The codes and data for "A Graph Convolutional Neural Network-based Method for Predicting Computational Intensity of Geocomputation"

A Graph Convolutional Neural Network-based Method for Predicting Computational Intensity of Geocomputation
This is the implementation for the paper "A Graph Convolutional Neural Network-based Method for Predicting Computational Intensity of Geocomputation".The framework is Learning-based Computing Framework for Geospatial data(LCF-G).Prediction, ParallelComputation and SampleGeneration.This paper includes three case studies, each corresponding to a folder. Each folder contains four subfolders: data, CIThe data folder contains geospatail data.The CIPrediction folder contains model training code.The ParallelComputation folder contains geographic computation code.The SampleGeneration folder contains code for sample generation.Case 1: Generation of DEM from point cloud datastep 1: Data downloadDataset 1 has been uploaded to the directory 1point2dem/data. The other two datasets, Dataset 2 and Dataset 3, can be downloaded from the following website:OpenTopographyBelow are the steps for downloading Dataset 2 and Dataset 3, along with the query parameters:Dataset 2:Visit OpenTopography Website: Go to Dataset 2 Download Link.https://portal.opentopography.org/lidarDataset?opentopoID=OTLAS.112018.2193.1Coordinates & Classification:In the section "1. Coordinates & Classification", select the option "Manually enter selection coordinates".Set the coordinates as follows: Xmin = 1372495.692761,Ymin = 5076006.86821,Xmax = 1378779.529766,Ymax = 5085586.39531Point Cloud Data Download:Under section "2. Point Cloud Data Download", choose the option "Point cloud data in LAS format".Submit:Click on "SUBMIT" to initiate the download.Dataset 3:Visit OpenTopography Website:Go to Dataset 3 Download Link: https://portal.opentopography.org/lidarDataset?opentopoID=OTLAS.052016.26912.1
Coordinates & Classification:
In the section "1. Coordinates & Classification", select the option "Manually enter selection coordinates".Set the coordinates as follows:Xmin = 470047.153826,Ymin = 4963418.512121,Xmax = 479547.16556,Ymax = 4972078.92768Point Cloud Data Download:
Under section "2. Point Cloud Data Download", choose the option "Point cloud data in LAS format".Submit:Click on "SUBMIT" to initiate the download.step 2: Sample generationThis step involves data preparation, and samples can be generated using the provided code. Since the samples have already been uploaded to 1point2dem/SampleGeneration/data, this step is optional.
cd 1point2dem/SampleGenerationg++ PointCloud2DEMSampleGeneration.cpp -o PointCloud2DEMSampleGenerationmpiexec -n {number_processes} ./PointCloud2DEMSampleGeneration ../data/pcd path/to/outputstep 3: Model trainingThis step involves training three models (GAN, CNN, GAT). The model results are saved in 1point2dem/SampleGeneration/result, and the results for Table 3 in the paper are derived from this output.cd 1point2dem/CIPredictionpython -u point_prediction.py --model [GCN|ChebNet|GATNet]step 4: Parallel computationThis step uses the trained models to optimize parallel computation. The results for Figures 11-13 in the paper are generated from the output of this command.cd 1point2dem/ParallelComputationg++ ParallelPointCloud2DEM.cpp -o ParallelPointCloud2DEMmpiexec -n {number_processes} ./ParallelPointCloud2DEM ../data/pcdCase 2: Spatial intersection of vector datastep 1: Data downloadSome data from the paper has been uploaded to 2intersection/data. The remaining OSM data can be downloaded from GeoFabrik. Below are the download steps and parameters:Directly click the following link to download the OSM data: GeoFabrik - Czech Republic OSM Datastep 2: Sample generationThis step involves data preparation, and samples can be generated using the provided code. Since the samples have already been uploaded to 2intersection/SampleGeneration/data, this step is optional.cd 2intersection/SampleGenerationg++ ParallelIntersection.cpp -o ParallelIntersectionmpiexec -n {number_processes} ./ParallelIntersection ../data/shpfile ../data/shpfilestep 3: Model trainingThis step involves training three models (GAN, CNN, GAT). The model results are saved in 2intersection/SampleGeneration/result, and the results for Table 5 in the paper are derived from this output.cd 2intersection/CIPredictionpython -u vector_prediction.py --model [GCN|ChebNet|GATNet]step 4: Parallel computationThis step uses the trained models to optimize parallel computation. The results for Figures 14-16 in the paper are generated from the output of this command.cd 2intersection/ParallelComputationg++ ParallelIntersection.cpp -o ParallelIntersectionmpiexec -n {number_processes} ./ParallelIntersection ../data/shpfile1 ../data/shpfile2Case 3: WOfS analysis using raster datastep 1: Data downloadSome data from the paper has been uploaded to 3wofs/data. The remaining data can be downloaded from http://openge.org.cn/advancedRetrieval?type=dataset. Below are the query parameters:Product Selection: Select LC08_L1TP and LC08_L1GTLatitude and Longitude Selection:Minimum Longitude: 112.5,Maximum Longitude: 115.5, Minimum Latitude: 29.5, Maximum Latitude: 31.5Time Range: 2013-01-01 to 2018-12-31Other parameters: Defaultstep 2: Sample generationThis step involves data preparation, and samples can be generated using the provided code. Since the samples have already been uploaded to 3wofs/SampleGeneration/data, this step is optional.cd 3wofs/SampleGenerationsbt packeagespark-submit --master {host1,host2,host3} --class whu.edu.cn.core.cube.raster.WOfSSampleGeneration path/to/package.jarstep 3: Model trainingThis step involves training three models (GAN, CNN, GAT). The model results are saved in 3wofs/SampleGeneration/result, and the results for Table 6 in the paper are derived from this output.cd 3wofs/CIPredictionpython -u raster_prediction.py --model [GCN|ChebNet|GATNet]step 4: Parallel computationThis step uses the trained models to optimize parallel computation. The results for Figures 18, 19 in the paper are generated from the output of this command.cd 3wofs/ParallelComputationsbt packeagespark-submit --master {host1,host2,host3} --class whu.edu.cn.core.cube.raster.WOfSOptimizedByDL path/to/package.jar path/to/outputStatement about Case 3The experiment Case 3 presented in this paper was conducted with improvements made on the GeoCube platform.Code Name: GeoCubeCode Link: GeoCube Source CodeLicense Information: The GeoCube project is openly available under the CC BY 4.0 license.The GeoCube project is licensed under CC BY 4.0, which is the Creative Commons Attribution 4.0 International License, allowing anyone to freely share, modify, and distribute the platform's code.Citation:
Gao, Fan (2022). A multi-source spatio-temporal data cube for large-scale geospatial analysis. figshare. Software. https://doi.org/10.6084/m9.figshare.15032847.v1Clarification Statement:The authors of this code are not affiliated with this manuscript. The innovations and steps in Case 3, including data download, sample generation, and parallel computation optimization, were independently developed and are not dependent on the GeoCube’s code.RequirementsThe codes use the following dependencies with Python 3.8torch==2.0.0torch_geometric==2.5.3networkx==2.6.3pyshp==2.3.1tensorrt==8.6.1matplotlib==3.7.2scipy==1.10.1scikit-learn==1.3.0geopandas==0.13.2