update README and minor code cleaning

Author: magehrig

README.md

@@ -26,12 +26,11 @@ Our current conversion code supports 2 event file formats:
 1. Rosbags with [dvs\_msgs](https://github.com/uzh-rpg/rpg_dvs_ros/tree/master/dvs_msgs)
 2. Prophesee Raw Format: Metavision 2.2
 
-First,
-* install [Metavision 2.2](https://docs.prophesee.ai/2.2.0/installation/index.html), if you want to use prophesee raw files.
+First, install [Metavision 2.2](https://docs.prophesee.ai/2.2.0/installation/index.html), if you want to use prophesee raw files.
 Second,
 ```bash
 pip3 install --no-cache-dir -r requirements.txt
-pip3 install --extra-index-url https://rospypi.github.io/simple/ rospy rosbag
+pip3 install --extra-index-url https://rospypi.github.io/simple/ rospy rosbag sensor-msgs
 pip3 install dataclasses # if your system Python version is < 3.7
 ```
 
@@ -55,6 +54,15 @@ The reconstruction code uses events saved in the h5 file format to reconstruct i
 ToDo: link the test file
 
 
+### Reconstructions to Rosbag
+If you want to use [kalibr](https://github.com/ethz-asl/kalibr), you may want to create a rosbag from the reconstructed images.
+To achieve this, additionally install (outside of the conda environment)
+
+```bash
+pip3 install tqdm opencv-python
+pip3 install --extra-index-url https://rospypi.github.io/simple/ sensor-msgs
+```
+
 ## Calibration Procedure
 
 The calibration procedure is based on three steps:

python/images_to_rosbag.py

@@ -1,16 +1,16 @@
-import rosbag
-import cv2
-# from cv_bridge import CvBridge
-from os.path import join
-import rospy
 import argparse
-import os
 import glob
-import tqdm
+import os
+from os.path import join
+
+import cv2
+import rosbag
+import rospy
 from sensor_msgs.msg import Image
+import tqdm
 
-if __name__ == "__main__":
 
+if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("--rosbag_folder", required=True,
                         type=str, help="Path to the base folder containing the rosbags")
@@ -25,7 +25,6 @@
     print('Images to process: {}'.format(args.image_folder))
     reconstructed_images_folder = args.image_folder
 
-    
     if not os.path.exists(args.rosbag_folder):
         os.makedirs(args.rosbag_folder)
 
@@ -58,4 +57,4 @@
                 outbag.write(args.image_topic, rosimage, stamp_ros)
                 pbar.update(1)
             except:
-                print("error in reading file ", image_path)
+                print("error in reading file ", image_path)

python/offline_reconstruction.py

@@ -1,17 +1,18 @@
-import os
-import tqdm
-import torch
-import urllib
 import argparse
+import os
+os.environ['CUDA_DEVICE_ORDER'] = 'PCI_BUS_ID'
 from pathlib import Path
+import urllib
+
+import torch
+import tqdm
+
 from data.provider import DataProvider
 from data.rectimestamps import TimestampProviderFile, TimestampProviderRate
 from reconstruction.utils.loading_utils import load_model, get_device
 from reconstruction.image_reconstructor import ImageReconstructor
 from reconstruction.options.inference_options import set_inference_options
 from reconstruction.utils.voxelgrid import VoxelGrid
-os.environ['CUDA_DEVICE_ORDER'] = 'PCI_BUS_ID'
-
 
 
 def download_checkpoint(path_to_model):
@@ -21,6 +22,7 @@ def download_checkpoint(path_to_model):
         f.write(e2vid_model.read())
     print('Done with downloading!')
 
+
 if __name__ == "__main__":
 
     parser = argparse.ArgumentParser(
@@ -44,7 +46,7 @@ def download_checkpoint(path_to_model):
     if not os.path.isfile(args.h5file):
         print('h5 file not provided')
         exit()
-    
+
     h5_path = Path(args.h5file)
     freq_hz = args.freq_hz
     timestamp_provider = None
@@ -69,7 +71,7 @@ def download_checkpoint(path_to_model):
         os.makedirs(args.output_folder)
     else:
         assert os.path.isdir(args.output_folder)
-    
+
     image_reconstructor = ImageReconstructor(model, args.height, args.width, model.num_bins, args)
     print('== Image reconstruction == ')
     print('Image size: {}x{}'.format(args.height, args.width))