibs['Rt']))
-    if not isinstance(calibs['K'], torch.Tensor):
-        calibs['K'] = torch.from_numpy(np.array(calibs['K']))
-    calibs['Rt'] = calibs['Rt'].reshape(3, 3).float().cpu()
-    calibs['K'] = calibs['K'].reshape(3, 3).float().cpu()
 
     # first reverse the data transformations
     xyz_depth = apply_3d_transformation(
         points_3d, 'DEPTH', img_metas, reverse=True)
 
-    # then convert from depth coords to camera coords
-    xyz_cam = Coord3DMode.convert_point(
-        xyz_depth, Coord3DMode.DEPTH, Coord3DMode.CAM, rt_mat=calibs['Rt'])
-
     # project to 2d to get image coords (uv)
-    uv_origin = points_cam2img(xyz_cam, calibs['K'])
+    uv_origin = points_cam2img(xyz_depth,
+                               xyz_depth.new_tensor(img_metas['depth2img']))
     uv_origin = (uv_origin - 1).round()
     imgfov_pts_2d = uv_origin[..., :2].reshape(num_bbox, 8, 2).numpy()
 

mmdet3d/datasets/pipelines/formating.py

@@ -100,6 +100,7 @@ class Collect3D(object):
         - 'ori_shape': original shape of the image as a tuple (h, w, c)
         - 'pad_shape': image shape after padding
         - 'lidar2img': transform from lidar to image
+        - 'depth2img': transform from depth to image
         - 'pcd_horizontal_flip': a boolean indicating if point cloud is \
             flipped horizontally
         - 'pcd_vertical_flip': a boolean indicating if point cloud is \
@@ -134,7 +135,7 @@ class Collect3D(object):
     def __init__(self,
                  keys,
                  meta_keys=('filename', 'ori_shape', 'img_shape', 'lidar2img',
-                            'pad_shape', 'scale_factor', 'flip',
+                            'depth2img', 'pad_shape', 'scale_factor', 'flip',
                             'cam_intrinsic', 'pcd_horizontal_flip',
                             'pcd_vertical_flip', 'box_mode_3d', 'box_type_3d',
                             'img_norm_cfg', 'rect', 'Trv2c', 'P2', 'pcd_trans',