Data for: Classification methods for point clouds in rock slope monitoring: a novel machine learning approach and comparative analysis
Description
Data for the following submission: Title: Classification methods for point clouds in rock slope monitoring: a novel machine learning approach and comparative analysis Weidner, L.1*, Walton, G.1, Kromer, R.1 1Colorado School of Mines, Golden, USA *Corresponding author email: weidner@mines.edu In the event the manuscript is unavailable, please reach out to us for a copy. The main contents of this file are as follows: -Supplementary figures referenced in the manuscript -All processed point clouds used in the through-time analysis. (~9.3 GB) -Scripts used to calculate the results shown in Figures 11, 12 and 13. (~1.6 GB) -Numeric data in other tables, graphs, and figures. Due to the nature of the research, many large point clouds are created, too many to be all uploaded to this repository. If you are looking for data that is not provided in this dataset, please reach out to the authors and we would be happy to provide any additional data. Scripts labeled "RUNME" are found in the main file directory for creating the ML method results ('tests_RUNME.m'), and for hybrid and masking results. For the most part, scripts can be run without modification and should provide results (assuming the required MATLAB toolboxes are installed) Note that for hybrid and masking, multiple runs of the script are required, changing the filenames at the beginning of the script for each of the four dates calculated. The Random Forest TreeBagger object ('tb_t14_jun16dec18') is also included and all the feature sets used for training and validation ('date_struct.mat').
Files
Steps to reproduce
For the most part, results are obtainable by running the .m scripts without modification (with the exception of the masking and hybrid scripts). Raw point cloud collection and processing are described in the associated manuscript, as well as random forest training hyperparameters. The hyperparameters are also encoded in the TreeBagger object, so you can see exactly what values were used. Note that the Matlab Statistics and Machine Learning toolbox is required to run most if not all of the scripts. If you would like a python version of the code please reach out to us.