facebookresearch / consistent_depth

We estimate dense, flicker-free, geometrically consistent depth from monocular video, for example hand-held cell phone video.
MIT License
1.63k stars 236 forks source link

Sometimes it finishes. Sometimes not. #12

Open noobtoob4lyfe opened 4 years ago

noobtoob4lyfe commented 4 years ago

I was able to get one video to go all the way through the pipe and output successful video results. Since then I get this error on every other video I try. All my clips are shot with the same camera and are roughly 4 seconds long. Any advice would be appreciated.

"Elapsed time: 16.860 [minutes] 2020-08-26 19:16:37,511 - INFO - #models = 0 Traceback (most recent call last): File "main.py", line 13, in dp.process(params) File "/content/consistent_depth/process.py", line 117, in process return self.pipeline(params) File "/content/consistent_depth/process.py", line 62, in pipeline valid_frames = calibrate_scale(self.video, self.out_dir, frame_range, params) File "/content/consistent_depth/scale_calibration.py", line 184, in calibrate_scale video.path, colmap.sparse_dir(colmap_dir, 0) File "/content/consistent_depth/scale_calibration.py", line 74, in make_camera_params_from_colmap cameras, images, points3D = load_colmap.read_model(path=sparse_dir, ext=".bin") File "/content/consistent_depth/third_party/colmap/scripts/python/read_write_model.py", line 416, in read_model cameras = read_cameras_binary(os.path.join(path, "cameras" + ext)) File "/content/consistent_depth/third_party/colmap/scripts/python/read_write_model.py", line 135, in read_cameras_binary with open(path_to_model_file, "rb") as fid: FileNotFoundError: [Errno 2] No such file or directory: 'results/GH017205/colmap_dense/sparse/0/cameras.bin' Done. Your results are saved at /content/consistent_depth/. Video results are at /content/consistent_depth/results/GH017205/R_hierarchical2_mc/videos/. Disparity maps are at /content/consistent_depth/results/GH017205/R_hierarchical2_mc/B0.1_R1.0_PL1-0_LR0.0004_BS4_Oadam/depth"

roxanneluo commented 4 years ago

Hi,

models=0 means that COLMAP fails to register any model. Typically, COLMAP can fail when the sequence doesn't have enough baseline between views, scenes is too dynamic or lack features. You may also consider pre-calibrate your camera intrinsics to facilitate camera registration.