# Processing point cloud data

 - Beagle use Hadoop map-reduce for processing point cloud data, read data from HDFS, generate LOD PCD.
 - Old documentation [confluence - Beagle data storage model and implimentation](https://conworth.atlassian.net/wiki/spaces/CONWORTH/pages/244383745/B-Eagle+s+Data+Storage+Model+Implementation)

## Procedure:
 
### Upload file to HDFS:

 - For text file you can upload to HDFS directly. See  [how to upload](./storing_data.md)
 - For las/laz file you have to split your data first ( <4000000 points for each), then upload splitted data.
 - Run hadoop map-reduce job
 
#### Split files

 - You can use lassplit to split las/laz files

```bash
lassplit "D:\<folder_name>\*.las" -merged -split 4000000 -o ./<file_name>.las
```

### Run hadoop map-reduce job
```
 ![beagle tiler parameters](//_static/_images/beagle_tiler_params.png)
```bash
/hadoop/bin/hadoop jar /hadoop/share/hadoop/beagle/beagle-tiler-2.0.jar com.beagle.tiler.task.BTask_Tiler2 -u m -so false -sp 0.0001 -sa POISSON -i hdfs://165.132.137.159:9000/beagle/Test.bprj/data/SDV_disseration/pcd_skew_10m.pts -sc true -bs 128
```