Didi Challenge Helper Scripts

Helper Scripts for the Didi Challenge (Team: Transformers!!)

These Scripts use Python2.x

Dataset1

rosbag_cooking.py

• Extracts a sample image from a bag file in the same directory and prints out a summary and information about the bag file.
• Prints a lot of information on the file's PointCloud2 Topic and a sample of the data to terminal.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 rosbag_cooking.py [file_name.bag]

bag2images.py

• Extracts all the images from a given bag file to a folder in the directory of a given name in grayscale or RGB colour.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 bag2images.py [bag_file_name] [new_images_folder_name] ["c" or "g" for colour or gray]

bag2video.py

• Extracts the images in the bag in video format in grayscale or RGB clolour.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 bag2video.py [file_name.bag] [fps] ["c" or "g" for colour or gray]

bag2pointcloud_xyzir.py

• Extracts the PointCloud2 frames in the bag file as an array of frames(topic messages).
• Each member of the output array contains an array of the X - Y - Z - Intensity-Ring values in this order.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 bag2pointcloud_xyzir.py [file_name.bag]
• The output file can be loaded using: numpy.load([npy_file_path])

pointcloud2plot_visual.ipynb

• Extracts the PointCloud2 frames in the .NPY file as an array of frames values (X,Y,Z,Intensity,Ring).
• Plots a bird's-eye view of a sample frame and saves it to an image at 1400dpi.
• Note! VTK and MayaVI need to be on compatible versions with each other.

bag2velo_xy_images.py

• Extracts all the xy axis (bird's-eye view) images from a given bag file to a folder in the directory.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 bag2velo_xy_images.py [bag_file_name] [int_dpi]

Feature_extraction_from_velo_points.ipynb

• A Notebook exploring how to extract features from the velodyne data provided

npy2images.py

• Extracts all the points form the .NPY file, converts them to top-down images, saves the images and a folder of their image cropped into chuncks of 230x230 to manually label the car/notCar data for Classification purposes

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 npy2images.py [npy_file_name]

Dataset2

rosbag_cooking2.py

• Extracts a sample image from a bag file in the same directory and prints out a summary and information about the bag file.
• Prints a lot of information on the file's PointCloud2 Topic and a sample of the data to terminal.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 rosbag_cooking2.py [file_name.bag]

bag2images.py

• Extracts all the images from a given bag file to a folder in the directory of a given name in grayscale or RGB colour.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 bag2images.py [bag_file_name] [new_images_folder_name] ["c" or "g" for colour or gray]

bag2pointcloud_xyzir.py

NOTE: for Dataset2 bag file, you must first use the conversion technique below to convert from velodyne_packets in the available bag file to bag files with point_cloud2

• Extracts the PointCloud2 frames in the bag file as an array of frames(topic messages).
• Each member of the output array contains an array of the [X - Y - Z - Intensity - Ring] values in this order.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 bag2pointcloud_xyzir.py [file_name.bag]
• The output file can be loaded using: numpy.load([npy_file_path])

npy2images.py

• Extracts all the points form the .NPY file, converts them to top-down images, saves the images and a folder of their image cropped into chuncks of 230x230 to manually label the car/notCar data for Classification purposes

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 npy2images.py [npy_file_name]

tinynpy2images.py

• Extracts all the points form the .NPY file, converts them to top-down images, saves the images and a folder of their image cropped into chuncks of 230x230 to manually label the car/notCar data for Classification purposes
• This is useful for using a method outlined by the Captain of the Transformers for Image-by-Image extraction.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 tinynpy2images.py [npy_file_name]

the_extractor_2.py

• An attempt at extracting all the available topics and data into npy, csv, text formats for the sake of simplicity or study.

Usage:

• Run from the command line as follows without brackets and parenthesis: python2 the_extractor_2.py [bag_file_name]

Currently Supports: -extract_tracks() -extract_steering_report() -extract_brake_report() -extract_twist() -extract_objects_gps_fix() -extract_wheel_speed_report() -extract_objects_gps_rtkfix() -extract_time()

Also Checkout...

bag_to_csv

• Converts the topics and types of the rosbag into csv format.

Origin: link to clearpathrobotics!