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

These Scripts use Python2.x

New Addition: Feature_extraction_from_velo_points.ipynb

New Addition: npy2images.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.

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

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

• 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]

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

• 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]

• 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.

• 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])

• 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.

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

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

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

• Extracts all the points form the .NPY file convert to a top-down image, save it and a folder of the images cropped into chuncks of 230x230 to manually label the car/notCar data from the images

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

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